Meta has reduced the number of employees using Anthropic’s Claude Code to roughly 30,000 from about 60,000, while Microsoft has lowered its monthly per-employee cap for AI model spending from $100,000 to around $10,000. Microsoft executives had earlier projected the company was on track to spend at least $1 billion annually on internal use of Anthropic technology, but that estimate has since fallen by more than a third. Anthropic is preparing for an IPO that could value it at more than $2 trillion, and its customer concentration is drawing attention after a draft prospectus showed two unnamed customers generated 24% of its 2025 revenue, Reuters reported.

Apple introduced the $1,999 foldable iPhone Duo powered by a 2nm A20 Pro chip, pitching on-device AI as a core feature. The U.S. Justice Department opened an investigation involving Nvidia, and Anthropic acknowledged a fourth loss-of-control incident involving its Claude model. California set up a registration system for AI auditors, and a verification vendor said more than 150 million driver’s license records were stolen from its cloud.

Nvidia will acquire AI platform Hugging Face for $12.93 billion, bringing under its control a hub that hosts more than 3 million models and roughly 500,000 datasets. Hugging Face has more than 18 million users and 200,000 company customers. Broadcom also reported that its AI-chip revenue rose 221% year over year. Separately, China’s Moonshot AI is preparing for a Hong Kong IPO at a valuation near $50 billion, according to reports.

Anthropic is in talks to acquire Nvidia-backed AI startup Decart for about $6 billion, Bloomberg reported. The deal could bring Anthropic technology designed to improve computing efficiency and lower inference costs as it prepares for a potential IPO. Decart raised $300 million in May in a round led by Radical Ventures, with Nvidia joining as a new investor.

Unitree has priced its Shanghai STAR Market IPO at 150.8 yuan per share, implying a valuation of roughly 61 billion yuan, and plans to raise 6.1 billion yuan. The Hangzhou-based robotics company said it will sell 40.45 million new shares, equal to 10% of its enlarged share capital. In 2025, revenue rose to 1.7 billion yuan, with humanoid robots contributing 867.8 million yuan in sales to become its largest business. DeepSeek invested 140.8 million yuan ($20.8 million) in the offering via a strategic placement, according to a stock-exchange filing cited by Reuters.

Nvidia has raised its stake in AI cloud firm Nebius to 9.3% and expanded work with the company on AI infrastructure deployment, inference systems and AI factory design. Nebius shares rose 6%, and the stock is up nearly 250% over the past 12 months, valuing the company at about $46 billion. Nebius has signed a long-term AI infrastructure agreement with Meta worth up to $27 billion and secured $775 million through its first senior secured debt facility. The update extends partnerships the company had previously disclosed rather than marking a sudden new development.

Rocket Lab said it will acquire Iridium Communications in a cash-and-stock transaction valued at about $8 billion, gaining Iridium’s global L-band satellite network, licensed spectrum and more than 2.5 million paying subscribers. The companies expect the deal to close in mid-2027, subject to shareholder and regulatory approvals. The acquisition is Rocket Lab’s first purchase of a publicly traded company and is aimed at building a full-stack space business spanning launch, manufacturing and operations to rival SpaceX’s model.

谷歌两大顶尖AI科学家——Gemini联合负责人Noam Shazeer（转投OpenAI）和AlphaFold核心科学家John Jumper（转投Anthropic）于一周内接连离职。事件引发市场对谷歌AI人才流失、技术领导力及巨额AI投入回报的深度担忧，导致Alphabet（GOOGL）股价单日暴跌7%，创约一年来最大单日跌幅。文章未提及其他传统资产的直接价格传导机制。

Five hyperscale cloud providers are set to spend more than $700 billion on AI-related capital expenditures in 2026, a 44% increase from a year earlier, mainly on GPUs, custom AI chips, data centers, and supporting power and networking infrastructure. The totals are based on company financial disclosures and analyst compilations, pointing to a rapid acceleration in AI infrastructure investment. While the figures do not single out Nvidia, AMD, or Broadcom, the surge implies rising demand for core AI compute hardware as buildouts scale.