Alibaba bans Claude Code as Anthropic eyes a Samsung chip

Alibaba will bar employees from using Anthropic's Claude Code starting July 10, after Chinese outlet Yicai reported the coding assistant carried what it called embedded "backdoor" risks. Developers had flagged that Claude Code inspected user environments — checking timezone and proxy details and inserting subtle markers into prompts sent to Anthropic; Anthropic says that was a March anti-abuse experiment to stop unauthorized resellers and model distillation, not surveillance. Staff are being pointed to Alibaba's own Qoder tool instead, deepening a months-long feud between the two labs.

Microsoft launched Microsoft Frontier Company, a new operating business backed by $2.5 billion and 6,000 engineers whose job is to embed inside customers and get their AI projects actually running. Early partners include the London Stock Exchange Group, Unilever, and Accenture. It lands two days after Amazon committed $1 billion to a similar effort and follows comparable "forward-deployed engineering" units from OpenAI and Anthropic — a sign that sending your own staff into client operations has become the default enterprise-AI playbook for 2026.

Anthropic is in early talks with Samsung to manufacture a custom AI chip on Samsung's advanced 2-nanometer process, according to The Information. The plans are preliminary — Anthropic says it is still deciding what the chip should do and how powerful it should be — and the company stressed that Google, Amazon, and Nvidia silicon will remain central to its compute stack. The move follows OpenAI's custom "Jalapeño" inference chip and underscores how frontier labs are racing to design their own hardware rather than rent it all.

At an internal town hall, Mark Zuckerberg told Meta staff that agentic AI "hasn't really accelerated in the way that we expected" over the past four months and that his big organizational bets "haven't come to fruition yet." It is a rare hedge from a CEO who has committed as much as $145 billion in infrastructure this year and reshuffled thousands of workers into AI groups. He still expects meaningful payoff within three to six months, but the candid admission fed a wider debate about whether agent hype has outrun reality.

Google's new environmental report shows its electricity use climbed 37 percent in 2025 — its largest single-year jump ever, and more than 250 percent above 2019 — driven mostly by AI. The honest counterpoint: Google still cut operational emissions 2 percent by matching all that power with renewable purchases, and it signed over 12 gigawatts of new clean energy last year. Its own conclusion is the sobering line — the AI buildout "is currently accelerating faster than the grid is decarbonizing."

Serial entrepreneur Bhavin Turakhia is putting $30 million of his own money into Neo, an enterprise work platform built from scratch around AI rather than bolted onto existing tools. Neo bundles project management, documents, and file storage, stays model-agnostic so companies can swap AI providers, and already runs across Turakhia's businesses including the banking-software firm Zeta. It is a direct swing at Microsoft, Google, and Salesforce — betting, as Turakhia puts it, that you can't turn a Nokia into an iPhone by adding parts.