The jury trial between Elon Musk and OpenAI entered its third week. Greg Brockman took the stand, with diary entries surfacing about alleged physical threats. Judge Yvonne Gonzalez Rogers is watching the remedies closely. The outcome could redefine how AI companies structure equity and mission obligations.

Bret Taylor's enterprise AI agent company Sierra closed a $950M round, valuing it at $15.8 billion. This is one of the largest enterprise AI raises of the year. Taylor pointed to strong ARR growth and expanding deployment across Fortune 500 customers. The round signals that enterprise AI infrastructure still commands premium valuations — even as the broader market debates an AI bubble.

Cloudflare cut approximately 20% of its workforce this week, citing restructuring around AI-native operations. In an unusual moment of candor, an exec acknowledged internally that some prior product announcements leaned into "AI washing." A cautionary signal for the wave of companies repositioning as AI plays without genuine technical depth.

A busy week on the model front. OpenAI quietly released GPT-5.5 Instant (faster, cheaper inference). xAI pushed Grok 4.3 with improved reasoning benchmarks. Apple previewed iOS 27's on-device model upgrades. AWS launched AgentCore, a new managed runtime for multi-agent workloads. The pace of releases is accelerating — infrastructure players are increasingly the winners.

Goldman Sachs reiterated AI infrastructure as a multi-decade supercycle. Michael Burry (of The Big Short fame) flagged GPU inventory buildup as a warning sign. Warren Buffett reiterated skepticism about any sector he can't fully model. Three perspectives, zero consensus. The debate around AI capex sustainability isn't going away — and this week added more fuel to both sides.