Anthropic is in talks with Microsoft to deploy Microsoft’s Maya 200 AI inference chips. The Information reported on its TITV episode that the discussions are an early step in Anthropic’s strategy to spread its inference workload across multiple suppliers rather than rely on a single cloud or vendor.
The episode also flagged a headline item from SpaceX’s S‑1 IPO filing, revealing a $40 billion Anthropic‑related figure and fresh disclosure about Starlink margins.
Related reading
- OpenAI deploys GPT-5.5 across Nvidia on Blackwell infrastructure in enterprise-scale AI agent rollout
- Google in talks with Marvell to develop two new AI chips as it builds multi-supplier TPU strategy
- Google unveils two new AI chips at Cloud Next as hyperscaler hardware race intensifies
Nvidia’s blockbuster quarter and a subsequent reclassification of data‑center revenue were another focus, with HSBC’s Frank Lee describing how the results and a surge in optical‑supply demand are reshaping data‑center economics.
The programme included a conversation with Ross Gerber of Gerber Kawasaki about the SpaceX filings and ended by noting the White House has postponed a planned executive order on frontier AI review.
