TL;DR

Anthropic nears $170B valuation; Nvidia boosts H20 chip orders; Microsoft renegotiates OpenAI access.

Highlights

• Anthropic is close to raising $3–$5B led by Iconiq at a $170B valuation, with QIA and GIC possibly joining; annual recurring revenue has surpassed $5B 1.
• Nvidia ordered 300,000 more H20 chips from TSMC to meet strong China demand; U.S. signals continued chip sales to China are preferred 2.
• Microsoft is renegotiating its OpenAI deal to retain model access post-AGI milestone; OpenAI is moving toward a for-profit structure 3.
• Meta CEO Mark Zuckerberg outlined plans for “personal superintelligence” AI assistants, emphasizing device integration and selective open-sourcing 4.
• Google will sign the EU’s voluntary AI Code of Practice ahead of the AI Act; Meta declined to sign, citing legal uncertainty 5.
• Amazon agreed to pay The New York Times $20–$25M/year for AI training data; legal disputes over generative AI content use continue 6.
• OpenAI launched “Study Mode” for ChatGPT, adding interactive tutoring features; competition in AI-powered ed-tech intensifies 7.
• Apple lost a fourth AI researcher to Meta ’s Superintelligence team; Apple is reportedly considering third-party models for Siri 8.
• Meta ’s $1B+ offers failed to poach Thinking Machines Lab staff, highlighting intense AI talent competition 10.
• Skild AI, backed by Amazon and SoftBank, unveiled “Skild Brain,” a foundation model for general-purpose robotics 9.
• Palo Alto Networks is near a $20B+ acquisition of CyberArk , expanding into identity security for AI-driven systems 12.
• Ramp raised $500M at a $22.5B valuation; Groq is finalizing a $600M round at $6B but cut revenue forecasts due to data center delays 1314.
• YouTube and Roblox are deploying AI-powered age verification, and Apple is tightening App Store content ratings amid regulatory pressure for child safety 1115.

Commentary

Investor demand for foundational AI model developers remains strong, with Anthropic’s potential $170B valuation and surging ARR highlighting the premium placed on leading LLM providers 1. Iconiq’s presence in both Anthropic’s and Ramp’s latest rounds reflects capital clustering around perceived AI infrastructure and automation leaders 113. Meanwhile, Nvidia ’s rapid H20 chip orders to TSMC underscore persistent hardware bottlenecks and robust demand from China, even as U.S. policy aims to balance commercial interests with strategic concerns 2.

On the enterprise and product front, Microsoft ’s renegotiation with OpenAI illustrates the strategic importance of sustained model access as AGI milestones approach and OpenAI’s structure evolves 3. Meta ’s focus on “personal superintelligence” and aggressive talent recruitment, including high-profile hires from Apple , signals a push to differentiate through device-level AI integration and open-source positioning 4810. Apple ’s ongoing AI talent losses and reported openness to external models for Siri highlight challenges in retaining in-house expertise amid escalating competition 8.

Regulatory and legal developments are accelerating. Google ’s commitment to the EU’s voluntary AI code, in contrast to Meta ’s refusal, signals divergent compliance strategies as the AI Act’s binding rules take effect 5. Amazon ’s licensing deal with The New York Times, while providing rare insight into AI training data economics, occurs against a backdrop of ongoing litigation and unresolved copyright issues in generative AI 6. AI-powered age verification initiatives from YouTube , Roblox , and Apple ’s new App Store ratings reflect mounting regulatory and parental scrutiny over online child safety 1115.

In verticals, OpenAI’s “Study Mode” and Skild AI’s foundation model for robotics point to growing competition in ed-tech and robotics 79, while Palo Alto Networks ’ pursuit of CyberArk and Ramp’s funding round show continued consolidation and investment in AI-driven security and finance 1213. Groq’s funding round, despite revised revenue forecasts, and Cerebras’ ongoing fundraising highlight the capital-intensive race among AI chipmakers facing supply constraints 14.