Google rents SpaceX AI compute as Musk pitches orbital data centers ahead of a historic IPO

Google will pay SpaceX about $920 million a month from October 2026 through June 2029 — roughly $30 billion in total — for access to around 110,000 Nvidia chips, capacity SpaceX first built for its own xAI division. Google Cloud called it short-term "bridge capacity" for its Gemini Enterprise platform while its own data centers scale up. The deal lands days before SpaceX's IPO, where Google already holds about a 5 percent stake, and echoes an earlier SpaceX compute arrangement with Anthropic worth roughly $1.25 billion a month.

At a JPMorgan roadshow hosted by Jamie Dimon on June 4, Elon Musk made space-based AI data centers the centerpiece of his case for SpaceX's roughly $1.5 trillion IPO, which lists this week. Launching solar-powered compute into orbit, he argued, could become "the primary means by which AI can be expanded," sidestepping the power and permitting bottlenecks that limit data centers on the ground. SpaceX, which absorbed Musk's xAI in February, plans to spend about $12.7 billion on AI this year — though critics called the pitch long on vision and short on operational detail.

OpenAI rolled out Lockdown Mode, an opt-in setting that walls ChatGPT off from the web and other external data to blunt prompt-injection attacks, where hidden instructions in a webpage or uploaded file hijack the model. Turning it on disables live web access, Deep Research, Agent Mode, image responses, connectors, and file downloads. It is aimed at people and organizations that handle sensitive data, and ships alongside new "Elevated Risk" labels — though OpenAI warns even Lockdown Mode cannot fully eliminate injection risk.

Separately, OpenAI is planning its biggest ChatGPT redesign yet, reshaping the app into a "superapp" that pushes users toward its Codex coding tools, image generation, and partner services like Canva and Booking.com, the Financial Times reports. The revamp, rolling out in the coming weeks, is part of a broader pivot toward higher-margin enterprise revenue ahead of a possible IPO. Business customers already drive about 40 percent of OpenAI's revenue, a share the company expects to reach half by year-end.

Clive Chan, by his own account the second engineer hired into OpenAI's custom-chip program, said he has left to join Anthropic. Chan, who previously worked on Tesla's Autopilot chip and the OpenAI and Broadcom silicon partnership, signals Anthropic's deepening interest in designing its own AI hardware as both labs barrel toward IPOs. Reuters reported in April that Anthropic was weighing a custom-chip effort of its own, though the plans were still early and without a dedicated team.

Meta is weighing a price of up to $200 a month for Hatch, a planned AI agent that builds software tools and handles chores like scheduling and email from plain-language prompts, according to The Information. A consumer version of Meta's OpenClaw agent, Hatch would be the company's first paid AI product — a sharp turn for a firm that built its AI reputation on free, open-weight releases. A broader US launch with free and premium "Hatch Plus" tiers is expected in July.

Perplexity launched Search as Code, an architecture that lets AI agents query its search stack by writing Python instead of calling fixed APIs. Agents can orchestrate primitives — parallel fan-out, deduplication, filtering, and ranking — before results ever reach the model's context, which Perplexity says cuts latency and delivers about 2.5-times the performance of standard tool-calling on complex retrieval. It is now the default in Perplexity's Computer product and is available through the company's Agent API.