⚡ Uber and UK startup Wayve launch London's first commercial robotaxi service, beating Waymo to one of the world's most symbolic markets. 💰
Happy Thursday. Google dropped Gemini 3.8 Flash — a dual-variant model optimised for agentic tasks and cyber vulnerability hunting — as the model arms race keeps accelerating. Meanwhile, Broadcom raised its AI chip forecast on surging Big Tech spend, and the Trump administration's decision to back OpenAI in the NYT copyright case is reshaping the legal terrain for the entire industry. A busy, multi-front day with no single dominant story.
Autonomous vehicles made a landmark commercial move with Uber and Wayve launching London's first robotaxi service, while enterprise AI adoption is reshuffling workforce hiring priorities away from pure technologists toward domain experts.
Security vulnerabilities in AI platforms are emerging as a critical middleware concern, with stolen Claude session cookies enabling corporate account compromise; meanwhile, Meta enters the real-time transcription API market at disruptive pricing.
Google is iterating at a notably fast cadence, shipping Gemini 3.8 Flash just weeks after 3.7 Flash with purpose-built variants for agentic reasoning and offensive cyber tasks, while OpenAI's upcoming Astra continues to draw safety-community alarm over its 'recurrent depth' reasoning technique.
Chip demand signals remain bullish — Broadcom raised its AI forecast on relentless Big Tech spending — but enterprise buyers are quietly diversifying away from Nvidia, and the data centre supply chain is innovating around power and cooling.
The deal flow reflects the AI stack's full height: from model-layer talent (Anthropic hiring policy architect Matt Clifford and Stargate exec Shamez Hemani) to application-layer M&A (Palo Alto Networks' $500M Console buy, Adobe's Rilo acquisition), with India's Yotta Data Services eyeing a $1.5B IPO.
The Trump administration's intervention in the NYT vs. OpenAI copyright case is the week's most consequential legal signal — a federal endorsement of AI training practices that could set precedent industry-wide — while NYC's one-year student AI ban and Google's ad-tech ruling add to a busy regulatory day.
Enterprises put non-Nvidia chips 14 points ahead of Nvidia's next-gen GPUs on their evaluation lists — While Broadcom's raised forecast and Nvidia's Equinix deal dominate the headlines, this VB Pulse survey of 170 enterprise infrastructure decision-makers reveals a quiet but structurally important shift: buyers are actively de-risking Nvidia concentration before they're locked in. For any strategist making multi-year AI infrastructure commitments, understanding which alternative silicon is winning those evaluation slots — and why — is the most actionable data point of the week. Read →