AI News for March 3, 2026 — 7min.ai

[NEWS] Altman admits OpenAI's Pentagon deal was 'opportunistic and sloppy,' pledges contract amendments: OpenAI CEO Sam Altman acknowledged that his company's rush to sign a Pentagon contract after Anthropic's blacklisting looked "opportunistic and sloppy." He announced OpenAI is working with the Defense Department to add explicit language barring domestic surveillance and use by intelligence agencies like the NSA.
[NEWS] Cursor surpasses $2B in annualized revenue, doubling in three months: AI coding editor Cursor has crossed $2B in annualized revenue, doubling its run rate in roughly three months, according to Bloomberg.
[NEWS] Nvidia invests $4B in photonics companies Lumentum and Coherent for AI data centers: Nvidia announced $2B investments each in Lumentum and Coherent, both developing photonics technology for data centers including optical transceivers, circuit switches, and lasers.
[NEWS] Apple tasks Google with hosting new Siri's AI models as Private Cloud Compute sits 90% idle: Apple has asked Google to set up servers for the next-generation Gemini-powered Siri inside Google's data centers while meeting Apple's privacy requirements, according to The Information.
[NEWS] Claude experiences widespread outage as user surge overwhelms infrastructure: Anthropic's Claude chatbot went down for thousands of users on Monday, affecting Claude.ai, Claude Code, and the Cowork platform.
[NEWS] Senate's Wyden pledges battle over Pentagon's unprecedented Anthropic blacklisting: Congressional Democrats are vowing to contest the Trump administration's actions against Anthropic, including a prohibition on all federal government use and a ban for Pentagon contractors.
[NEWS] US considers capping Nvidia H200 chip sales at 75,000 per Chinese customer: US officials are weighing caps on the number of H200 AI accelerators Nvidia can export to any single Chinese company, further constraining the chipmaker's reentry into the market.
[NEWS] Anthropic pitched Pentagon drone swarm contest while feuding over military AI terms: Anthropic submitted a proposal for a $100M Pentagon prize challenge to produce voice-controlled autonomous drone swarming technology, according to Bloomberg.
[NEWS] Meta AI smart glasses investigation reveals workers reviewing intimate user recordings: A joint investigation by Swedish newspapers Svenska Dagbladet and Göteborgs-Posten revealed that human data workers in Kenya review recordings from Meta's Ray-Ban smart glasses, including footage of people undressing and using the toilet.
[RESEARCH] Alibaba releases Qwen3.5 small model series, with 9B outperforming OpenAI's 120B open-source model: Alibaba's Qwen team released the Qwen3.5 Small Model Series, a family of open-source models ranging from 0.8B to 9B parameters.

About 7min.ai · Subscribe to newsletter