Nvidia's first post-Jensen earnings call delivered record beats and strong guidance — remarkable given the SEC is now investigating the company over Huang's exit. Meanwhile OpenAI launched GPT-6.2 Astra and Sol, two new frontier models, and Google opened its agentic Gemini Tasks assistant to free users globally. The day's signal: AI infrastructure spending remains unshaken by leadership chaos, while the model and application layers keep accelerating.
Nvidia's first earnings call without Jensen Huang could have been a crisis moment; instead it was almost aggressively normal — revenue beat, guidance beat, stock set for a $335B options-implied swing. New board chair Harvey Jones opened by closing the door firmly on a Huang return and refusing all further questions on the topic. That calculated silence is itself a signal: the board is betting the AI infrastructure supercycle is bigger than any single founder, and so far the numbers agree. The wrinkle is the SEC probe. An investigation into whether Huang's resignation — triggered by an affair with an alleged Chinese spy — involved any undisclosed material information is not a small thing. Combined with the ongoing political sensitivity of Nvidia chip sales to China, this story has regulatory and geopolitical dimensions that earnings alone can't dispel. Smart observers should watch the SEC timeline, not the quarterly beat.
ChatGPT for Sales and Alphabet's Isomorphic drug-discovery push show AI moving deeper into high-value professional workflows, while Waymo's freeway suspension by California's DMV is a stark reminder that autonomous AI applications face serious real-world safety accountability.
Google's Gemini Tasks rollout to free users marks a meaningful shift from chat assistants to agentic platforms capable of real-world actions, while Kubernetes orchestration challenges from AI agent workload spikes are becoming an operational reality for platform teams.
OpenAI's dual GPT-6.2 launch — a professional-grade Astra and an everyday Sol — signals the company's move toward tiered model offerings with dramatically improved cost efficiency, while Microsoft quietly expanded its own first-party model lineup for coding and cybersecurity.
Nvidia's post-Huang earnings dominance proves the AI hardware supercycle is institution-proof, while new data center projects from Israel to Australia signal continued global buildout — though a Trump administration curb on China-linked data centers is quietly advancing over Big Tech objections.
A $5B raise for Alphabet's Isomorphic Labs and a $500M robotics round for Genesis AI dominate a funding day that also sees Anthropic's IPO ownership structure crystallize — together signaling that capital is flowing heavily into AI's most capital-intensive frontiers.
The SEC opening a probe into Nvidia — tied directly to Jensen Huang's departure and his alleged affair with a Chinese intelligence operative — adds a serious regulatory dimension to what had seemed like a contained corporate scandal, while the Waymo freeway ban shows AI regulators are willing to act with real force.
Google's rollout of Gemini Tasks in India signals the country is now a first-wave market for agentic AI, while Indian IT services face a structural squeeze as AI adoption erodes traditional contracts and clients push for 10–30% fee discounts on renewals.
Token sticker shock pushes enterprises from frontier labs toward open-weight AI factories — While everyone debates model capabilities, this SC26 dispatch identifies the actual production bottleneck: token economics and data sovereignty are killing more AI pilots than technical limitations. A senior strategist building an enterprise AI roadmap needs to understand why the make-vs-buy calculus is shifting toward on-prem open-weight models right now. Read →