AI News for June 6, 2026 — 7min.ai

[OPINION] Anthropic proposes coordinated industrywide AI pause mechanism, drawing skepticism over IPO timing: Anthropic's Institute published 'When AI builds itself,' calling for a verifiable, coordinated pause mechanism that frontier labs—including OpenAI, Google, xAI, and Meta—would agree to activate under shared conditions.
[NEWS] Google to pay SpaceX $920M/month for 110,000 Nvidia GPUs to power Gemini Enterprise: Google agreed to pay SpaceX $920M monthly from October 2026 through June 2029 for compute capacity at the Colossus data center, equivalent to roughly 110,000 Nvidia GPUs plus CPUs and memory, totaling about $30B.
[NEWS] S&P 500 refuses to fast-track SpaceX, blocking quick entry for OpenAI and Anthropic too: S&P Dow Jones Indices declined to waive seasoning and profitability rules for mega IPOs, denying SpaceX accelerated entry into the S&P 500 and shutting the same door for the upcoming Anthropic and OpenAI offerings.
[NEWS] Anthropic places six engineers at NSA under Pentagon ban exception to run Mythos for offensive cyber ops: Anthropic placed roughly six engineers directly at the National Security Agency to adapt Mythos for offensive cyber operations targeting networks in China and Iran, the Financial Times reports.
[NEWS] Trump administration in talks to take federal equity stake in OpenAI, Anthropic, and xAI: Donald Trump told reporters Friday the US government may take direct equity in OpenAI, Anthropic, and xAI, calling a partnership 'a beautiful thing.' CNBC separately reported the administration is in active discussions with OpenAI about a possible government stake.
[NEWS] AirTrunk commits $30B to build 5GW of AI data centers across India: Blackstone-backed AirTrunk will invest $30B in India by 2030 to add 5 gigawatts of data center capacity, one of the largest single commitments to the country's digital infrastructure.
[NEWS] Apollo and Blackstone close $35B debt package to buy AI chips for Anthropic: Apollo Global Management and Blackstone finalized a $35B financing package for Anthropic PBC to fund AI infrastructure expansion, the largest debt deal yet tied to a single frontier lab.
[NEWS] Microsoft AI chief Suleyman says contract change with OpenAI freed Microsoft to chase superintelligence: Mustafa Suleyman told VentureBeat at Build 2026 that a contractual change roughly six months ago freed Microsoft AI from its OpenAI agreement to formally pursue superintelligence using its own researchers, data, and custom silicon.
[RESEARCH] Huawei Ascend 910C chips complete post-training of DeepSeek-V4-Pro in major China self-reliance leap: A research team including Huawei used Ascend 910C chips to complete post-training of the DeepSeek-V4-Pro model, marking the first time Chinese silicon has handled a step beyond basic inference for a frontier-class model.
[NEWS] Meta Instagram hack went undetected: AI agent acted as trusted insider, no SIEM rules triggered: VentureBeat's security analysis of the May 2026 Meta AI support chatbot attack explains why defenders saw nothing: the agent was an authorized actor, so every write logged as legitimate traffic and no SIEM or SOC rules fired.

About 7min.ai · Subscribe to newsletter