AI Lessons

Understand the definition of AI, its history from the Turing Test to modern LLMs, and how AI differs from machine learning and deep learning.

Explore supervised, unsupervised, and reinforcement learning — and why data quality determines AI quality.

How neural networks are structured, how they learn through backpropagation, and how CNNs, RNNs, and Transformers differ.

How attention mechanisms, tokenization, pretraining, fine-tuning, and RLHF combine to create the large language models powering modern AI.

Master the core techniques for communicating effectively with AI — from zero-shot prompting to chain-of-thought reasoning and system prompts.

Explore how bias enters AI systems, what responsible AI principles look like in practice, and the key governance frameworks shaping the industry.

A hands-on guide to having your first conversation with an AI chatbot. Learn how to craft effective prompts, compare responses across tools, and start using AI in your daily work.

How AI is transforming cybersecurity — from threat detection and behavioral analysis to AI-powered attacks and the escalating defender-attacker dynamic.

How AI is transforming medicine — from AlphaFold's protein structure breakthrough to AI medical imaging, clinical documentation, and drug discovery at scale. May 2026 saw the launch of Medicare ACCESS, the first US payment model designed around AI agents in clinical care.

How AI is transforming financial services — from algorithmic trading and AI-powered banking to insurance underwriting, claims automation, and computer vision for damage assessment.

How AI is transforming sales and marketing — from AI-powered CRM and revenue intelligence to personalization at scale — and AI's role in the cryptocurrency ecosystem.

How AI is enabling autonomous vehicles, delivery drones, and humanoid robots — from Waymo's commercialized robotaxis to Tesla's Optimus and the companies racing to build general-purpose physical AI.

How AI is transforming the energy sector — from grid optimization and renewable energy forecasting to nuclear fusion timelines and the massive energy demand that AI itself is creating.

How AI is transforming military and national security operations — from Palantir's battlefield intelligence to Anduril's autonomous systems, and the profound ethical questions that AI-enabled warfare raises.

How AI is reshaping four major industries — adaptive learning platforms in education, AI-native legal research and contract tools, personalization and logistics in retail, and predictive analytics in real estate.

Understand how AI is transforming the nature of work — which tasks are being automated, which roles are resilient, and why augmentation is the most likely outcome for most workers.

A detailed look at which roles face the highest automation risk, which are being augmented, and which new roles are growing because of AI.

A practical framework for evaluating your own role's AI exposure and making strategic career decisions — including the T-shaped professional model and a 2x2 automation risk matrix.

Understand which human capabilities AI cannot easily replicate — and how to intentionally develop them as your most durable professional assets.

Eight concrete, actionable strategies for building a career that thrives alongside AI — not despite it — including portfolio development, network building, and the deliberate cultivation of AI-resilient skills.

A comprehensive overview of OpenAI's model portfolio — from GPT-5.5 (April 23, 2026, the new agentic flagship) through Codex and GPT Image 2 — and the company's unique position as the maker of both the most-used AI product and the most capable frontier models.

Understand Anthropic's origin, its safety-first research mission, the Claude model family — Opus 4.7, Sonnet, and Haiku — and the latest developments including Claude Mythos Preview, Claude Cowork, Managed Agents, and Claude Design.

Explore Google DeepMind's comprehensive model portfolio — Gemini 3 Pro and Flash, Gemma 4, Nano Banana image generation, and Veo 3 video — and understand Google's unique advantages in AI infrastructure and data.

Explore xAI's approach to AI — now merged with SpaceX in a $1.25 trillion combined entity — the Colossus data center, real-time X/Twitter data, the Grok model family including Grok 4.20's multi-agent capabilities, the new Terafab chip foundry joint venture, and the April 2026 SpaceX option to acquire Cursor that extends the AI coding stack.

Understand Microsoft's dual AI strategy — its OpenAI partnership (now restructured by the April 2026 amendment as a primary-not-exclusive cloud relationship), and its own Phi series of small, highly capable models — plus Microsoft 365 Copilot as the most-deployed enterprise AI product globally.

Explore Meta's evolving AI strategy — from open-source Llama models to the proprietary Muse Spark flagship from Meta Superintelligence Labs — and how Llama 4 Maverick, Llama 4 Scout, and Llama 3.3 70 billion fit different deployment scenarios.

Explore Amazon's AI model portfolio — the Nova series — and Amazon Bedrock's role as the enterprise multi-model platform hosting 50+ models including Claude, Llama, and Amazon's own models.

Understand what open and closed source actually mean for AI models, the strategic reasons companies choose each approach, and a practical framework for deciding which to use in different situations.

Explore Mistral AI — Europe's leading foundation model company — its open-source models, GDPR-compliant hosted API, and its strategic role in EU AI sovereignty.

Survey China's leading foundation models — DeepSeek, Qwen, Kimi, Ernie, GLM, Doubao, and MiniMax — and understand how the DeepSeek + Moonshot $45-$20 billion mega-rounds in May 2026 reshaped global AI economics.

Survey international AI models from Canada, the UAE, the UK, and beyond — understand the US-China AI race, EU regulation, and develop a framework for thinking about AI development as a geopolitical contest.

The primary chat interfaces for US-based foundation models — ChatGPT, Claude, Gemini, and peers — with a practical guide to their distinct strengths and when to use each.

AI chatbots from outside the US — from China's DeepSeek and Qwen to France's Mistral and Canada's Cohere — offer distinct capabilities, different data governance, and in some cases dramatically lower costs.

AI image generation has advanced dramatically — from text-to-image pioneers to reasoning-native models that plan compositions before they draw, render text accurately across scripts, and search the web for facts they don't know. OpenAI's GPT Image 2 (April 2026) tops the Image Arena leaderboard by +242 points, the largest recorded lead.

AI video generation has split into three distinct categories: text-to-video cinematic models, AI avatar and presenter tools for corporate content, and AI-powered editing tools that transform existing footage — each with different quality levels and use cases.

AI voice and audio tools span voice cloning, speech recognition, dictation, music generation, and audio enhancement — with ElevenLabs leading TTS, OpenAI Whisper leading open-source transcription, Wispr Flow leading consumer dictation, and Suno/Udio transforming music creation.

AI productivity tools have moved from novelty to daily necessity — with M365 Copilot, Google Workspace AI, Claude Cowork, Claude Design, and Perplexity Computer reshaping how professionals work, while NotebookLM and Otter.ai solve specific high-value workflows.

Cloud storage tools are increasingly adding AI-powered search and summarization, while developer-oriented object storage (S3, Backblaze B2) remains the foundation for large-scale data storage in AI applications.

AI research tools span consumer search assistants, autonomous multi-source research agents, academic literature tools, and developer APIs — with Perplexity and ChatGPT Deep Research leading the consumer category while Elicit and Consensus serve the scientific research community.

AI-powered web scraping tools have simplified data extraction from structured and unstructured web content — with Firecrawl optimized for LLM pipelines and Apify providing enterprise-scale scraping infrastructure.

Automation tools range from no-code workflow builders (Zapier, Make, n8n) to developer frameworks for multi-step AI agents (LangChain, AG2) — with the right choice depending entirely on whether you're connecting existing tools or building new AI systems.

Computer control AI — systems that can operate a full desktop by seeing screenshots and taking actions — represents one of the most powerful and most security-sensitive categories of AI capability.

Browser-integrated AI tools bring AI assistance directly into your web browsing — reading page content, taking actions on your behalf, and connecting what you're looking at to your broader workflow.

Vector databases and Retrieval-Augmented Generation (RAG) solve LLMs' most practical limitation — the inability to access your specific data — by enabling AI to search and reason over your documents, knowledge bases, and custom content.

Foundation models are the large-scale AI systems trained on massive datasets that power chatbots, coding tools, and creative applications. This category covers models you can download, run locally, or access via API — from open-source options like Gemma and Phi to enterprise platforms like Amazon Bedrock.

ChatGPT is OpenAI's flagship AI chat interface — the world's most-used AI product with 400 million+ weekly users — now defaulting to GPT-5.5 Instant across every tier, with image generation, voice conversation, Deep Research, an opt-in Trusted Contact self-harm safeguard, Plaid-routed bank-account access for Pro subscribers, and content-provenance signals on every AI-generated image via C2PA metadata and Google's SynthID watermark.

Claude.ai is Anthropic's flagship AI interface — combining a 1 million-token context window, exceptional coding and document analysis, and a safety-first design that makes it the go-to choice for long-context and professional workflows. Anthropic has expanded into vertical product lines (Claude for Legal, Claude for Small Business), is reportedly in talks for a Series H at a $900 billion+ valuation, and is the defendant in Bartz v. Anthropic — the largest known US copyright settlement at $1.5 billion.

Gemini is Google's flagship AI interface — combining a 1 million-token context window, native multimodal inputs, deep Google Workspace integration, and built-in access to Google Search, making it the strongest choice for users embedded in the Google ecosystem. The May 2026 Google I/O rollout defaults Google Search's AI Mode to Gemini 3.5 Flash worldwide across 98 languages and rebuilds Search around agents — adding always-on information watchers, agentic booking, generative UI mini-apps, and a Universal Cart that follows a single shopping session across the open web.

Grok is xAI's flagship AI interface — combining a 2 million-token context window, real-time X/Twitter data access, native Aurora image generation, and a direct, unfiltered communication style that sets it apart from the major AI chatbots.

Microsoft Copilot is Microsoft's AI assistant — available free at Copilot.microsoft.com and deeply embedded across Windows, Edge, and Microsoft 365 — making it the most widely pre-installed AI interface on the planet.

Meta AI is Meta's AI assistant — built on the Llama 4 and Muse Spark model family and embedded across Facebook, Instagram, WhatsApp, and Messenger. The casual-use tier remains free with no usage cap, with optional Meta One Plus and Premium paid tiers that unlock deeper reasoning and expanded image and video generation.

Perplexity AI is the answer engine that cites every claim — combining real-time web search with LLM synthesis to deliver sourced, verifiable answers in a format that directly competes with Google Search for research-heavy queries.

Mistral Vibe is the unified agent platform that replaced Le Chat — combining a Work mode that plans tasks across Google Workspace, Outlook, SharePoint, Slack, and GitHub with a Code mode that runs coding agents in isolated sandboxes from a web UI, the CLI, or a VS Code extension. It is the most capable end-user AI platform built and hosted entirely within the European Union.

Qwen is Alibaba's open-weight AI model family — with the Qwen 3.5 flagship reaching 397 billion parameters (17 billion active) using a novel Gated DeltaNet+MoE architecture and supporting 100+ languages — making it one of the most versatile and widely accessible international AI systems available.

DeepSeek is the Chinese AI lab that shocked the industry in early 2025 by releasing a frontier-class reasoning model trained for a fraction of the cost of comparable US models. April 2026: DeepSeek shipped V4-Pro (1.6 trillion-parameter MoE, 1 million-token context) and V4-Flash (284 billion total / 13 billion active) — both MIT-licensed and priced well below frontier rivals.

Kimi is Moonshot AI's AI assistant — now powered by Kimi K2.6 (May 2026), the second-most-used model on OpenRouter, shipped alongside Moonshot's $2 billion raise at a $20 billion valuation. The K2 family combines a 256K context window, natively multimodal 1 trillion-parameter MoE architecture, and the open-source Kimi Code coding tool.

Ernie Bot is Baidu's flagship AI assistant — China's most-used AI chatbot among the general public, now powered by ERNIE 5.0 (2.4 trillion parameters, unified multimodal) and deeply integrated with Baidu Search for real-time, grounded Chinese-language responses.

Cohere Coral is the enterprise AI assistant built on Cohere's Command A model — purpose-built for retrieval-augmented generation (RAG) in regulated industries, with the North agentic platform, native multi-cloud deployment, and connectors to enterprise data sources.

GPT Image 2 (also branded ChatGPT Images 2.0) is OpenAI's current flagship image model — the first image generator with built-in O-series reasoning, character-level multilingual text rendering, and web-search grounding before it draws a single pixel.

Nano Banana 2 is Google's real-time AI image synthesis model — capable of generating images at 30 fps with sub-500ms latency, accurate text rendering, character consistency, and Personal Intelligence integration with Google Photos for personalized image generation.

Adobe Firefly is Adobe's AI image generation suite, built into Photoshop, Illustrator, and Adobe Express — trained exclusively on licensed content to be commercially safe, making it the go-to choice for professional designers who need to own their output.

Canva AI brings image generation directly into the world's most popular design platform — letting anyone create professional-quality visuals without design experience, backed by Canva's vast library of templates and design elements.

Flux is Black Forest Labs' family of open-weight and commercial image generation models — offering best-in-class photorealism, strong prompt adherence, and flexible deployment options from fully open-source local use to API access via third-party platforms.

Midjourney is the gold standard for AI-generated art and illustration — producing images with exceptional aesthetic quality and stylistic depth that no other model consistently matches, available exclusively via a paid subscription.

Stable Diffusion is the foundational open-source image generation model — free to download, self-host, and fine-tune, with a vast ecosystem of thousands of community-trained model variants that cover virtually every visual style imaginable.

Ideogram is an AI image generation platform with a particular strength in accurate text rendering within images — making it a practical choice for generating social media graphics, posters, and any design where readable on-image text is required.

NightCafe is a community-centered AI art creation platform that provides access to multiple underlying image generation models — including Stable Diffusion and DALL-E — through a beginner-friendly interface with social sharing, daily free credits, and an active creator community.

Recraft is an AI design tool built for professional creatives — offering vector image generation, brand style consistency, and a design-oriented workflow that goes beyond simple text-to-image to produce assets ready for professional use.

Sora was OpenAI's text-to-video model — capable of generating cinematic-quality video with synchronized audio. OpenAI announced on March 24, 2026 that Sora is being shut down, along with the iOS app, API, and sora.com.

Veo 3 is Google DeepMind's text-to-video model — producing highly photorealistic video with native audio synthesis, and available through Google's VideoFX, Gemini, and Vertex AI for enterprise deployment.

Runway ML is a professional AI video creation platform built for creative teams — offering text-to-video, image-to-video, video editing, and a growing suite of AI media tools in a single web-based workspace used by filmmakers and studios worldwide.

HeyGen is the leading AI avatar video platform — enabling businesses and creators to produce professional talking-head videos with realistic AI presenters, instant video translation, and custom avatar cloning without cameras, studios, or editing software.

Kling AI is a text-to-video and image-to-video model from Chinese tech company Kuaishou — notable for producing long-form video clips with strong physical realism, available globally via a free-to-access web platform.

Synthesia is an enterprise AI video platform for corporate training, onboarding, and communications — offering a large library of professional AI avatars, multi-language support across 140+ languages, and a template-based editor designed for non-video-production teams.

Pika Labs is a fast, accessible AI video generation platform — popular with social media creators and casual users for quick text-to-video and image-to-video generation, with a generous free tier and a Discord-based community that accelerated its early growth.

Descript is an AI-powered video and podcast editor that lets you edit audio and video by editing text — combining transcription, multi-track editing, AI voice cloning (Overdub), filler word removal, and screen recording in a single desktop application used by podcasters, YouTubers, and content teams.

ElevenLabs is the leading AI voice platform — offering ultra-realistic text-to-speech in 30+ languages, instant voice cloning from a one-minute sample, and real-time voice conversion for creators, publishers, and developers.

OpenAI offers two complementary audio models: Whisper for best-in-class speech-to-text transcription in 99 languages (open-source and API), and TTS for natural text-to-speech synthesis — both accessible as APIs and embedded in ChatGPT.

Suno AI is the leading AI music generation platform — describe a song in plain language and get a complete, original track with vocals, instrumentation, and structure in seconds, across any genre.

Adobe Podcast Enhance is a free AI audio tool that removes background noise and makes any microphone recording sound studio-quality in seconds — the fastest way to clean up podcast, video, or voice-over audio without audio engineering experience.

Udio is an AI music generation platform that competes directly with Suno — known for high production quality and strong stylistic precision, particularly for specific subgenres and niche musical styles, with a generous free tier.

Murf AI is a professional text-to-speech platform built for voiceover production — offering 120+ studio-quality AI voices across 20+ languages, a built-in script editor, and media sync for corporate training, e-learning, and video content teams.

Microsoft 365 Copilot embeds GPT-5.5 directly into Word, Excel, PowerPoint, Outlook, and Teams — turning the world's most widely used office suite into an AI-powered productivity platform for hundreds of millions of enterprise and business users.

NotebookLM is Google's AI research and note-taking tool that lets you upload your own documents, PDFs, and URLs and then have a deeply grounded conversation with those sources — all answers are cited, hallucinations are minimized, and your source material stays private.

Google Workspace AI brings Gemini-powered intelligence directly into Gmail, Docs, Sheets, Slides, Drive, and Meet — making the world's most widely used cloud productivity suite an AI-first platform for over 3 billion users.

Claude Cowork is Anthropic's desktop AI agent for non-technical knowledge work — running on your computer, navigating local files and applications autonomously, and connecting to enterprise tools like Google Drive, Gmail, and DocuSign.

Notion AI brings writing assistance, summarization, Q&A, and document generation directly inside Notion — the all-in-one workspace used by millions of teams for notes, wikis, project management, and documentation.

Perplexity Personal Computer is an autonomous Mac agent — released to general availability for all Pro and Max subscribers on May 7, 2026 with 400+ connectors, native Mac app control, sandboxed server execution, and remote iPhone trigger, plus a hybrid local-cloud inference orchestrator (Computex 2026) that auto-routes each task between device and cloud.

Gamma is an AI-native presentation and document builder that generates beautiful, structured decks, documents, and webpages from a short text prompt — with no design skills required and results in under a minute.

Grammarly is the world's most widely used AI writing assistant — checking grammar, clarity, tone, style, and now generating full drafts across browsers, Microsoft Office, Google Docs, and mobile, with over 40 million daily active users.

Fireflies.ai is an AI meeting assistant that automatically joins video calls to transcribe, summarize, and extract action items — then makes every meeting searchable and shareable, eliminating the need for manual note-taking entirely.

Jasper AI is an enterprise AI writing platform designed specifically for marketing teams — generating on-brand blog posts, social content, email campaigns, and ad copy with brand voice controls and team workflow features built in from the ground up.

Otter.ai is an AI meeting transcription and note-taking tool that records, transcribes, and summarizes meetings in real time — with a particular strength in personal note-taking, voice capture, and live caption display during calls.

Durable is an AI website builder that generates a complete, professional business website in under 30 seconds from a short business description — then provides hosting, a CRM, invoicing, and basic business tools in a single platform.

Taskade is an AI-native project management and collaboration tool that combines tasks, notes, documents, and AI agents in one workspace — letting teams automate repetitive workflows and manage projects with AI built into every feature.

Google Drive is Google's cloud storage platform that has become deeply integrated with Gemini AI — enabling document summarization, natural language search, and AI-assisted writing directly within Docs, Sheets, and Slides through Google Workspace AI features.

Microsoft OneDrive is Microsoft's cloud storage platform, tightly integrated with Microsoft 365 Copilot AI across Word, Excel, PowerPoint, Outlook, and Teams — making it the AI-powered storage backbone for the world's largest enterprise productivity suite.

Dropbox is a veteran cloud storage and collaboration platform that has invested heavily in AI-native features — including Dash AI universal search across all your files and apps, AI-generated document summaries, and an AI assistant that understands your entire file library.

Amazon S3 (Simple Storage Service) is the world's most widely used object storage platform — the infrastructure backbone of the web that powers data lakes, AI training datasets, application assets, backup systems, and static website hosting for millions of applications globally.

Box is an enterprise-grade cloud content management platform that has built deep AI capabilities through Box AI — offering document summarization, intelligent Q&A, contract analysis, and AI-powered workflows with a strong emphasis on compliance, security, and regulated industry requirements.

Backblaze B2 is a low-cost S3-compatible cloud object storage service offering storage at $6/TB/month — roughly one-quarter the price of Amazon S3 Standard — with no egress fees when paired with Cloudflare, making it a popular choice for media storage, backups, and cost-sensitive developer applications.

ChatGPT Deep Research is OpenAI's agentic research mode that autonomously browses dozens of web sources over several minutes to produce comprehensive, cited research reports — designed for complex questions that require synthesizing information across many sources rather than answering from a single search result.

Gemini Deep Research is Google's agentic research mode within the Gemini AI platform that autonomously conducts multi-step web research over several minutes, leveraging Google Search infrastructure to produce comprehensive cited reports — and uniquely generates an Audio Overview of the report for passive listening.

Tavily is an AI-optimized search API designed specifically for integration into AI agents and RAG pipelines — returning clean, structured, LLM-ready search results that allow AI applications to access real-time web information without the noise and formatting issues of standard search APIs.

Consensus is an AI-powered academic search engine that searches over 200 million scientific papers, uses AI to extract key claims from research, and synthesizes what the scientific literature actually says about a question — giving you the 'scientific consensus' on empirical topics backed by citations.

Elicit is an AI research assistant purpose-built for academic literature that automatically finds relevant papers, extracts structured data from each study, and builds comparison tables — making systematic literature review faster by automating the most tedious parts of academic research workflows.

You.com is an AI-powered search engine and assistant that combines real-time web search with multi-model AI chat — allowing users to choose from multiple LLMs (GPT-5.5, Claude, Gemini, and others) while getting cited answers grounded in current web search results.

Firecrawl is an AI-optimized web scraping API that converts any website into clean, LLM-ready Markdown — handling JavaScript rendering, pagination, authentication, and anti-bot measures so developers can feed web content directly into AI pipelines without building a custom scraper.

Apify is a comprehensive web scraping and automation platform with a marketplace of 2,000+ pre-built scrapers (Actors), cloud execution infrastructure, and a full suite of tools for building, deploying, and scheduling web data extraction pipelines — the enterprise standard for production web scraping.

Apollo.io is a sales intelligence and engagement platform with a database of over 275 million professional contacts — combining web data scraping with AI-powered outreach automation, email sequencing, and CRM enrichment to help sales and marketing teams find and engage prospects at scale.

Browse AI is a no-code web scraping and monitoring platform that allows non-technical users to extract structured data from any website using a visual point-and-click interface — and then schedule automatic monitoring to track changes over time, with no coding required.

SerpAPI is a real-time Google search results API that returns structured, parsed JSON data from Google, Bing, YouTube, Amazon, Google Maps, and 25+ other search engines — allowing developers to integrate live search engine data into applications without HTML scraping or IP blocking.

Diffbot is an AI-powered web data extraction company that uses computer vision and machine learning to automatically understand and extract structured data from any web page without CSS selectors or XPath — and provides a Knowledge Graph of 10 billion entities built from continuously crawling and structuring the entire web.

n8n is an open-source workflow automation platform with a visual node-based editor that connects 400+ apps and services — offering the power of Zapier with the flexibility of code, self-hosting for data privacy, and deep AI agent and LLM integration for building intelligent automation workflows.

Make (formerly Integromat) is a visual workflow automation platform known for its beautiful no-code canvas interface, advanced data manipulation capabilities, and support for complex multi-branch scenarios — positioned between Zapier's simplicity and n8n's developer power.

Zapier is the world's most widely used workflow automation platform — connecting 7,000+ apps with a simple trigger/action model that requires no technical setup, making it the fastest path to automating repetitive tasks for non-technical business users.

LangChain is the most widely used open-source framework for building LLM-powered applications — providing composable building blocks for chains, agents, RAG pipelines, tool use, and memory that let developers build sophisticated AI workflows in Python or JavaScript.

CrewAI is an open-source multi-agent AI framework that lets developers define a 'crew' of specialized AI agents — each with a specific role, backstory, and tools — that collaborate to complete complex tasks, with simpler syntax and better role clarity than AG2.

AG2 (formerly AutoGen) is the community-governed open-source multi-agent framework originally created by Microsoft Research — now independently maintained, with Microsoft having retired AutoGen in favor of its new Microsoft Agent Framework.

Relevance AI is a no-code/low-code platform for building, deploying, and managing AI agents and multi-agent teams — enabling business users to create production AI agents through a visual builder without programming, with tools, memory, and integrations built in.

LlamaIndex is an open-source data framework for building RAG (Retrieval-Augmented Generation) applications — specializing in connecting LLMs to any data source with best-in-class indexing, retrieval, and query pipeline tools, and extending into multi-agent workflows through its LlamaAgents and workflows system.

The OpenAI Agents SDK is OpenAI's official open-source Python framework for building production AI agents — providing handoffs between agents, guardrails, tracing, and a typed function tool system built natively for GPT models with the Model Context Protocol.

ChatGPT Operator is OpenAI's agentic AI that uses a cloud-hosted browser to complete real-world tasks on your behalf — booking reservations, filling forms, making purchases, and navigating the web autonomously, all from a simple natural language instruction.

Claude Computer Use is Anthropic's API capability that lets Claude control a real computer — taking screenshots, moving the cursor, clicking, typing, and executing shell commands — enabling developers to build autonomous agents that operate any desktop or web application; reserve for third-party apps you cannot modify, since structured APIs cost roughly 45 times less for the same task.

Perplexity Assistant is Perplexity's AI-powered mobile assistant that combines real-time web search with agentic task execution — answering questions, completing tasks, and taking actions on your phone using live information rather than a static knowledge cutoff.

Microsoft Copilot Vision is a screen-aware AI capability that lets Copilot see and understand what's on your screen in real time — providing contextual help, explanations, and actions based on what you're looking at in your browser or Windows environment.

OpenAI's browser agent capabilities — spanning ChatGPT's web browsing, the Atlas browser integration, and Operator's web navigation — bring AI-powered browsing, real-time information retrieval, and autonomous web task completion directly into web-based workflows.

Claude for Chrome is Anthropic's browser extension that brings Claude's AI capabilities directly into the Chrome browser — letting you ask questions about any webpage, summarize content, get writing help, and access Claude without switching tabs.

Microsoft Edge Copilot is AI built natively into the Microsoft Edge browser sidebar — providing page summarization, contextual Q&A, writing assistance, and shopping help without any extension to install, powered by GPT-5.5 for Copilot Pro subscribers.

Perplexity Comet is Perplexity AI's dedicated browser built with AI search at its core — replacing the traditional address bar with an AI interface that answers questions, retrieves current information, and navigates the web with real-time search grounding built into every interaction.

Pinecone is the leading managed cloud vector database — purpose-built for AI applications that need fast similarity search at scale, with a serverless architecture, metadata filtering, hybrid search, and seamless integration with every major AI framework.

Supabase Vector brings vector similarity search to PostgreSQL via the pgvector extension — enabling AI applications to store embeddings alongside their regular application data, eliminating the need for a separate vector database service for most use cases.

MongoDB Atlas Vector Search adds native vector similarity search to MongoDB Atlas, enabling teams already using MongoDB to build RAG applications and semantic search without adopting a separate vector database — with HNSW indexing, metadata filtering, and Atlas's global cluster infrastructure.

Chroma is an open-source vector database designed for rapid prototyping and local AI development — with an in-memory or persistent mode, a simple Python-first API, and built-in embedding functions that make it the fastest way to add vector search to a local AI project.

Qdrant is a high-performance open-source vector database written in Rust — offering advanced filtering, quantization for memory efficiency, multi-vector support, and a cloud-hosted or self-hosted deployment model with strong performance benchmarks for production AI applications.

Weaviate is an open-source vector database with a built-in multi-modal data model, schema-based organization, generative search capabilities, and flexible deployment — cloud-hosted or self-hosted — making it particularly strong for applications that combine vector search with structured data retrieval and AI-generated responses.

OpenAI Codex is OpenAI's most powerful coding agent, now defaulting to GPT-5.5. It can build full applications from natural language, run code in sandboxed environments, and handle complex multi-file development tasks autonomously — and is now expanding beyond code into white-collar roles like analytics, sales, and finance via role-specific plug-ins. OpenAI's acquisition of Ona (formerly Gitpod) adds persistent cloud sandboxes that keep agents running on multi-day tasks even after a developer's laptop shuts down.

GPT-OSS is OpenAI's first open-weight model release, available under the Apache 2.0 license. With over 20 billion parameters, it brings OpenAI-quality reasoning to on-premise, edge, and custom deployment scenarios.

Gemini 3.5 Flash is Google DeepMind's new default agentic model, unveiled at Google I/O 2026 — posting 76.2% on Terminal-Bench 2.1, 83.6% on MCP Atlas, 84.2% on CharXiv Reasoning, and 1656 Elo on GDPval-AA. Google claims roughly four-times faster output than other frontier models at less than half the cost. It is also the default model behind Google Search's AI Mode worldwide across 98 languages and the backbone of Google Antigravity 2.0's parallel-subagent stack.

Gemma 4 is Google's open-weight model family under a permissive Apache 2.0 license — E2B and E4B for mobile and edge, the new multimodal 12 billion model for laptops, and a 26 billion mixture-of-experts variant for advanced reasoning. The 12 billion model is Gemma's first with native audio input and runs on a 16-gigabyte laptop.

Phi-4 is Microsoft's open-source small language model, released under the MIT license. It delivers exceptional reasoning and coding performance at a fraction of the size of larger models, making it ideal for on-device and resource-constrained deployments.

Amazon Nova is Amazon's family of foundation models — Premier, Lite, and Micro — designed for deep integration with AWS services. They power enterprise AI applications through Amazon Bedrock with optimized cost and performance tiers.

Amazon Bedrock is AWS's fully managed multi-model AI platform, offering access to foundation models from Amazon, Anthropic, OpenAI (added April 28, 2026), Meta, Mistral, Cohere, and others through a single API — with enterprise security, RAG, and agent capabilities built in.

Mistral Large 3 is Mistral AI's flagship open-weight model — a 675 billion MoE architecture with 41 billion active parameters per forward pass, 256K context, and multimodal capabilities. The April 2026 lineup also includes Mistral Medium 3.5 — a 128 billion-parameter dense model that runs on as few as 4 GPUs at $1.50/$7.50 per million tokens, posting 77.6% on SWE-Bench Verified. Mistral expanded into industrial engineering in May 2026 by acquiring Austrian physics-AI company Emmi AI, layering real-time simulations and digital twins onto the Mistral platform for aerospace, automotive, semiconductor, and energy customers.

Devstral 2 is Mistral AI's dedicated software engineering model, achieving 72.2% on SWE-bench Verified with up to 7x greater cost efficiency than comparable models. Available in 123 billion and 24 billion sizes with open weights.

DeepSeek R1 is the first open-source reasoning model to match OpenAI's o1, released under the MIT license. It introduced chain-of-thought reasoning to the open-source ecosystem, with distilled variants as small as 1.5 billion parameters.

GLM-5 is a 744 billion open-source MoE model from Zhipu AI, built entirely on Huawei Ascend chips with zero NVIDIA dependency. The first model from a publicly listed Chinese AI company, it claims to surpass Gemini 3 Pro on coding and agentic benchmarks.

QwQ-32 billion is Alibaba's reasoning-specialized open-source model under the Apache 2.0 license. Part of the Qwen3.5 family, it brings chain-of-thought reasoning capabilities to a size that runs on a single high-end GPU.

Command A is Cohere's flagship enterprise LLM — 111 billion parameters with 256K context, running on just 2 GPUs. It powers the North agentic AI platform behind customer firewalls for enterprise RAG and agentic workflows.

Aya Expanse is Cohere's open-source multilingual model family, spanning from the original 23-language text model to Aya Vision (multimodal) and Tiny Aya (70+ languages on edge devices). It advances language equity in AI.

Aurora is xAI's native image generation model, integrated directly into the Grok chatbot and the X (Twitter) platform. It generates images through conversational prompts within the Grok interface.

Nano Banana Pro is Google's state-of-the-art image generation model, producing 4K resolution images with accurate text rendering and SynthID watermarking. It represents Google's most capable image generation offering.

Claude Code is Anthropic's autonomous coding agent CLI. It runs in your terminal, reads your entire codebase, writes and edits files across your project, runs tests, manages git workflows, and connects to external tools via MCP — all powered by Claude Opus and Sonnet.

GitHub Copilot is an AI pair programmer integrated into VS Code, JetBrains, and other IDEs that suggests code completions, entire functions, and documentation as developers type. At Build 2026 it added a dedicated desktop app for orchestrating multiple coding agents in parallel.

ERNIE 5.0 is Baidu's 2.4 trillion parameter unified multimodal foundation model — integrating text, image, video, and audio in a single framework, comparable to Gemini-2.5-Pro and GPT-5-High on 40+ benchmarks.

Kimi K2.6 is Moonshot AI's current flagship — a 1 trillion parameter mixture-of-experts model with a 262K context window, native INT4 quantization, and an Agent Swarm system that scales to 300 sub-agents and 4,000 coordinated steps in 12-hour autonomous coding sessions. Open-weights released April 20, 2026 on Hugging Face; commercial launch alongside Moonshot's $2 billion funding round at a $20 billion valuation on May 7, 2026. Ranks as the second-most-used model on OpenRouter, and now powers Kimi Work, a local desktop agent for macOS and Windows that runs the 300-sub-agent swarm on your own machine.

Kimi K2.5 (January 2026) is Moonshot AI's 1 trillion parameter MoE model — natively multimodal with 256K context, beating GPT 5.2 on SWE-Bench Multilingual and Gemini 3 Pro on SWE-Bench Verified. Superseded as flagship by Kimi K2.6 (April-May 2026); kept here as the previous-generation reference. For the current Moonshot flagship, see the Kimi K2.6 page.

Falcon 3 is the Technology Innovation Institute's open-source model family from Abu Dhabi — leading Hugging Face leaderboards for models under 13 billion parameters, with specialized variants for Arabic, multimodal, and ultra-low-power edge deployment.

Doubao is ByteDance's AI chatbot — the most-used in China with over 100 million daily active users. Doubao 2.0 (Seed 2.0) matches GPT 5.2 and Gemini 3 Pro at roughly one-tenth the cost.

MiniMax is a Chinese AI company that IPO'd in Hong Kong in January 2026, raising $620 million with its stock surging 109% on debut. Long known for consumer AI products with over 100 million users, MiniMax has now turned to the open-weight frontier with M3 — a model combining frontier coding, agentic execution, native multimodality, and a 1-million-token context window — while M2.7 remains its prior-generation foundation model.

Salesforce Agentforce is an autonomous AI agent platform with 12,000+ customers live and $500 million+ annualized revenue — deploying AI agents for sales, service, and marketing directly within the world's largest CRM.

ServiceNow Now Assist is an AI assistant integrated across IT, HR, customer service, and security workflows — enhanced by the $2.85 billion Moveworks acquisition and used by 85% of the Fortune 500.

Cohere North is an enterprise agentic AI platform deployed entirely behind customer firewalls — built on Command A and designed for organizations that cannot let data leave their infrastructure.

Microsoft Dragon Copilot (formerly Nuance DAX) is a clinical AI assistant that listens to patient-physician encounters and auto-generates documentation — deployed at 600+ health systems.

Amazon Nova Forge is a platform for building custom frontier models from Nova checkpoints using proprietary enterprise data — a unique offering among cloud providers for organizations that need specialized AI.

Google Gemini 3.1 Pro is Google DeepMind's most advanced reasoning model — scoring 94.3% on GPQA Diamond (highest ever) with 1 million token context, native multimodal input, and Deep Think mode.

Palantir Maven Smart System is an AI-powered battlefield intelligence platform integrating computer vision, multi-source intelligence fusion, and targeting support — adopted by NATO and the US Army.

Cohere Tiny Aya is an open-weight 3.35 billion-parameter multilingual model family supporting 70+ languages — designed for laptops, edge devices, and resource-constrained deployment with regional variants for Africa, South Asia, and Asia-Pacific.

HubSpot Breeze AI is a platform of 20+ AI agents embedded across CRM, marketing, and sales — including Customer Agent, Prospecting Agent, and Breeze Studio for custom agent building, serving 288,000+ paying customers.

Gong Revenue AI records and analyzes every sales conversation — using AI Call Reviewer, deal intelligence, and Mission Andromeda's open MCP connections to transform revenue operations. $300 million+ ARR, 5,000+ customers.

Zendesk AI Agents resolve 80% of customer support issues autonomously with outcome-based pricing — replacing the former Answer Bot and strengthened by the $200 million+ Forethought acquisition.

Intercom Fin 2 is a customer support AI agent that resolves up to 82% of support volume with 99.9% accuracy — featuring omnichannel support, Procedures for multi-step workflows, and $0.99 per resolution pricing.

Workday Illuminate is an AI platform with 7+ autonomous agents for HR and finance — enhanced by the $1.1 billion Sana acquisition, Workday Build for custom solutions, and Flex Credits for transparent AI consumption.

Figma Make is an AI prompt-to-code tool that turns text descriptions or designs into working React/Tailwind prototypes — part of Figma's expanded product suite after its $19.3 billion IPO.

LangGraph is LangChain's production-grade agent orchestration framework that models agent logic as stateful directed graphs — with durable execution, human-in-the-loop checkpoints, and persistent memory. v1.0 reached early 2026.

Gemini CLI is Google's open-source AI coding agent that runs in your terminal, automatically routing between Gemini 3.5 Pro and Flash models with a 1 million token context window, plan mode, subagent support, and a free daily usage limit. Important — Gemini CLI sunsets June 18, 2026 in favor of Antigravity CLI, the Go-based rewrite that ships alongside Google Antigravity 2.0.

AWS Kiro is Amazon's AI-native IDE built on Code OSS (VS Code base), featuring spec-driven development with EARS notation, agent hooks that trigger on file events, native MCP support, and GovCloud availability for regulated environments.

Mistral Vibe CLI is the open-source terminal entry point to the unified Mistral Vibe agent platform — powered by Devstral 2 with custom subagents via TOML configuration, slash-command skills, MCP support, and on-premise deployment. Sits alongside the Vibe web UI, mobile apps, and the new VS Code extension as one of several Vibe entry points; remote coding agents run on the Mistral Medium 3.5 backend.

Cursor is an AI-native code editor built on VS Code that combines inline completions, multi-file editing, and autonomous agent mode into a single IDE — the fastest-growing AI coding tool with over 360K paying customers. In June 2026, SpaceX agreed to acquire Cursor outright for $60 billion in stock.

Windsurf is an AI-native IDE featuring the Cascade agent that deeply understands your codebase and proactively suggests multi-step changes — ranked number one in developer power rankings.

Bolt.new by StackBlitz is a browser-based AI development environment that generates, runs, and deploys full-stack web applications from a natural language description — the fastest path from idea to deployed app.

v0 is Vercel's AI-powered development environment that generates full-stack applications from natural language, with GitHub integration, git workflows, and one-click Vercel deployment.

Replit is a cloud-based development platform with Agent 4 — an AI that builds full applications on an infinite interactive canvas with parallel execution, real-time collaboration, and built-in hosting.

Google Antigravity is an agent-first IDE from Google DeepMind that lets developers dispatch and manage multiple AI coding agents simultaneously through a Manager View — with support for multi-model selection. Antigravity 2.0 (May 2026) adds parallel subagents that compress multi-week workflows into hours, a unified backend across desktop and a new Go-based Antigravity CLI, and default Gemini 3.5 Flash + Gemini 3.5 Pro routing — also absorbing the Gemini CLI installed base ahead of its June 18, 2026 sunset.

Aider is an open-source, git-integrated CLI coding assistant that works with any LLM — every AI change is auto-committed to git with an explanatory message, giving you a clean history and easy rollback.

Goose is an open-source, MCP-native coding agent donated to the Linux Foundation — model-agnostic, community-governed, and backed by AWS, Anthropic, Google, Microsoft, and OpenAI.

Continue.dev is an open-source Continuous AI platform that runs async agents on every pull request, integrates with GitHub, Sentry, and CI/CD, and offers both headless cloud and interactive TUI modes.

Ollama is the most popular local model runner, letting you download and run open-source LLMs with a single command on Mac, Linux, and Windows — with a REST API, web search, and support for hundreds of models.

Vercel is the leading frontend hosting platform, offering zero-config deployment, preview deployments for every PR, Edge Functions, a global CDN, and v0 integration for design-to-production workflows.

Netlify is a web platform for modern applications, offering an AI Gateway for LLM API routing, edge-rendered prerendering, split testing, and built-in identity and form handling.

Neon is a serverless PostgreSQL platform with scale-to-zero economics, git-like database branching, and modern developer tooling including Neon Auth and a Rust-powered Data API.

Stripe is the dominant payment platform for internet businesses, now expanding into agentic commerce, multi-processor orchestration, stablecoin accounts, AI-powered financial tools, and a Stripe Projects beta (May 2026) that orchestrates AI agents buying their own cloud infrastructure.

Firebase is Google's application development platform, now featuring Firebase Studio (an agentic dev environment), AI Logic with Gemini 3 models, and a comprehensive suite of backend services.

Tesla Full Self-Driving (FSD) is a vision-only autonomous driving system deployed across ~2 million vehicles, and Cybercab is Tesla's purpose-built robotaxi entering production in April 2026. Tesla's AI5 chip — taped out April 15, 2026 and dual-sourced at Samsung Taylor + TSMC Arizona — will power FSD v15 and transition long-term to the Terafab foundry.

Tesla Optimus is a general-purpose humanoid robot now in Gen 3 with over 1,000 units deployed in Tesla factories, targeting mass production at $20,000 per unit and external customer deliveries by late 2026. Next-gen Optimus runs on Tesla's AI5 brain (taped out April 2026) — the same silicon path as FSD — and will transition long-term to the Terafab foundry.

Waymo One is the world's most commercially advanced autonomous robotaxi service, operating fully driverless rides in 10+ US cities with 170 million+ autonomous miles driven and 91% fewer serious-injury crashes than human drivers.

Zoox is Amazon's purpose-built autonomous robotaxi with a bidirectional design, operating in San Francisco and Las Vegas and expanding to Austin and Miami in 2026.

Aurora Driver is the first commercial driverless trucking platform operating at scale in the US, with 10 active routes, 250,000+ driverless miles, and partnerships with FedEx, Werner, and Uber Freight.

Neuralink N1 is a brain-computer interface implant with ~20 patients as of early 2026, enabling thought-controlled cursor movement, internet browsing, and communication via 1,024 electrode threads.

Figure 03 is Figure AI's third-generation humanoid robot, targeting home use at $20,000 after Figure 02 completed an 11-month BMW manufacturing deployment producing 30,000+ vehicles.

Boston Dynamics Atlas is an all-electric humanoid robot with 56 degrees of freedom, now commercially deployed in warehouse and logistics operations and shipping to partners including Hyundai and Google DeepMind.

Synchron's Stentrode is an endovascular brain-computer interface threaded through blood vessels — requiring no open brain surgery — that enables patients with paralysis to control digital devices through thought.

Precision Neuroscience's Layer 7 Cortical Interface is a thin-film brain-computer interface with 1,024 electrodes that conforms to the brain surface without penetrating tissue, with FDA 510(k) clearance and a Medtronic partnership.

OpenClaw is a free, open-source autonomous AI agent with 247,000+ GitHub stars that runs locally and uses messaging platforms as its interface — supporting 100+ built-in skills, persistent memory, proactive scheduling, and an extensible skill marketplace.

Nemotron is NVIDIA's family of open-weight large language models — now led by the Nemotron 3 generation, whose 550 billion parameter Ultra flagship launched at Computex as the top-ranked US open-weights model, alongside the efficient Nano and Super sizes, the Elastic 30 billion checkpoint that packs three nested model sizes into one, and the Diffusion 14 billion tri-mode model that switches between autoregressive, diffusion, and self-speculation decoding — designed for enterprise customization, synthetic data generation, and production deployment on NVIDIA hardware.

NVIDIA NIM (Neural Inference Microservices) packages optimized AI models into production-ready Docker containers with a single API call — dramatically simplifying model deployment while delivering peak inference performance on NVIDIA GPUs.

NVIDIA NeMo is an open-source framework for building, customizing, and deploying generative AI models at scale — including tools for data curation, fine-tuning, safety guardrails, and evaluation that cover the full LLM lifecycle.

NVIDIA Jetson is the dominant edge AI computing platform — from the $499 Orin Nano developer kit to the 275 TOPS AGX Orin module — powering robotics, autonomous machines, industrial inspection, and on-device AI where cloud connectivity is impractical.

NVIDIA Isaac is the leading robotics AI platform (simulation, ROS packages, and manipulation tools), while Omniverse provides the 3D simulation and digital twin infrastructure that powers it — together enabling teams to train, test, and deploy robot AI in photorealistic virtual environments before touching real hardware.

NVIDIA DRIVE is the end-to-end autonomous vehicle computing platform — from the DRIVE Orin SoC already in production vehicles to the next-generation DRIVE Thor — providing the AI compute brain, simulation tools, and software stack used by major automakers worldwide.

CUDA is NVIDIA's parallel computing platform — the foundational software layer that nearly all AI frameworks, models, and tools are built on. Its 20-year ecosystem of optimized libraries, developer tools, and community knowledge is NVIDIA's deepest competitive moat.

TensorRT-LLM is NVIDIA's open-source library for optimizing large language model inference — delivering best-in-class throughput through automatic quantization, in-flight batching, KV cache optimization, and multi-GPU parallelism on NVIDIA GPUs.

GPT-5.5 is OpenAI's flagship model (April 23, 2026), built for agentic work; on May 5, 2026 GPT-5.5 Instant became ChatGPT's default model across every tier with reduced hallucinations on law / medicine / finance and AIME 2025 math 81.2 vs the prior 65.4.

Claude Opus 4.6 is Anthropic's flagship model — the most capable Claude for complex reasoning, agentic coding, and long-document analysis, with a 1 million token context window, leading OSWorld benchmark scores (72.7%), and the highest retrieval accuracy among frontier models.

The Claude Agent SDK is Anthropic's official framework for building custom AI agents in Python and TypeScript — providing built-in file operations, shell commands, web search, and MCP integration out of the box, with the same architecture that powers Claude Code.

Grok is xAI's model family — now part of SpaceX ($1.25 trillion combined) — featuring the largest context window of any major model (2 million tokens), real-time X/Twitter data access, and the Grok 4.20 multi-agent system. xAI is a founding partner in Terafab, the $20-25 billion chip foundry joint venture that positions Grok's long-term non-NVIDIA training hardware supply.

The Microsoft Agent Framework is Microsoft's enterprise platform for building, deploying, and governing AI agents — replacing AutoGen and Semantic Kernel as the company's unified approach to agentic AI across Azure, Microsoft 365, and Dynamics 365.

Llama 4 is Meta's open-weight foundation model family — featuring the Mixture-of-Experts architecture, with Maverick (400 billion total / 17 billion active parameters, 1 million context) and Scout (10 million token context) variants that are the most downloaded open-weight frontier models in the world.

PyTorch is the dominant open-source machine learning framework — originally created by Meta AI Research, now governed by the Linux Foundation — used by the majority of AI researchers and increasingly in production, powering everything from LLM training to computer vision to robotics.

SAM 2 (Segment Anything Model 2) is Meta's open-source foundation model for image and video segmentation — capable of identifying and segmenting any object in any image or video with a single click, point, or text prompt, without requiring task-specific training.

Amazon Q Developer is AWS's AI coding assistant — rebranded from CodeWhisperer — offering code suggestions, chat-based debugging, code transformation (automated Java upgrades), and AWS infrastructure troubleshooting, with a generous free tier for individual developers.

Amazon SageMaker is AWS's fully managed machine learning platform — covering the entire ML lifecycle from data preparation and model training to deployment and monitoring, used by thousands of enterprises to build and operate production ML systems at scale.

AWS PartyRock is Amazon's free, no-code AI app builder — powered by Amazon Bedrock — that lets anyone create AI-powered applications (chatbots, text generators, image creators) through a visual interface without writing code or needing an AWS account.

Cerebras Inference is an AI inference platform powered by the world's largest chip — the Wafer-Scale Engine 3 — delivering the fastest token generation speeds for large language models, with partnerships from OpenAI, AWS, and Meta. Cerebras completed the largest US tech IPO of 2026 on May 14, pricing at $185 and closing +108% at $311 for a $66 billion market cap.

Groq Cloud (GroqCloud) is an ultra-fast AI inference platform powered by custom Language Processing Unit (LPU) chips — purpose-built for real-time AI applications. Following NVIDIA's $20 billion December 2025 deal that absorbed founder Jonathan Ross and senior chip leadership, Groq has refocused on its inference neocloud — the on-demand cloud platform sitting on top of the existing LPU fleet — and is raising roughly $650 million to fund the pivot under interim CEO Adam Winter and interim CFO Matt Eng.

Together AI is an AI cloud platform for running, fine-tuning, and training 200+ open-source models — offering competitive inference speeds, GPU rentals, and pioneering research like FlashAttention and RedPajama.

Scale AI is the leading AI data infrastructure company — providing data annotation, model evaluation (SEAL benchmarks), and enterprise AI deployment services for frontier labs, enterprises, and the US Department of Defense.

Databricks is a unified data intelligence platform combining a data lakehouse with AI and machine learning — used by 60% of Fortune 500 companies for data engineering, model training, fine-tuning, and AI-powered analytics.

IBM watsonx is an enterprise AI platform spanning model training (watsonx.ai), data management (watsonx.data), AI governance (watsonx.governance), and agentic automation (watsonx Orchestrate) — built for regulated industries with hybrid cloud deployment.

Snowflake Cortex AI is a suite of AI and machine learning features built directly into the Snowflake data cloud — enabling SQL-first AI on governed enterprise data without moving data out of the platform.

Baseten is a high-performance AI model inference platform — backed by NVIDIA — that lets teams deploy, serve, and scale custom and open-source AI models in production with auto-scaling GPU infrastructure.

Dell AI Factory is an end-to-end enterprise AI infrastructure platform from the world's number one server vendor — combining PowerEdge AI servers, storage, networking, and as-a-service options to help organizations build and run AI on-premises.

Datadog LLM Observability monitors AI application performance, cost, and quality — tracking LLM calls, token usage, latency, and error rates alongside full-stack infrastructure metrics in a single platform.

The Huawei Ascend 950PR is China's most powerful domestically produced AI accelerator — featuring 1.56 petaFLOPS of FP4 compute and in-house HBM memory — built under US sanctions to reduce China's dependence on NVIDIA.

Samsung Mach-1 is Samsung's first proprietary AI inference accelerator — targeting edge computing and lightweight data center inference at approximately one-tenth the cost of NVIDIA hardware, using LPDDR instead of HBM memory.

IonQ builds the world's most powerful trapped-ion quantum computers — from the 36-qubit Forte to the 100-qubit Tempo — accessible via AWS, Azure, and Google Cloud for drug discovery, optimization, and quantum machine learning.

Lightmatter Passage is a photonic interconnect platform that uses light instead of electricity to connect AI processors — dramatically reducing power consumption and enabling thousands of GPUs to communicate in a single domain.

Hugging Face Hub is the world's largest open-source AI platform — hosting over 2 million models, 500,000 datasets, and 300,000 demo applications — serving as the central infrastructure for the open-source AI ecosystem.

Jamba is AI21 Labs' hybrid SSM-Transformer model family — combining Mamba's memory efficiency with Transformer quality for enterprise-grade long-context AI with 256,000 token context windows and Apache 2.0 licensing.

CoreWeave is the leading independent GPU cloud provider — purpose-built for AI workloads with 250,000+ NVIDIA GPUs, Kubernetes-native infrastructure, and the distinction of being first to deploy every new NVIDIA GPU generation.

Samsung Galaxy AI is a suite of on-device and cloud AI features built into Galaxy smartphones, tablets, and foldables — reaching 400 million+ devices with live translation, photo editing, writing assistance, and agentic automation.

Oracle Cloud Infrastructure (OCI) AI provides GPU superclusters, generative AI services, and enterprise AI capabilities — positioned as the price-performance leader for AI workloads at 40-70% less than AWS and Azure.

Sarvam-105B is India's first sovereign AI model — a 105 billion parameter open-source model supporting all 22 official Indian languages, trained on government-subsidized GPUs under the IndiaAI Mission.

Solar Pro 2 is Upstage's 31 billion parameter model that scored 58 on the Artificial Analysis Intelligence Index — beating GPT-4.1 — positioning South Korea as a serious player in the global AI model race.

NTT tsuzumi 2 is a 30 billion parameter Japanese-optimized language model that runs on a single H100 GPU — selected for Japan's government Gennai AI platform serving 180,000 government staff.

Devin is the first fully autonomous AI software engineer — planning, writing, debugging, and deploying code end-to-end from natural language. Cognition Labs is now valued at $26 billion post-money on $492 million in annualized revenue, with enterprise customers including Mercedes-Benz, NASA, Goldman Sachs, and Santander. CEO Scott Wu disclosed in late May 2026 that Devin commits 89 percent of Cognition's own engineering output — and explicitly frames the product as augmentation, not replacement.

Sierra Agent is an enterprise AI customer service platform by Bret Taylor (ex-Salesforce CEO) — handling customer interactions across voice, chat, and email with ~70% auto-resolution. Sierra closed a $950 million round at over $15 billion valuation on May 4, 2026, with $150 million in ARR and 40 percent of the Fortune 50 as customers.

pi-0 is Physical Intelligence's foundation model for general-purpose robotics — a single AI model that controls 7+ different robot types across 68+ manipulation tasks, from folding laundry to assembling boxes in a real factory.

Augment Code is an enterprise AI coding assistant backed by Eric Schmidt — featuring a Context Engine that understands entire codebases, an AI agent (Auggie) with a 70.6% SWE-bench score, and SOC 2 + ISO 42001 compliance.

Kimi Code is Moonshot AI's open-source coding agent — now powered by the Kimi K2.7 Code model (June 2026), a 1 trillion parameter mixture-of-experts model with a 256K context window. Moonshot reports double-digit benchmark gains over K2.6 and roughly 30 percent fewer reasoning tokens, shipped under a Modified MIT license.

Magic is building AI coding models with 100 million+ token context windows — capable of understanding entire codebases at once — backed by $466 million from Eric Schmidt, Atlassian, and Sequoia, though the product has not yet publicly launched.

Poolside is building code foundation models from scratch using reinforcement learning from code execution — trained across 130,000+ real codebases, backed by up to $1 billion from NVIDIA, and valued at up to $12 billion.

Character.AI is the largest AI character platform — with 20 million monthly active users creating and chatting with 18 million+ AI characters for roleplay, creative writing, and entertainment; faces a May 2026 Pennsylvania lawsuit over a chatbot that impersonated a licensed psychiatrist with a fabricated medical license number.

SAP Joule is an AI copilot embedded across SAP's enterprise suite — ERP, HR, finance, and supply chain — with 2,100+ AI skills, 14+ agents, and Joule Studio for building custom enterprise AI workflows.

Writer is an enterprise AI content platform with its own Palmyra LLM family — delivering near GPT-4.1 performance at 75% lower cost with brand governance, compliance controls, and domain-specific models for finance and healthcare.

Airtable AI is a no-code platform that lets anyone build AI-powered databases, workflows, and applications through natural language — with Omni conversational builder and Superagent multi-agent research, used by 80% of Fortune 100 companies.

Lovable is an AI web app builder that generates complete full-stack applications from natural language — with nearly 8 million users, $400 million ARR, and a $6.6 billion valuation, making it the leader in 'vibe coding.'

Mercor is an AI-powered talent matching and contractor marketplace — connecting 30,000+ domain experts with AI companies for model training, reaching $500 million ARR and a $10 billion valuation in under two years.

Copy.ai is an AI-powered Go-to-Market platform with Content Agent Studio, prospecting agents, and sales workflow automation — now part of Fullcast's RevOps ecosystem after its October 2025 acquisition.

Klarna AI Assistant is the most cited example of enterprise AI customer service — handling two-thirds of all customer chats, saving $60 million, but also a cautionary tale after aggressive AI job cuts led to customer satisfaction declines and a rehiring pivot.

Shopify Magic is a suite of free AI tools built into every Shopify plan — including product description generation, AI image editing, SEO optimization, and the Sidekick conversational business assistant for 1.75 million+ merchants.

Glean is an enterprise AI search platform that connects 100+ business apps — Slack, Google Workspace, Salesforce, Jira, and more — delivering permissions-aware AI answers grounded in your company's knowledge. $300 million annualized revenue (tripled in fifteen months), $7.2 billion valuation, and a cost-cutting context-graph pitch as its primary differentiator.

Hebbia Matrix is an AI-powered document analysis platform for finance and legal — using multi-agent swarms to process hundreds of documents in parallel, adopted by 33% of top global asset managers.

Stable Diffusion 3.5 is Stability AI's latest open-weight image generation model — available in Large (8.1 billion parameters), Large Turbo, and Medium (2.5 billion) variants that run on consumer GPUs.

Dream Machine is Luma AI's text-to-video and image-to-video generation platform — powered by the Ray3 model with native 1080p output, character consistency, and studio-grade HDR, backed by $1 billion in funding at a $4 billion valuation.

Elastic AI Search (ESRE) combines traditional keyword search with vector semantic search in a single platform — enabling hybrid retrieval-augmented generation (RAG) for enterprise AI applications, backed by 20+ years of Elasticsearch infrastructure.

CrowdStrike Falcon is the world's leading AI-native cybersecurity platform — protecting 29,000+ organizations with endpoint detection, Charlotte AI agentic analyst, and AgentWorks for building custom security agents.

Darktrace is a self-learning AI security platform that models normal behavior across enterprise environments and autonomously neutralizes cyber threats — acquired by Thoma Bravo for $5.3 billion in October 2024.

SentinelOne is an AI-powered endpoint protection platform with Purple AI — a generative AI security analyst offering autonomous investigation, attack timeline visualization, and agentic detection and response at $1 billion+ ARR.

Cortex XSIAM is Palo Alto Networks' AI-native security operations platform — replacing traditional SOCs with agentic AI, autonomous investigation, and 98% reduction in mean time to respond, with 470+ customers each spending over $1 million ARR.

Wiz is a cloud security platform using agentless scanning and a Security Graph to visualize attack paths across multicloud environments — acquired by Google for $32 billion in March 2026, the largest cybersecurity acquisition ever.

Tempus AI is a clinical AI and precision medicine company with the largest multimodal clinical dataset in the industry — 40+ million records powering genomic testing, AI oncology tools, and data licensing at $1.27 billion revenue.

PathAI is a digital pathology AI company with FDA-cleared tools for primary diagnosis (AISight Dx), clinical trials (AIM-MASH), and dermatopathology — deployed at Labcorp and Quest Diagnostics nationwide.

Aidoc is an AI radiology triage platform with healthcare's first FDA-cleared foundation model — detecting 14 critical conditions from a single scan across 150+ health systems serving 45 million patients annually.

Recursion Pharmaceuticals is an AI drug discovery company using biology-scale datasets and machine learning to compress drug development from 12 years to 4-5 — with 5 clinical programs and partnerships with Sanofi, Roche, Bayer, and Merck.

Viz.ai is a medical AI care coordination platform with 50+ FDA-cleared algorithms — detecting strokes, pulmonary embolisms, and other time-sensitive conditions across 1,000+ hospitals, including the first AI to receive CMS reimbursement.

Harvey is the fastest-growing AI legal platform — used by 100,000+ lawyers across 50+ of the top 100 AmLaw firms for research, contract analysis, due diligence, and custom agentic workflows, valued at $11 billion. May 2026 brought Anthropic into the same market with Claude for Legal — the first frontier lab to ship a named vertical legal product line.

CoCounsel is Thomson Reuters' AI legal assistant with 1 million users across 107 countries, while Westlaw Advantage is the AI-powered evolution of the definitive legal research platform — featuring agentic Deep Research.

Ironclad is an AI-powered contract lifecycle management platform with Jurist agentic AI — featuring 6 specialized agents for intake, redlining, drafting, review, research, and search across 2 billion+ contracts processed.

Relativity is the dominant eDiscovery platform with the aiR suite — AI-powered document review, privilege detection, and case strategy included at no extra cost in RelativityOne, processing 3 million documents per day.

Palantir builds AI platforms for enterprise operations (AIP) and government intelligence (Gotham) — with $4.48 billion revenue, a $325+ billion market cap, and major defense contracts including the $10 billion Army enterprise agreement and Golden Dome missile shield.

AlphaSense is an AI market intelligence platform used by 88% of the S&P 100 — analyzing earnings calls, SEC filings, and expert transcripts for investment banks, hedge funds, and corporate strategy teams at $500 million+ ARR.

Intuit Assist is an agentic AI platform embedded across TurboTax, QuickBooks, Credit Karma, and Mailchimp — with specialized AI agents for tax filing, bookkeeping, and financial guidance serving 100 million consumers and businesses.

Lemonade is an AI-first insurance company where chatbot Maya binds policies in 90 seconds and claims bot Jim processes 27% of claims autonomously — serving 3 million+ policyholders across renters, homeowners, pet, car, and life insurance.

Tractable uses computer vision AI to assess vehicle and property damage from photos — reducing insurance appraisal times from weeks to minutes, processing $7 billion+ in annualized repairs for GEICO, Tokio Marine, and 20+ top insurers.

Coinbase Agentic Wallets are the first wallet infrastructure designed for AI agents — enabling autonomous spending, earning, trading, and transacting on-chain with programmable security guardrails, launched February 2026.

Khanmigo is Khan Academy's AI tutor that guides students to understanding through Socratic questioning — rather than giving answers — serving 2 million users including 700,000 K-12 students across 380+ school districts.

Duolingo is the world's most popular language learning app with 133 million monthly active users — using its proprietary Birdbrain AI system plus GPT-4 for personalized lessons, roleplay conversations, and AI video calls.

Gradescope is Turnitin's AI-assisted grading platform for higher education — using AI answer grouping and handwritten submission support to help professors grade faster and more consistently at scale.

Synthesis is an AI math tutoring and collaborative problem-solving platform for kids ages 8-14 — originally developed for SpaceX's school, now serving 21,500+ students with a vision to replace traditional K-12 education.

Lattice is Anduril Industries' AI autonomy and command-and-control platform for defense — selected for a $20 billion Army contract to integrate sensors, unmanned systems, and autonomous operations across all military domains.

ServiceNow Now Assist is an enterprise AI suite automating IT service management, HR operations, and business workflows — reducing resolution times by up to 4.5 hours and saving agents 6 hours per week, embedded in the $13+ billion ServiceNow platform.

UiPath is the leading robotic process automation platform — now pivoting to agentic AI that unifies AI agents, robots, and human workers for enterprise automation, with $1.69 billion ARR and 950+ companies building AI agents.

AspenTech is an industrial AI software company optimizing operations for energy, chemical, and manufacturing companies — now a wholly owned subsidiary of Emerson after a $17 billion acquisition, with ~$1.1 billion annual revenue.

Uplight is an AI-powered energy intelligence platform helping utilities manage demand response, customer engagement, and grid flexibility — valued at $1.5 billion as data center AI demand transforms the energy landscape.

Duetto is an AI revenue management platform for hotels — using Open Pricing to dynamically optimize room rates per room type, segment, and channel across 6,000+ hotel properties worldwide.

Prospera is a computer vision AI platform for real-time crop health monitoring — acquired by Valmont Industries for $300 million and now integrated into the Accent 365 precision agriculture platform.

Taranis is an AI-powered aerial crop intelligence platform using drone and satellite imagery with submillimeter resolution — trained on 50+ million images and partnered with Syngenta for Midwest distribution.

Climate FieldView is Bayer's digital farming platform using AI to optimize planting, fertilization, and harvest across 250+ million subscribed acres in 23 countries — from $299/year.

Trimble Agriculture provides GPS guidance, auto-steer, and AI analytics for precision farming — now largely operating as PTx Trimble (85% AGCO / 15% Trimble joint venture) serving the $17 billion precision agriculture market.

Augury is an AI-powered machine health platform using vibration sensors and IoT to predict equipment failures before they cause downtime — a $1 billion+ unicorn with 99.9% failure detection accuracy.

Sight Machine is a manufacturing analytics platform that creates digital twins of production processes — using AI for anomaly detection, root cause analysis, and process optimization, named to Fast Company's Most Innovative Companies 2026.

Tulip is a no-code factory operations platform letting frontline workers build custom manufacturing apps — a $1.3 billion unicorn with 43,000 apps built across 1,000 customer sites in 45 countries.

CoStar is the dominant commercial real estate data and analytics platform — now expanding with Homes AI conversational search — with $3.2 billion revenue and an $18 billion market cap.

HouseCanary provides the most accurate residential real estate AI valuations — with 2.8% median error across 136+ million properties, used by lenders, investors, and iBuyers for automated property assessment.

Procore is the leading construction project management platform — adding AI intelligence (Helix), custom agent building, and conversational search to its $1.3 billion revenue platform serving the global construction industry.

Dynamic Yield is Mastercard's AI personalization platform for e-commerce and retail — delivering individualized recommendations, content, and pricing across web, mobile, email, and in-store displays, with 7 consecutive years as Gartner Leader.

Nosto is an AI e-commerce personalization platform delivering product recommendations, personalized search, and dynamic merchandising to 3,500+ brands — with 323% growth in AI search queries and performance-based pricing.

project44 is an AI-powered supply chain visibility platform tracking 1.5 billion shipments annually — using multi-agent orchestration for predictive ETAs, delay detection, and automated decision-making across global logistics.

Samsara is a fleet intelligence platform using AI dashcams, GPS tracking, and predictive analytics — with $1.46 billion ARR and data showing AI-enabled fleets reduce crash rates by nearly 75% over 30 months.

Paperclip is an open-source orchestration platform that manages teams of AI agents like employees in a company — with org charts, budgets, governance, and audit trails for fully autonomous business operations.

autoresearch is Andrej Karpathy's open-source AI agent that autonomously runs ML experiments on a single GPU — modifying training code, keeping improvements, and discarding failures at a rate of approximately 100 experiments overnight.

Claude Mythos 5 is Anthropic's restricted frontier model — the unsafeguarded version of the same Mythos-class model the public reaches as Claude Fable 5. Launched June 9, 2026 alongside Fable 5, it is delivered through Project Glasswing for defensive cybersecurity and biomedical research, the program that has surfaced over 10,000 critical vulnerabilities across roughly 50 partner organizations, including 271 in Firefox 150 as the first peer-validated deployment.

Muse Spark is Meta's flagship AI model — the first product from Meta Superintelligence Labs, accepting voice, text, and image inputs. Code-named Avocado and built over 9 months, it powers Meta AI across WhatsApp, Instagram, Facebook, and Ray-Ban glasses.

Claude Managed Agents is Anthropic's public beta offering composable APIs for building and deploying cloud-hosted AI agents at scale — handling orchestration, scaling, and monitoring through Anthropic's infrastructure.

Voxtral TTS is Mistral AI's open-source text-to-speech model — 4 billion parameters that run on consumer hardware, supporting 9 languages with natural prosody, available under an open license and via API at $0.016 per 1,000 characters.

Gemini Computer Use is Google DeepMind's agentic capability that allows Gemini 3 Pro and Flash to interact with graphical user interfaces — taking screenshots, clicking, typing, and navigating applications autonomously.

Codex Security is OpenAI's dedicated security agent — an extension of the Codex coding platform that automatically scans codebases for vulnerabilities, suggests fixes, and integrates with CI/CD pipelines for continuous security analysis.

Mistral Small 4 is Mistral AI's efficient Mixture-of-Experts model — 119 billion total parameters with 128 experts and approximately 6 billion active per token, released under the Apache 2.0 license with a 256,000 token context window.

Claude Opus 4.7 is Anthropic's most capable generally available model — leading SWE-bench Verified (87.6%), with 3.75 megapixel vision, a new xhigh effort level for agentic work, and task budgets for controlling autonomous agent spend.

Claude Design is Anthropic's AI-powered design tool for quickly generating prototypes, presentation slides, one-pagers, and UI mockups from natural language — built on Claude Opus 4.7, with automatic brand system integration. As of April 28, 2026, it sits inside the broader Claude for Creative Work suite alongside MCP connectors for Adobe Creative Cloud, Canva Affinity, Ableton, Autodesk Fusion, Blender, and more.

Terafab is the $20-25 billion Tesla/SpaceX/xAI/Intel chip foundry joint venture announced March 2026 — targeting 1 terawatt of AI compute output per year, roughly 50× current global AI chip production, to vertically integrate silicon supply for FSD, Optimus, Dojo3, xAI training, and SpaceX orbital inference.

Starlink V3 is SpaceX's third-generation satellite platform — 1 Tbps download, terabit-class laser mesh, 60 per Starship launch — designed to double as distributed orbital AI compute nodes. In June 2026 SpaceX revealed AI1, its first dedicated orbital data center satellite, and the Gigasat factory in Bastrop, Texas that will build it, then completed a landmark Nasdaq IPO as SPCX.

Wispr Flow is an AI voice keyboard from Wispr AI — press a hotkey, dictate into any app, and Flow transcribes, removes filler words, and applies context-aware formatting in real time. Not affiliated with OpenAI's Whisper model.

Salesforce Einstein is the AI layer embedded across the world's largest CRM — predictive analytics, generative content, and conversational AI for sales, service, and marketing — now unified under the Einstein 1 Platform alongside Agentforce autonomous agents.

HubSpot AI — branded Breeze since 2024 — is the AI layer embedded across HubSpot's CRM platform, spanning Breeze Copilot, autonomous Breeze Agents (now on GPT-5), and Breeze Intelligence enrichment from a 200 million+ company database, with novel outcome-based pricing for agent work.

AutoCAD AI — Autodesk's machine-learning features inside the world's most-used 2D drafting software, anchored by Smart Blocks Detect and Convert, Markup Assist, Activity Insights, and the conversational Autodesk Assistant in AutoCAD 2026.

Autodesk Revit is the dominant BIM software for building design; Autodesk Insight (now Forma Carbon Insights) is the AI-powered building-performance analysis layer that adds energy simulation, embodied + operational carbon analysis, and generative-design alternatives directly inside the Revit ribbon.

Bentley iTwin is the open, cloud-based digital-twin platform for infrastructure — the engineering equivalent of a living model that fuses BIM design, reality capture, IoT sensor streams, and AI analytics for the entire lifecycle of bridges, roads, rail, water, energy, and substations.

ASML's extreme ultraviolet (EUV) lithography systems are the only machines on Earth capable of fabricating advanced AI chips at 3nm and below — making ASML the silent gatekeeper of the entire frontier-AI compute supply chain.

ARM is the chip architecture under most mobile AI inference globally; the Neoverse server platform has captured nearly 50 percent of hyperscaler shipments, and the new AGI CPU (March 2026) is ARM's first in-house data-center processor designed specifically for AI inference orchestration.

Broadcom's Tomahawk 5 (BCM78900) is the industry-leading 51.2 Tbps Ethernet switch ASIC powering AI data center fabrics — supports 64 ports of 800GbE and is now the dominant silicon in commercial AI Ethernet switches from Arista, FS, NADDOD, and others.

Cloudflare Workers AI is the serverless GPU inference platform running open-source AI models at 300+ global edge locations — pay-per-Neuron pricing with a free daily allocation, near-zero cold starts, tight integration with the Cloudflare developer platform, and a new May 2026 Stripe Projects beta that lets AI agents create Cloudflare accounts, buy domains, and deploy Workers without a human in the loop.

Docker Hub is the world's largest container registry with millions of AI/ML images; the new MCP Catalog adds 200+ verified Model Context Protocol servers as Docker images, making MCP integrations into Claude, Cursor, and other AI clients as simple as docker pull.

Extropic's Thermodynamic Sampling Unit (TSU) is a novel probabilistic AI chip using p-bits that fluctuate between states — a fundamentally different computing paradigm targeting up to 10,000x energy efficiency vs GPUs for diffusion-style AI workloads.

Fly.io GPU Machines are persistent global edge VMs with NVIDIA A10, L40S, and A100 GPUs deployable across 35 regions — designed for low-latency AI inference at the edge rather than centralized training.

G42 Cloud is the UAE-based AI cloud platform behind Stargate UAE — a $30 billion 1-gigawatt AI campus in Abu Dhabi co-built with OpenAI, Oracle, and NVIDIA, broken ground March 2026, with first 200MW cluster going live 2026 as the largest AI infrastructure project outside the United States.

HPE GreenLake for AI is the enterprise AI cloud platform combining Cray supercomputers, NVIDIA GPU systems, and as-a-service infrastructure — anchored by the new Cray GX5000 with GX240 liquid-cooled blades and Private Cloud AI for on-premises generative AI.

Intel Gaudi 3 is a 128GB HBM2e AI accelerator competing with NVIDIA H100/H200 — slower in raw throughput but priced at roughly half the cost per accelerator, making it the value-tier challenger for AI training and inference at scale.

Lambda Cloud is the GPU cloud built specifically for AI training and inference — on-demand NVIDIA H100, H200, and B200 access at competitive hourly rates with pre-installed AI software stacks, persistent storage, and high-speed networking.

Micron HBM3e is the high-bandwidth memory powering NVIDIA H200 and B200 AI accelerators — the third major HBM supplier alongside SK hynix and Samsung, with Micron capturing approximately 20-25 percent global market share by end of 2025 and HBM4 transition arriving late 2026.

NuScale VOYGR is the NRC-design-approved 77 MWe Small Modular Reactor — factory-built nuclear modules that scale up to 924 MW per plant, with a landmark 6 GW deployment program with TVA and ENTRA1 Energy targeting AI data centers and other critical infrastructure.

Oklo's Aurora powerhouse is a compact 75 MWe fast neutron microreactor designed for off-grid AI data center deployment — backed by Sam Altman, partnered with Switch for up to 12 GW of power supply, with first commercial deployment targeting early 2028.

Railway is a modern PaaS for full-stack AI applications — instant deployment from GitHub or Docker, built-in PostgreSQL/MySQL/Redis/MongoDB, persistent servers, per-second billing — popular with AI agent builders, though GPU instances are not currently offered.

Render is a modern PaaS popular with AI startups — auto-deploy from GitHub, free PostgreSQL/Redis/background workers, plus 50+ GPU models from RTX 3060 to B200 starting at $0.04/hour with no minimums or contracts.

Supermicro is the #2 AI server vendor globally, building liquid-cooled rack solutions optimized for NVIDIA HGX H200/B200/B300 — including the 4U liquid-cooled B200 system and ultra-dense 144-GPU per rack B300 deployments capturing 92% of heat for major data center power savings.

Tenstorrent Wormhole is the RISC-V AI accelerator from Jim Keller's chip design startup — open-source hardware architecture as a deliberate alternative to NVIDIA's closed CUDA ecosystem, with chiplet-based scaling and successor Blackhole already shipping.

TSMC fabricates over 90 percent of cutting-edge AI chips for NVIDIA, AMD, Apple, Broadcom, and Google — running 3nm at 180,000 wafer/month capacity by end-2026 and ramping five 2nm fabs simultaneously, with advanced-chip capacity booked through 2028.

VMware Private AI Foundation with NVIDIA is the enterprise platform for deploying generative AI on private infrastructure — combining VMware Cloud Foundation, NVIDIA AI Enterprise, GPU virtualization, RAG knowledge bases, and air-gapped deployment for regulated industries.

C3 AI Platform is the enterprise AI application platform with 100+ pre-built applications for predictive maintenance, supply chain, and fraud detection — and the new C3 Code agentic development tool that turns natural language into production-grade enterprise AI applications, scoring 9.2/10 in Anthropic's Claude evaluation.

Coframe is an AI-powered web optimization platform that autonomously generates and tests UI variations using generative AI and Thompson sampling — driving conversion lifts of 26-352 percent across customers, with the 2026 HaystacksAI acquisition expanding into agentic growth automation.

DataRobot is the unified Agent Workforce Platform for enterprise AI — combining AutoML, MLOps, and agentic orchestration into a single governed system, with the 2026 platform pivot from model builder to AI workforce orchestration layer for frontline business teams.

Fujitsu Kozuchi is Fujitsu's enterprise AI platform anchored by Takane LLM (world-class Japanese language proficiency, co-developed with Cohere) plus Digital Annealer quantum-inspired optimization and Data Intelligence PaaS — with a dedicated autonomous AI lifecycle platform launching July 2026.

Greenhouse AI Recruiting is the structured-hiring ATS used by approximately 4,000 companies — adding AI-powered resume screening, candidate summaries, scorecard generation, bias detection, and resume anonymization, with monthly third-party bias audits and ISO 42001 certification for AI governance.

Limitless was a wearable AI pendant for continuous conversation capture and intelligent recall — acquired by Meta in December 2025, with new Pendant sales discontinued and the company's technology and team being absorbed into Meta's smart-glasses and AI wearables roadmap.

Shield AI Hivemind is the world's first AI pilot with sustained real-world combat operational use — powering the V-BAT autonomous drone deployed by US Coast Guard, Indian Army, and allied nations, with Shield AI's Series G valuing the company at $12.7 billion in 2026 and projected revenue exceeding $540 million.

Twilio Voice AI is the conversational AI layer of Twilio's communications platform — ConversationRelay for natural voice agents with any LLM, Conversational Intelligence for call analysis, and tight Segment CDP integration for AI-driven personalization.

Autodesk Forma is the AI-driven early-stage architecture platform — descended from the 2020 Spacemaker acquisition — combining generative site automation, environmental impact analysis (sun, wind, noise, daylight, carbon), and the new Forma Building Design module entering closed beta late 2025.

TestFit is a real estate feasibility platform that auto-generates building configurations optimized for FAR, parking ratios, yield on cost, and unit mix — generating thousands of solves in seconds, with the 2026 Generative Design release allowing teams to explore solutions across multi-family, industrial, and mixed-use programs.

Trunk Tools is the construction-document AI for active job sites — TrunkSubmittal automates submittal review, TrunkText provides conversational access to drawings/specs/RFIs in seconds, and TrunkReview produces visual overlays of drawing changes — founded by Sarah Buchner in 2021.

AMI Labs is Yann LeCun's Paris-based AI startup founded March 2026, developing world models built on Joint Embedding Predictive Architecture (JEPA) — an alternative to autoregressive LLMs that learns abstract embeddings rather than predicting pixels or tokens, with 2026's LeWorldModel demonstrating up to 48x faster planning.

Apple Intelligence is the privacy-first personal AI built into iOS, iPadOS, and macOS — Writing Tools, Image Playground, Genmoji, notification summaries, and a rebuilt Siri whose cloud brain now runs on a custom Google Gemini model, processed on-device via Apple Silicon with Private Cloud Compute for complex requests.

Core ML is Apple's on-device machine learning framework for deploying trained models across iPhone, iPad, Mac, Apple Watch, and Apple TV — supporting PyTorch and TensorFlow conversion via coremltools, automatically routing inference across CPU, GPU, and Neural Engine, with low-bit quantization for compact deployments.

MLX is Apple's open-source machine learning framework purpose-built for Apple Silicon — exploiting unified memory architecture so CPU, GPU, and Neural Engine share the same memory pool, with M5 hardware now delivering under 10s time-to-first-token for dense 14B models on a MacBook Pro.

Sakana AI is the Tokyo-based foundation model startup co-founded by Transformer paper author Llion Jones — using nature-inspired evolutionary methods and swarm collective intelligence to produce efficient AI models, valued at $2.65 billion after a February 2026 Series B.

Tencent Hunyuan is the multimodal foundation model family powering AI across Tencent's WeChat (1.3 billion+ users), QQ, and Tencent Cloud — with Hunyuan 3.0 (April 2026 launch) bringing 295 billion parameters total, 21 billion activated, and integrated text/image/video/audio generation.

Abridge is the leading ambient clinical documentation AI for physicians — generating structured notes from patient encounters in real time with deep Epic integration, expanding into nursing documentation and real-time prior authorization via the January 2026 Availity partnership.

Altos Labs is the $3 billion-funded longevity research company applying AI to cellular reprogramming via Yamanaka factors — testing rejuvenation therapies in organs perfused outside the body, with 2024 mouse lifespan extension results pointing toward eventual clinical translation.

Augmedix is the ambient AI clinical documentation platform deployed across acute care, ambulatory, and skilled-nursing facilities — with bespoke Emergency Medicine app and AI fine-tuned for oncology and other specialties — now operating as a wholly-owned subsidiary of Commure following 2025 acquisition.

Bonterra is the social-good technology company building Bonterra Que (launched October 2025), the first agentic AI platform designed specifically for nonprofits, foundations, and government agencies — covering case management, fundraising, grant distribution, and program-outcome reporting.

Calico is the Alphabet-backed longevity biotech now operating as an independent clinical-stage company — applying AI and machine learning to aging biology with five clinical-stage candidates plus ~20 preclinical programs, including 2026 FDA Orphan Drug Designation for an Autosomal Dominant Polycystic Kidney Disease treatment.

Casebook is the cloud-based, configurable case-management platform created by the Annie E. Casey Foundation for human-services agencies — supporting caseworkers in child welfare, homelessness, housing, family services, and community-based initiatives with AI-assisted documentation.

Epic Cosmos is Epic's research database of 270 million patient records from 13 billion encounters — combined with embedded AI features (In-Basket ART message drafting at 1 million+ drafts/month, note summarization, AI agents for pre-visit prep) inside the Epic EHR used by most US health systems.

GE Healthcare Edison is the AI-enabled imaging platform integrated with GE's CT, MRI, X-ray, and ultrasound devices — including 2026 launches of the FDA-cleared Photonova Spectra photon-counting CT and AIRx AI-automated MRI workflow, with deep NVIDIA partnership for autonomous diagnostic imaging.

Innovaccer is the agentic AI cloud for healthcare — managing 80 million+ patient records across 130+ organizations including 6 of the top 10 US health systems, with $250 million in 2026 platform investment, the new Atlas Population Health OS, and CMS ACCESS model first-cohort participation.

IsoDDE is Isomorphic Labs' Drug Design Engine — DeepMind's spin-off's proprietary 'AlphaFold 4'-class AI for drug design, achieving 50 percent accuracy where AlphaFold 3 managed 23.3 percent on novel protein-ligand structures, with up to $1.7 billion in milestones from Eli Lilly and $1.2 billion from Novartis.

Manas AI is the AI-driven drug discovery platform co-founded by Reid Hoffman and Dr. Siddhartha Mukherjee — building a physics-based atomic 'world model' for drug binding, with $50+ million in seed funding and a January 2026 strategic agreement granting access to Schrödinger's physics-based computational platform.

MatrixCare AI is the cloud-based EHR for skilled nursing and senior living organizations — featuring nutrition management, financial and operations management, voice-enabled documentation, and integrations with emerging technologies like Tallio for automation.

Overjet is the industry-leading dental AI platform with 10 FDA-cleared modules covering caries and calculus detection (pediatric and adult), periapical radiolucency, automated dental charting, image enhancement, and CBCT Assist — used by dental service organizations and insurers.

Paige.AI is the digital pathology AI platform — first to receive FDA de novo approval for AI-based pathology with Paige Prostate, increasing pathologist sensitivity from 89.5 to 96.8 percent on prostate cancer detection — and now FDA Breakthrough Designation for the multi-cancer Paige PanCancer Detect.

PathAI is the AI-powered digital pathology platform helping pathologists and pharma companies analyze tissue samples with greater speed and diagnostic accuracy using deep learning — with a multi-year Memorial Sloan Kettering deployment for research and clinical applications.

Pearl AI is the FDA-cleared dental AI platform with seven FDA-cleared modules — covering multi-condition detection in bitewing and periapical images, periodontal bone level measurement, CBCT segmentation, and most recently FDA clearance for AI-assisted detection of pathologies on panoramic dental radiographs.

PointClickCare AI is the EHR and care-coordination platform for skilled nursing and senior living — KLAS-rated #1 Long-Term Care Software Provider for 6 consecutive years — with machine learning for resident risk prediction, regulatory compliance, and revenue cycle automation.

Suki AI is the voice-driven AI assistant for clinical documentation supporting 100+ specialties — the broadest specialty coverage of any ambient AI scribe — with deep real-time integrations into Epic, Oracle Health, athenahealth, and MEDITECH, cutting note time by 41 percent.

Wysa is the conversational AI for mental health support — clinically validated CBT-based chat assistant with over 6 million users across 95 countries, 45+ peer-reviewed studies, 31 percent average improvement in depression and anxiety scores, and the new Wysa Gateway streamlining patient access to therapy providers and health plans.

Applied Epic is the cloud-based insurance agency management system from Applied Systems — managing P&C and Benefits in a single platform — with the 2026 Applied Epic Max upgrade adding Epic Max Search for AI-powered information retrieval and embedded AI across client communications and producer workflow.

Aladdin is BlackRock's portfolio management and risk-analytics platform managing approximately $25 trillion in assets across 1,000+ organizations — with new AI-powered Auto Commentary tool launched at Morgan Stanley and December 2025 AWS partnership for 2H 2026 GA alongside existing Azure deployment.

Bloomberg AI is the AI feature suite integrated across the Bloomberg Terminal — earnings call summaries, AI-powered news summaries, document Q&A across earnings transcripts and research, and the new ASKB conversational interface coordinating AI agents to dynamically access Bloomberg data, news, and research.

Guidewire is the dominant P&C insurance core-system platform — PolicyCenter, ClaimCenter, BillingCenter, UnderwritingCenter — with 2026 AI launches including ProNavigator AI assistant, Claims Intel built on Industry Intel dataset, and LLM-powered submission processing in UnderwritingCenter.

JPMorgan LLM Suite is the in-house enterprise AI platform deployed to 200,000+ employees daily across 450+ AI use cases — from fraud detection saving $1.5 billion to risk modeling, trading, credit underwriting, and compliance — with plans to expand to 1,000 use cases by end of 2026 under a $1.8 billion AI investment program.

MindBridge is the AI-powered audit and risk-analytics platform that analyzes 100 percent of transactions for anomalies — replacing traditional manual sampling — used by audit firms and corporate finance teams with the Ensemble AI engine assigning risk scores to every transaction.

NICE Actimize is the AI platform for anti-money-laundering (AML), fraud detection, and financial-crime compliance — used by 1,000+ organizations across 70+ countries — with the Xceed AI FRAML platform combining fraud and AML, and the January 2026 Insights Network for real-time counterparty risk visibility.

Sage Copilot is Sage's AI-powered productivity assistant embedded across Sage Accounting, Sage Sole Trader, Sage X3, and other ERP products — surfacing anomalies, automating data categorization and receipt processing, with early adopters reporting 5+ hours per week saved on manual tasks.

Shift Technology is the AI-powered insurance fraud detection and claims-decisioning platform used by global P&C and health insurers — applying machine learning across claims, underwriting, and SIU (Special Investigations Unit) workflows to identify suspicious claims and accelerate legitimate claim processing.

Verisk AI is the insurance analytics platform embedded across most US P&C carriers — AI for underwriting, claims, fraud detection, and catastrophe modeling — with 2026 GenAI-powered Augmented Underwriting transforming unstructured submissions into structured information.

Vertafore is the leading insurance technology platform for agencies — competing with Applied Epic — with the 2026 launch of Velocity AI Platform unveiled at Accelerate 2026 (2,000+ attendees) embedding AI across the insurance lifecycle for client retention, document automation, and producer productivity.

Vic.ai is the AI-first accounts payable automation software achieving 5x efficiency, 99 percent accuracy, and 85 percent no-touch invoice processing — built by accountants and trained on over 1 billion invoices for accounting firms and corporate finance teams.

Xero is the cloud-based small-business accounting platform that introduced Xero OS in April 2026 — an AI-native operating system powered by JAX (Just Ask Xero), the AI assistant designed to bring CFO access to businesses at every stage by automating bills, payments, and reconciliation end-to-end.

AlphaFold is DeepMind's protein-structure prediction AI — predicting structures for over 200 million proteins via the AlphaFold Protein Structure Database — with AlphaFold 3 (open-sourced 2024) extending to predict structures of proteins, DNA, RNA, small molecules, and ions in a single pass.

Anduril Lattice is the defense AI operating system for autonomous sensors, drones, and command-and-control — selected by the US Army in March 2026 for an enterprise contract with a $20 billion ceiling over 10 years and an initial $87 million counter-UAS task order.

BenevolentAI is the AI drug-discovery platform combining knowledge graphs and machine learning to identify novel drug targets and repurposing opportunities — with deep partnerships including Sanofi for AI-driven target identification across multiple disease areas.

Cadence Cerebrus is the AI-driven chip implementation tool using reinforcement learning to automate physical-design closure — competing directly with Synopsys DSO.ai for the floorplan, placement, and routing automation work that determines how chips are physically built.

CheggMate is the AI-powered learning assistant from Chegg built in partnership with OpenAI — providing homework help, step-by-step explanations, AI tutoring, and writing assistance for college students alongside Chegg's existing question-and-answer database.

Datadog AI is the cloud monitoring and security platform's AI layer — anomaly detection, automated incident correlation, and intelligent alerting for engineering teams managing complex distributed systems — with growing AI workload observability for LLM applications.

Duolingo Max is the AI-powered language learning tier on Duolingo's platform serving 100 million+ monthly users — featuring AI roleplay conversations, explain-my-answer, personalized lessons via Birdbrain AI engine, and speech recognition across 40+ courses.

GitLab Duo is GitLab's AI-powered DevSecOps suite — code suggestions, vulnerability resolution, test generation, autonomous Planner Agent — embedded across the full software development lifecycle in a single platform.

Insilico Medicine is the AI-driven drug discovery platform (Pharma.AI suite) covering target identification, generative chemistry, and clinical-trial design — with multiple AI-discovered candidates already in Phase II clinical trials, validating the AI-to-clinic translation.

JetBrains AI Assistant is the AI coding assistant integrated across all JetBrains IDEs (IntelliJ IDEA, PyCharm, WebStorm, GoLand, etc.) — featuring code completion, explanation, refactoring, Recap, Insights, Next Edit Suggestions, and Claude Agent integration.

Kakao Kanana is the Korean-language AI model designed for on-device deployment and KakaoTalk integration — bringing AI assistant capabilities (search, scheduling, task management) to 50 million+ KakaoTalk users across South Korea's dominant messaging platform.

Lexis+ AI is LexisNexis's AI legal-research and drafting platform — competitive with Westlaw and Harvey for case-law search, summarization, document drafting, and regulatory analysis — built on LexisNexis's massive legal corpus including case law, statutes, secondary sources, and regulations.

Mobileye is the vision-based ADAS and autonomous-driving AI platform (Intel-owned) supplying technology to over 30 global automakers — covering Advanced Driver Assistance Systems and progressing toward higher levels of vehicle autonomy.

Naver Agent N is the AI agent platform for South Korea's dominant search engine — AI shopping agents, search assistants, and autonomous task completion for 60 percent+ of Korean search traffic — launching H1 2026.

Redis Vector Search adds in-memory vector similarity search to the world's most popular data store — ultra-low-latency semantic search, LLM response caching, and real-time recommendations using Redis's familiar APIs across millions of existing deployments.

Schrödinger is the computational platform combining physics-based simulation and machine learning for drug discovery and materials science — used across pharmaceutical R&D for decades and now augmented with AI, including the January 2026 strategic agreement granting Manas AI significant access.

scispace (formerly Typeset) is the AI tool for scientific paper analysis, literature review, and writing — extracting answers from research PDFs and helping researchers explore citation networks across hundreds of millions of published papers.

Siemens Industrial Copilot is the generative AI assistant for shop-floor engineers and operators — integrated across Siemens automation, PLM, and digital-twin platforms (TIA Portal, NX, Teamcenter, Xcelerator) for industrial manufacturing AI.

Spellbook is the AI suite for commercial law operating entirely within Microsoft Word — automating contract redlines, suggesting clauses, flagging risks, and drafting assistance — used by 4,000+ in-house teams and law firms with starting price of $99/user/month.

Synopsys DSO.ai is the AI-driven chip-design optimization platform applying reinforcement learning to floorplan, placement, and routing — used by major fabless and IDM semiconductor firms — and the direct competitor to Cadence Cerebrus for AI-driven physical design closure.

Tabnine Enterprise is the privacy-first AI code completion tool with on-premise deployment for regulated industries — supporting 80+ languages across VS Code, JetBrains, and Neovim, with custom model fine-tuning on proprietary code for enterprises that can't use cloud-only AI.

Tegus is the expert-call platform with AI-generated transcripts and summaries — used by management consultants, hedge funds, and corporate strategy teams for primary research — providing access to thousands of expert-call transcripts plus AI-augmented analysis.

Wiz is the cloud security platform using AI to identify and prioritize critical risks across AWS, Azure, Google Cloud, and Oracle Cloud — completing Google's $32 billion acquisition March 11, 2026 (Google's largest-ever) while continuing as a multicloud platform with expanded AI-era security capabilities.

Wealthfront is the longest-running independent direct-to-consumer robo-advisor — automated portfolio management, daily tax-loss harvesting, and AI-driven cash management for retail investors with $50+ billion in assets under management.

Betterment is the largest independent robo-advisor in the US — automated investing for retail clients, a 401(k) platform for small employers (Betterment for Business), and a B2B platform serving thousands of independent financial advisors (Betterment for Advisors).

Orion Advisor Tech is the dominant AI-augmented platform for registered investment advisors — portfolio management, financial planning, CRM, compliance, and reporting across thousands of RIA firms managing roughly $3 trillion in assets.

eMoney Advisor is the dominant retail financial-planning platform — Fidelity-owned, AI-driven cash-flow analysis and retirement scenario modeling used by approximately 95,000 financial advisors and powering planning conversations for millions of households.

Skydio is the leading US-built autonomous drone manufacturer — its AI-driven flight stack handles obstacle avoidance, path planning, and target tracking without a pilot in the loop, deployed across industrial inspection, public safety, and defense customers.

DroneDeploy is the leading drone-software platform for surveying, mapping, and construction site monitoring — AI-powered photogrammetry turns drone flights into 2D orthomosaics, 3D models, and BIM-integrated progress reports.

Pix4D is the survey-grade photogrammetry standard — Swiss-built desktop and cloud software for converting drone imagery into orthomosaics, digital surface models, 3D point clouds, and survey-precision deliverables for surveying, agriculture, and construction.

OpenSpace is the leading 360° construction-site capture platform — workers walk job sites with helmet-mounted cameras and AI automatically maps every photo to BIM coordinates, surfacing deviations between as-built reality and the design model.

Buildots is the schedule-grade AI construction-progress platform — 360° helmet-camera site walks compared against the BIM schedule to flag delays and rework risk before they cascade into budget overruns.

Esri ArcGIS AI brings generative AI to the dominant geographic information system platform — natural-language spatial queries, automated feature extraction from imagery, and predictive analytics across utilities, government, and infrastructure use cases.

Hinge Health is the dominant digital MSK platform — smartphone-based exercise programs, AI motion tracking, and care navigators for back, neck, joint, and pelvic-floor pain delivered to 19+ million people through employer benefit programs.

Kaia Health is the broad digital therapeutics platform — AI-driven exercise programs for back pain, knee pain, and COPD pulmonary rehab with FDA clearance for several indications, deployed across major US and European employer health benefits.

SWORD Health is the FDA-cleared sensor-driven digital physical therapy platform — motion sensors plus licensed PT supervision deliver in-clinic-quality MSK, women's health (Bloom), and post-surgical rehab care to thousands of employer and health plan partners covering millions of members.

CarePredict is the wearable AI for elderly activity, fall, and behavior-pattern monitoring — ML models on continuous accelerometer data detect routine changes that predict urinary tract infection, falls, and depression days to weeks ahead of human caregivers.

Inspiren is the AI vision platform (AUGi) for senior-living facilities — tracks falls, exit-seeking, and clinical events from in-room cameras without requiring residents to wear devices, used by Brookdale, Atria, and other large senior-living operators.

Sudowrite is the AI writing assistant built specifically for fiction authors — generates plot ideas, expands prose, brainstorms character details, and rewrites passages with novel-craft conventions baked into the model behavior.

Automated Insights is the natural-language-generation pioneer — its Wordsmith platform automatically writes data-driven narratives from structured datasets, used by AP, Yahoo, and enterprise customers to scale routine reporting at thousands of stories per minute.

Reflection AI is an open-source frontier AI lab building large-scale Mixture-of-Experts foundation models with reinforcement learning, focused initially on autonomous coding. Founded by ex-DeepMind and ex-Google Brain researchers; raised $2 billion in October 2025. In May 2026 it secured two federal deployments: the Pentagon for Impact Level 6 and 7 classified networks, and the US Department of Energy as primary model provider across all 17 national laboratories for the Genesis Mission.

Legora is the Stockholm-based legal AI platform serving 1,000-plus law firms across 50 markets — collaborative document review, contract analysis, due diligence, and AI-native research. April 2026 Series D extension led by NVentures and Atlassian valued the company at $5.6 billion post-money on more than $100 million ARR, positioning Legora as Harvey's most direct global competitor. May 2026 brought Anthropic into the same market with the first frontier-lab vertical legal product line.

AMD Instinct is the company's datacenter AI accelerator line — currently the MI355X (CDNA 4, 288 GB HBM3e, GA Q3 2025), with the MI400 series announced at CES 2026 to compete with NVIDIA Vera Rubin. Major customer wins include Oracle, Microsoft Azure, OpenAI, and Meta.

AMD ROCm is the company's open, permissively licensed AI compute software stack — the deliberate counterweight to NVIDIA's CUDA. ROCm 7.2.2 ships with vLLM 0.10.1, native PyTorch on Windows + Linux, and a vLLM CI pass rate that jumped from 37 percent to 93 percent across late 2025.

AMD Ryzen AI brings on-device AI to laptops and mini-desktops via the XDNA 2 NPU. The flagship Ryzen AI Max+ 395 (Strix Halo) delivers 50 TOPS NPU, 126 TOPS system-wide, and up to 128 GB unified memory — enough to run open-weight LLMs up to roughly 70 billion parameters locally.

AMD EPYC AI is the server-CPU side of the AI stack — currently shipping EPYC 9005 'Turin' (Zen 5, full-width AVX-512 with VNNI), with Venice (Zen 6, up to 256 cores, H2 2026) pairing 1.6 TB/s memory bandwidth with the new x86 ACE matrix-AI standard AMD co-authored with Intel.

Genesis AI is a Khosla- and Eclipse-backed full-stack robotics startup that pairs the GENE-26.5 robotics foundation model with proprietary human-anatomy-mimicking robotic hands and a sensor-laden data collection glove — demoed cooking, playing piano, and solving Rubik's cubes.

ZAYA1-8B is San Francisco lab Zyphra's 8.4 billion-parameter mixture-of-experts model with only 760 million active parameters at inference — matching DeepSeek-R1 on math benchmarks while staying competitive with Claude Sonnet 4.5 on reasoning. Trained entirely on AMD Instinct MI300X GPUs with no NVIDIA dependency. Apache 2.0.

XFRA is Span's distributed AI inference data center, built from 16-GPU NVIDIA Blackwell compute nodes installed in homes and small commercial sites. By tapping the roughly 60 percent of residential electrical capacity that sits idle behind a Span Panel, XFRA aims to deliver gigawatt-scale inference compute by 2027 — pitched as roughly six times faster and five times cheaper than the equivalent centralized data center buildout.

AlphaEvolve is Google DeepMind's Gemini-powered coding agent that designs and optimizes algorithms across scientific research and infrastructure — with concrete deployments at Klarna, FM Logistic, Schrödinger, WPP, plus measurable wins in quantum circuits, DNA sequencing, Google Spanner, and TPU compiler design.

OpenAI's Realtime API is the unified WebSocket-based voice surface for builders — now hosting GPT-Realtime-2 (GPT-5-class voice reasoning), GPT-Realtime-Translate (real-time translation across 70+ languages), and GPT-Realtime-Whisper (live speech-to-text), all launched May 7, 2026.

Mojo is Modular's AI-native programming language — a Python superset that targets CPUs, GPUs, and AI accelerators from a single codebase, with native Python interop. Hit 1.0 Beta on May 7, 2026.

Ant Group / InclusionAI's 1 trillion parameter open-weights model under MIT license — hybrid Multi-head Latent Attention plus Linear Attention, 262,144-token context, 72.2 on SWE-bench Verified.

Vapi is the voice-agent platform that processes over 1 billion calls — chosen by Amazon Ring over 40 competitors and serving Kavak, Instawork, New York Life, and Intuit. Its May 2026 Series B raised $50 million at a $500 million valuation, led by Peak XV with Microsoft's M12, Kleiner Perkins, and Bessemer participating.

Circle Agent Stack is an open-source set of primitives — Circle CLI, Agent Wallets, an Agent Marketplace, and gas-free Nanopayments — that let autonomous AI agents hold, discover, and spend USDC across blockchains, paired with Circle's new Arc settlement layer.

SANA-WM is NVIDIA Labs' 2.6 billion parameter open-source video world model — Apache 2.0, 720p, one-minute generation with 6 degrees of camera-pose control, designed as a baseline for embodied-AI and robotics research at consumer-GPU compute budgets.

Semble is MinishLab's open-source code-search library purpose-built for AI coding agents — claims roughly 98% fewer tokens than grep-plus-read pipelines at higher recall (94% at 2,000 tokens versus 85% at a full 100,000-token context window), with sub-2-millisecond query latency. Combines tree-sitter chunking, Model2Vec static embeddings, BM25 lexical retrieval, and reciprocal-rank fusion.

Alexa+ Podcasts is Amazon's on-demand generative audio feature that turns any topic into an AI-narrated podcast episode in minutes, grounded in licensed content from the Associated Press, Reuters, The Washington Post, and over 200 local U.S. newspapers. Launched May 2026 and bundled into Alexa+ at no extra cost.

SandboxAQ's Large Quantitative Models (LQMs) bring simulation-grade quantum chemistry, molecular dynamics, and microkinetics into Claude as a conversational interface for drug discovery and materials science. Spun out of Alphabet in 2022 and led by Jack Hidary, SandboxAQ shipped its Claude integration in May 2026.

Agora-1 is Odyssey's learned world model — a real-time multi-agent simulator that lets up to four humans or AI agents share the same generated environment. Launched May 19, 2026, the system decouples simulation from rendering and demos as a four-player deathmatch, with longer-term applications in robotics, defense, and reinforcement learning.

Gemini Spark is Google's always-on personal agentic assistant, first unveiled at Google I/O 2026 and now live for US Google AI Ultra subscribers on web, Android, and iOS. Built on Gemini 3.5 Flash, Spark monitors user-set triggers around the clock and dispatches multi-step actions across Gmail, Calendar, Drive, and Workspace.

Gemini Omni is Google's multimodal generation model unveiled at Google I/O 2026 — it turns combinations of images, audio, and text into video. Positioned as the modal sibling to Veo 3 and Nano Banana Pro inside Google's generative media stack.

Ocean is an AI-native email security platform that launched from stealth in May 2026 with $28 million led by Lightspeed Venture Partners. Its agentic investigation engine, Ray, analyzes every incoming email in real time to flag fraud, impersonation, and AI-driven phishing — already protecting Kayak, Kingston Technology, and Headspace.

NVIDIA's first CPU purpose-built for agentic AI workloads — architected to pair with GPU inference for autonomous agents that behave like users. Positioned by CEO Jensen Huang as a new $200 billion total addressable market for NVIDIA, Vera CPU shipped to partner labs in mid-2026 with $20 billion in standalone year-to-date sales already booked.

Stability AI's four-model family for music generation — medium and large variants render compositions up to six minutes and twenty seconds, with three of the four variants shipping under open weights. Training data is fully licensed via direct partnerships with Warner Music Group and Universal Music Group — a deliberate contrast with Suno and Udio, both of which face major-label copyright suits.

Be My Eyes is the world's largest accessibility platform — a free mobile app and AI vision service connecting more than one million blind and low-vision users to GPT-4-powered descriptions and a community of more than ten million sighted volunteers, now integrated into Ray-Ban Meta and Oakley Meta smart glasses.

OpenRouter is a multi-model API gateway routing requests across 400+ LLMs from Anthropic, Google, OpenAI, xAI, DeepSeek, and others — letting developers optimize each request for cost, latency, or capability through a single unified API. CapitalG-led $113 million Series B in May 2026 at a $1.3 billion valuation; ~8 million users; 100 trillion tokens per month.

Mistral Physics AI is a class of data-driven foundation models for industrial engineering — learning from physics-solver outputs to predict the behavior of physical systems in seconds on a single GPU, replacing simulations that traditionally take hours or weeks. Production customers include Airbus, ASML, BMW Group, Safran, and Siemens Energy, with deployment backed by a new 10 megawatt Mistral-operated inference data center in Les Ulis south of Paris.

Robinhood Agentic Trading is a beta feature that lets users connect AI agents to a dedicated brokerage wallet over the Model Context Protocol (MCP) — letting agents analyze portfolios, look up analyst notes, and place stock trades within user-approved limits. It is the first retail-facing brokerage to expose live trading to third-party AI agents.

Claude Opus 4.8 is a flagship-tier Anthropic model, now positioned just below the new Claude Fable 5 as the economical default for routine professional work — and the model Fable 5 falls back to on sensitive requests. It ships with Dynamic Workflows to coordinate hundreds of parallel subagents on codebase-scale migrations, at $5 input / $25 output per million tokens, with a Fast mode at 2.5-times speed.

Liquid LFM is the family of open-weight Liquid Foundation Models from MIT CSAIL spinoff Liquid AI, engineered for on-device inference. Current flagship LFM2.5-8B-A1B is an 8 billion-parameter mixture-of-experts model with roughly 1 billion active parameters per token, pretrained on 38 trillion tokens, with a 128,000-token context window and day-one runtime support across llama.cpp, MLX, vLLM, SGLang, and ONNX — delivering around 253 tokens per second on an Apple M5 Max and roughly 30 tokens per second on flagship smartphones.

Step 3.7 Flash is the May 2026 flagship open-weights release from Shanghai-based StepFun — a 198 billion total parameter mixture-of-experts vision-language model with roughly 11 billion active parameters per token, a 256,000-token context window, up to 400 tokens per second of throughput, and Apache 2.0 weights. The model targets agentic coding and search workflows with reported parity against frontier models at a fraction of the per-task cost.

Assured Robot Intelligence is the humanoid-robotics foundation models startup acquired by Meta in May 2026 and folded into Meta Superintelligence Labs — co-founded by ex-NVIDIA researcher Xiaolong Wang and ex-Fauna Robotics co-founder Lerrel Pinto to build foundation models for whole-body humanoid control and self-learning.

AI @ Morgan Stanley is the firmwide internal generative-AI assistant deployed to Morgan Stanley wealth advisors — built on OpenAI models, it provides instant access to 100,000+ research reports plus client-meeting prep, Debrief meeting notes, and email/document drafting. The most-publicized example of a major wealth manager rolling out an AI copilot at scale.

WeatherMesh is WindBorne Systems' AI weather-forecasting model, trained on data from a global constellation of long-duration smart weather balloons. The June 2026 WeatherMesh-6 delivers hourly, 3-kilometer forecasts and, by the company's account, matches the accuracy of a traditional five-day forecast a full day earlier — outpacing the world's leading numerical weather model.

Intel Crescent Island is a forthcoming data-center GPU built for AI inference on the Arc Xe3P architecture. Detailed at Computex 2026 and expected to launch in the second half of 2026, it carries up to 480 gigabytes of cost-optimized LPDDR5X memory instead of scarce high-bandwidth memory — Intel's bid to undercut NVIDIA and AMD on inference cost rather than peak training performance.

MAI-Code-1-Flash is Microsoft's lightweight, in-house coding model, built end-to-end for GitHub Copilot. Microsoft says it outperforms Anthropic's Claude Haiku 4.5 across its coding benchmarks — including a roughly 16-point lead on SWE-Bench Pro — while using up to 60 percent fewer tokens, and it rolls out to Visual Studio Code Copilot users via the model picker.

Cisco AI Defense is an enterprise platform for securing the AI applications, models, and agents organizations build and run — combining algorithmic red-teaming, runtime guardrails, model and MCP-server scanning, and real-time inspection of agentic traffic against threats like memory poisoning, tool misuse, and intent hijacking.

Cisco Hypershield is a distributed, AI-native security architecture for AI-scale data centers — pushing enforcement into the Linux kernel via eBPF and into the network through Smart Switches, with an Autonomous Segmentation engine and self-qualifying policy updates that test changes against a digital twin before applying them.

Cisco Secure AI Factory with NVIDIA is a validated reference architecture for mass-scale AI data centers — combining Cisco Silicon One networking, 8000 Series and N9300 switches, and Hypershield security with NVIDIA accelerated computing to connect AI accelerators scale-up in the rack, scale-out across rows, and scale-across between distant data centers.

MAI-Thinking-1 is Microsoft's first in-house frontier reasoning model, unveiled at Build 2026. A roughly 35-billion-parameter model with a 256K-token context window, Microsoft says it matches Anthropic's Claude Opus 4.6 on the SWE-Bench Pro coding benchmark. It anchors a new family of seven first-party MAI models and is in private preview on Azure AI Foundry.

Flourish is a New York research lab building brain-inspired 'Cortex AI' models that aim to match today's AI on a fraction of the power — targeting roughly 20 to 50 watts. Co-founded by Internet Explorer creator and CTRL-labs founder Thomas Reardon, it emerged from stealth in 2026 with a $500 million round at a $2.5 billion valuation, backed by Jeff Bezos. No commercial product yet.

The Supabase MCP Server is Supabase's official Model Context Protocol server — it connects AI coding agents like Cursor, Claude, Windsurf, GitHub Copilot, and Cline directly to a Supabase backend, exposing more than 20 natural-language tools for the database, edge functions, storage, branching, and debugging.

database.build (formerly postgres.new) is Supabase's free, in-browser Postgres sandbox with an AI assistant — powered by PGlite, a WebAssembly build of Postgres, it lets you describe a database in plain language and have AI scaffold the tables, import CSVs, and run real SQL, entirely in the browser.

Claude Fable 5 is Anthropic's new public flagship — a safeguarded version of its Mythos-class model that is state-of-the-art on nearly every capability benchmark, from software engineering to scientific research. It works autonomously across millions of tokens of memory, ships at $10 input / $50 output per million tokens, and falls back to Opus 4.8 on sensitive cybersecurity and biology requests. Its restricted sibling, Mythos 5, lifts those safeguards for vetted partners.

North Mini Code is Cohere's first agentic coding model and its first open-weights release — a 30 billion-parameter mixture-of-experts design that keeps just 3 billion parameters active, handles a 256,000-token context, and ships under Apache 2.0 on Hugging Face. Cohere reports 33.4 on the Artificial Analysis Coding Index and up to 2.8-times the output throughput of Devstral Small 2 on comparable hardware.

Enflame CloudBlazer is the family of data-center AI accelerators from Shanghai chipmaker Enflame, built on its in-house DTU (Deep Thinking Unit) architecture with separate training and inference lines. Backed by Tencent and positioned as a domestic Chinese alternative to NVIDIA's data-center GPUs, it is tied to Enflame's mid-2026 Shanghai STAR Market IPO.

Neura 4NE-1 is German company Neura Robotics' humanoid robot, built for series production and powered by its cognitive-robotics stack and Neuraverse skills ecosystem. Designed for manufacturing and warehouse work, it sits at the center of Neura's 2026 funding round — one of the largest in robotics history, backed by Amazon, NVIDIA, Bosch, and others.

Qwen-Robot is Alibaba Tongyi Lab's suite of three foundation models for embodied AI — covering navigation, manipulation, and physics-aware world prediction — a software stack meant to act as a common operating layer for the next generation of robots.

Enterprise AI application platforms are the critical layer between foundation models and business value — translating raw model capability into measurable outcomes at organizational scale.

The most common way professionals encounter AI at work is through SaaS tools they already use — and most are now embedding AI directly into their core workflows, changing what's possible without switching tools.

AI agents go beyond chatbots by perceiving their environment, reasoning about what to do, and taking real-world actions — making them the dominant pattern for production AI deployments.

Every AI agent is built from five core components: a perception module, memory systems, planning and reasoning, tool use, and action output — understanding each is essential for building or evaluating agents.

MCP is the open standard that lets any AI agent connect to any tool or data source — ending the fragmented ecosystem of one-off integrations and enabling the truly composable AI stack.

A practical guide to the frameworks, patterns, and orchestration models developers use to build production AI agents — from the dominant ReAct pattern to multi-agent architectures and open-source frameworks.

Agentic AI introduces failure modes that don't exist in single-turn LLM interactions — from error compounding and prompt injection to the cost of autonomous action — and getting safety design right is what separates working production agents from dangerous ones.

A ranked, practical guide to the leading AI models for software development — what each one is best for, how they compare on benchmarks, and how to choose between them for your coding workflow.

MCP servers transform AI coding tools from text generators into full engineering collaborators — and project configuration files like AGENTS.md and SKILL.md give agents the context they need to be immediately productive in your specific codebase.

The AI-powered IDE landscape has split into two categories — AI-native editors rebuilt from scratch with AI as the primary interface, and AI-augmented editors that add AI layers to established foundations like VS Code. In April 2026, SpaceX secured a $60 billion option to acquire Cursor, the category leader.

Browser-based coding environments and AI-first app generation platforms let you build and deploy applications from a description — no local setup, no infrastructure configuration, from concept to live URL in minutes.

AI-powered CLI tools bring agentic coding capability directly to the terminal — integrating with Unix pipelines, git workflows, and CI systems in ways that browser-based tools cannot match.

GitHub has become the integration hub for AI coding tools — from GitHub Copilot across every IDE to autonomous agents that read issues, implement features, and open pull requests without human intervention at each step.

Modern hosting platforms have removed most infrastructure complexity from web application deployment — a git push to a connected repository is all it takes to deploy globally, with SSL, CDN, and preview environments included automatically.

AWS, Microsoft Azure, and Google Cloud — the big three cloud providers — offer distinct AI development ecosystems that reflect their unique advantages: AWS's breadth, Azure's OpenAI partnership, and Google's first-party Gemini models.

The physical infrastructure required to train and run frontier AI models is now a strategic constraint — from energy consumption measured in gigawatts to GPU supply chains and cooling systems that push the limits of known engineering.

The AI chip landscape spans NVIDIA's dominant GPU ecosystem, AMD's memory-rich challengers, Apple's unified silicon for local AI, and a growing array of custom ASICs from Google, AWS, and Cerebras — each with distinct tradeoffs in performance, cost, and ecosystem.

Edge AI — running model inference locally or on-device rather than in the cloud — addresses privacy, latency, cost, reliability, and data sovereignty requirements that cloud-only approaches cannot meet.

The right database and payment stack can be the difference between a weekend prototype and a production-ready SaaS — and AI coding tools have dramatically lowered the barrier to implementing both correctly.

Survey the AI developments we can be most confident about over the next two years — agentic AI proliferation, multimodal defaults, vertical-integration chip plays like Terafab, orbital AI data centers, physical AI, and the cost collapse of frontier intelligence.

Explore the more uncertain but grounded medium-term developments expected between 2028 and 2035 — scientific acceleration, personal AI with full life context, AI-designed AI, brain-computer interfaces, and the energy infrastructure AI demands.

A carefully framed exploration of long-term AI possibilities — AGI, superintelligence, the alignment problem, economic transformation, and AI-accelerated longevity — with appropriate uncertainty throughout.

Map the major schools of thought shaping the AI debate — techno-optimists, effective accelerationists, AI safety researchers, AI ethics scholars, and AI skeptics — with their real arguments, blind spots, and how to engage with each.

Explore AI's broad societal consequences — economic disruption, threats to democracy and information integrity, privacy and surveillance, and the copyright questions reshaping creative industries.

Understand the seven core principles of responsible AI — fairness, accountability, transparency, privacy, safety, human oversight, and accessibility — and the governance frameworks giving them legal force. May 2026 brought a third major AI-chatbot safety lawsuit (the teen ChatGPT drug-combination case), establishing a litigation pattern that's reshaping how foundation-model providers handle vulnerable users.

A practical resource guide for staying current in AI — the best newsletters, YouTube channels, courses, and policy reports for different learning goals, with guidance on building sustainable learning habits.

Six concrete first steps for beginning your AI journey, a reflection on what you have accomplished across the full curriculum, and a closing framework for staying curious in a field that never stops moving.

AI adoption is an execution challenge, not a strategy challenge — a practical guide for the people who make it actually happen.

KPIs for AI adoption, building a dashboard, reporting to leadership, and proving ROI with data your organization cares about.

A step-by-step guide to running a 30-day AI pilot with your team — scope, metrics, tools, timeline, and how to report results.

Why people resist AI adoption and exactly how to address each type of resistance — from fear to skepticism to legitimate concerns.

Design an effective AI training program for your team — what to teach, how to teach it, and the mistakes that kill adoption.

Build your first AI agent step by step — a practical, working agent that uses tools to accomplish real tasks.

Your starting point for building AI agents — what agents are, why they matter, and what you will build in this playbook.

The major agent architecture patterns — single agent, router, orchestrator, and pipeline — and when to use each one.

Taking your agent from a working prototype to a reliable production system — scaling, monitoring, cost management, and human-in-the-loop design.

The hardest part of agent development — how to test nondeterministic systems, common failure modes, and debugging strategies.

ROI frameworks for AI investment, build vs. buy decisions, and where AI delivers the fastest returns for most organizations.

Practical AI governance for business leaders — policies, vendor evaluation, data governance, and compliance without bureaucratic paralysis.

AI is a strategic business decision, not a tech project — a leadership playbook for executives navigating AI adoption.

The human side of AI adoption — change management, upskilling your team, handling AI anxiety, and restructuring roles.

A phased roadmap for AI adoption — from first pilot to organization-wide integration, with decision points at each stage.

A concrete 30-day plan with weekly milestones to start future-proofing your career — from assessment to action.

A personal framework to evaluate how AI will affect YOUR specific role — honest self-assessment with actionable categories.

How professionals are using AI to become more valuable, not less — real examples, mindset shifts, and practical strategies.

An honest starting point for understanding how AI affects your career — no hype, no panic, just a clear plan to stay ahead.

Practical guide to introducing AI at work — talking to your manager, navigating company policy, upskilling, and positioning yourself as a leader.

How AI bias shows up in the real world, how to recognize it, and what you can do about it — as a user, a professional, and a citizen.

Copyright, attribution, creative displacement, and disclosure — navigating the ethical landscape of AI-generated content.

Build your personal AI ethics framework — not adopting someone else's rules, but developing your own principled approach to AI decisions.

AI ethics is not abstract philosophy — it is about the decisions you make every day as an AI user, builder, or leader.

Making ethical AI decisions in your professional life — when to use AI, when to push back, and how to advocate for responsible practices.

Design a sustainable AI information diet — what to follow, what to ignore, and how to spend your 30 minutes per week wisely.

AI moves fast — but keeping up does not have to be a full-time job. Build a sustainable system for staying current.

A repeatable monthly routine to stay current with AI — try one tool, read one report, update one workflow. Under 2 hours per month.

How to interpret AI announcements, benchmark claims, and hype cycles — so you can separate genuine breakthroughs from marketing noise.

A practical framework for comparing AI models — benchmarks, context windows, pricing, and how to choose the right model for your needs.

The near-term future of AI models — multimodal, agents, reasoning chains, and what these trends mean for how you use AI.

Why understanding AI models matters — make better tool choices, have informed conversations, and see through the marketing hype.

The big picture of who is building AI models, how they compete, and the dynamics shaping the industry in 2026.

The rise of efficient AI models that run on your laptop — why smaller models matter and when to choose them over frontier giants.

Practical, age-appropriate AI rules for elementary, middle, and high school students — including homework policies, screen time, and supervision levels.

How to have productive conversations about AI with your kids — conversation starters, family activities, and building AI literacy together.

A parent's guide to the best AI tools for families — kid-friendly chatbots, educational AI, parental controls, and what to avoid.

Your starting point for understanding AI as a parent — what you will learn, why it matters for your family, and how to get the most from this playbook.

Discover how children and teens are already using AI tools for homework, creativity, and socializing — and what parents need to know about it.

Your guide to saving 5 to 10 hours per week with AI — concrete workflows for knowledge workers, not tool demos.

Track your AI time savings, build lasting habits, and continuously optimize your AI productivity workflow.

A concrete daily workflow for starting your day with AI — email triage, meeting prep, and task planning in under 30 minutes.

Use AI to research faster, analyze better, and make more informed decisions — practical workflows for knowledge workers.

Cut your writing time in half — AI-assisted emails, reports, presentations, and messages with prompts that actually work.

Cut through the hype — what AI can actually do today, which tools matter, and what you can safely ignore as a beginner.

Your starting point for learning AI — what to expect, how this playbook works, and why now is the perfect time to get started.

Five concrete, practical ways to start using AI in your daily life and work — starting today.

Your personalized guide to continuing your AI education — choose your path based on your interests, role, and goals.

A concrete, actionable AI safety checklist — settings to change, habits to build, and red flags to watch for.

Your practical guide to using AI safely and protecting your privacy — no policy jargon, just clear actions you can take today.

How to spot AI-generated misinformation and deepfakes — practical detection techniques and verification habits.