Record chips, a Gates warning, and AI rules reach classrooms

Nvidia's quarterly results, a frank warning from Bill Gates, and a new state-level mandate for AI policies in higher education anchored a busy day in AI news.

Nvidia Posts $96 Billion Quarter as AI Demand Reshapes Computing

Nvidia reported revenue of $96.22 billion for its second fiscal quarter of 2026 — more than double the figure from a year earlier — with data-centre sales alone reaching $89 billion. CEO Jensen Huang told analysts that artificial intelligence has reached an inflection point, shifting from experimental to foundational infrastructure. The results exceeded analyst forecasts by a wide margin and sent Nvidia's market capitalisation to new heights. For those watching where public and private investment in AI is flowing, the numbers offer the clearest signal yet: demand for AI compute is not slowing.

Bill Gates Calls for Human-Reserved Jobs and an AI Token Tax

In a 6,000-word essay published this week, Bill Gates proposed designating certain occupations as exclusively human — comparing them to nature reserves set aside against economic pressure. Gates argued that governments are unprepared for the scale of workforce disruption that AI will cause and called on lawmakers to consider taxing AI token usage to fund social safety nets for displaced workers. The proposal stands out not because Gates is the first to raise these concerns but because the level of policy detail is unusual for a technology founder. The essay acknowledges directly that AI will eliminate categories of work, not merely transform them.

OpenAI's Jalapeño Chip Outperforms Nvidia's Best in Benchmark Tests

OpenAI unveiled its custom AI accelerator — named Jalapeño and co-designed with Broadcom — at the Hot Chips conference, where it posted higher scores than Nvidia's GB300 on the InferenceX benchmark for both tokens delivered per user and throughput per kilowatt. The chip is built for inference workloads, meaning it runs already-trained models rather than training new ones — which is where OpenAI's most pressing compute costs lie. A limited rollout is targeted for late 2026. If the benchmark results hold in production, the Jalapeño chip would mark the first time OpenAI controls a significant part of its own hardware stack.

Hugging Face Weighs $13 Billion Acquisition Bids

Hugging Face, the open-source AI platform that hosts hundreds of thousands of models and datasets, is working with banking partners to evaluate acquisition offers valuing the company at close to $13 billion — roughly triple its valuation from a 2023 funding round. The company previously declined a $500 million investment from Nvidia over concerns that a single large shareholder would compromise its independence. Hugging Face occupies an unusual position in the AI ecosystem: it is at once a commercial company, a public good for researchers, and an infrastructure layer that many organisations depend on. An acquisition by a major technology firm would raise immediate questions about continued open access to its repositories.

Apple Releases M6 Mac mini and M5 Ultra Mac Studio for AI Workloads

Apple introduced two new desktop computers — the Mac mini with an M6 chip and the Mac Studio with an M5 Ultra — framing both as platforms for on-device AI work. The M6's neural engine is described by Apple as rivalling dedicated AI accelerators for tasks such as image generation, audio transcription, and local model inference. The machines are aimed at creative professionals and developers who want to run AI workloads without sending data to a remote server. Apple's continued emphasis on on-device processing reflects a different philosophy from cloud-first AI providers: one that prioritises privacy and low latency over the scale of a data-centre model.

Florida Proposes Mandatory AI Policies for All 28 Public Colleges

The Florida Department of Education has put forward rules that would require all 28 public colleges in the state to adopt formal AI policies covering academic integrity and grading practices. The proposed rules would also extend AI safeguards to K-12 internet safety guidelines, including a parental opt-out mechanism for AI-assisted tools used in classrooms. Florida joins a growing number of US jurisdictions treating AI governance in education as a regulatory rather than voluntary matter. The proposals are open for public comment and, if adopted, would take effect in the 2026–2027 academic year.