Read the latest
Friday, August 28, 2026HotTea archive editionVerified 2:33 AM PDT

6 minutes. Facts before narrative.

Reuters said Nvidia agreed to buy Hugging Face

Reuters put Nvidia's reported deal for Hugging Face at $12.9 billion. More than 100 organizations asked for stronger cyber defenses against AI-assisted attacks. The White House restricted some foreign grid equipment. Anthropic discussed a MatX deal. A surgeon used AI during live brain surgery. Bill Gates wants AI oversight across borders.

Start the briefingSee how we verify →

Published daily by 3:00 AM Pacific. No forced optimism. No manufactured panic.

Listen to today’s briefing

Reuters said Nvidia agreed to buy Hugging Face

The sourced HotTea edition, condensed into a chaptered morning podcast with verified audio and a full transcript.

Follow on Apple PodcastsFollow on SpotifyPodcast home →
01

The lead

Nvidia, chips, geopolitics, and open modelsNvidia may be trying to buy the model hub where many developers find and share open models.

Reuters says Nvidia's reported Hugging Face deal is worth $12.9 billion

Reuters said Nvidia agreed to buy Hugging Face for $12.9 billion, citing The Information and a person with knowledge of the deal. CNBC said another source confirmed talks. Nvidia and Hugging Face have not publicly confirmed a deal.

What happened

The reported price is nearly three times Hugging Face's $4.5 billion valuation from its 2023 funding round. Hugging Face hosts models, datasets, and tools for many open-model developers. Nvidia already sells much of the hardware those developers use to train and run models.

Why it matters

A deal would give Nvidia a place inside the hub where many developers find, test, and ship models. Nvidia would still sell much of the hardware those developers use. That puts Nvidia closer to both the model workflow and the chips, while large AI labs try to work around its hardware.

What to watch

Watch for company confirmation, a regulatory filing, and any promise that Hugging Face will stay neutral across chip vendors. If a deal closes, hosting terms, rankings, and deployment defaults will show how much control Nvidia gets.

The caveat

The report still rests on unnamed sources. Reuters said it could not verify The Information's account on its own. CNBC said talks were ongoing. Other reporting said a signed agreement was not certain. The companies have not published terms or a regulatory filing.

Reuters: Nvidia agrees to buy Hugging Face for $12.9 billion, The Information reports CNBC: Nvidia reportedly agrees to buy Hugging Face for $12.9 billion
Read this story on its own →

Worth knowing

The rest of the morning

Facts, pressure point, next evidence.

02
AI security, infrastructure, regulation, and modelsModel builders want governments and customers to protect the hospitals, water systems, and other targets that AI-assisted attackers could hit.

More than 100 organizations asked for defenses against AI-assisted attacks

OpenAI, Anthropic, Microsoft, Google, Amazon, chipmakers, security companies, banks, and other organizations signed a letter on August 27. CNBC counted 116 signatories. The letter asked companies to upgrade security, share verified fixes, and give hospitals, water systems, and other critical-infrastructure defenders AI tools and support.

Pressure point The letter does not include binding funding, delivery dates, or enforcement. Many signatories also build or sell the models behind the warning. The industry says attacks will spread in the coming months. Observed data does not show that yet.

Watch Watch for funded programs for hospitals, water utilities, and local governments. Delivery, access rules, and incident data matter more than another warning letter.

OpenAI ↗CNBC ↗
Read article →
03
Energy security, supply chains, and regulationAI power demand is making grid hardware a security fight.

The White House declared an emergency over foreign equipment in the U.S. bulk-power system

President Donald Trump signed an order on August 26. The order restricts purchases, imports, and installation of some foreign-made equipment and related software in the bulk-power system. It covers equipment judged to create sabotage, remote-access, supply-chain, or operational risks. It excludes local distribution facilities.

Pressure point The Energy Department still has to name the covered countries, companies, and equipment. Grid operators already wait a long time for transformers and other hardware. A broad rule could reduce one security risk. It could also slow reliability work and construction.

Watch Watch the Energy Department's rules within 120 days. The covered equipment list, installed hardware, waivers, grid reliability, and data-center construction will matter most.

The White House ↗The Business Times ↗
Read article →
04
AI chips, markets, and model economicsAnthropic discussed spending billions on a chip startup as it builds around Claude.

Anthropic discussed buying MatX for $7 billion, then moved toward partnership talks

Reuters reported that Anthropic discussed buying AI chip startup MatX for about $7 billion. The acquisition talks are no longer active. One source said the talks became partnership talks. Anthropic said it is expanding an internal chip team and keeping relationships with Nvidia, Google, and other suppliers.

Pressure point Reuters based the acquisition account on unnamed sources. Anthropic declined to discuss the deal talks, and MatX did not respond. A working chip can take years to design. Manufacturing and software costs come later. Those costs can run into hundreds of millions of dollars.

Watch Watch for a formal MatX partnership, new Anthropic hardware hires, and a disclosed training or inference chip plan. Useful proof would include working silicon, software support, cost per workload, and volume production.

Reuters ↗
Read article →
05
AI in medicine, products, regulation, and societyAI moved from reviewing scans before surgery to marking risky structures during an operation.

A London neurosurgeon used AI during live brain surgery

University College London Hospitals said a surgeon used an in-house AI system during a pituitary-tumor operation. The system analyzed live endoscopic video and marked nearby nerves, blood vessels, and other risky structures. The surgeon stayed in control, removed the tumor, and protected the patient's sight.

Pressure point This is one patient in a clinical trial. It does not show that the system improves outcomes across hospitals or surgeons. UCLH developed the system and announced the result. The operation shows that the system can work in surgery. Larger studies still need to measure safety, error rates, and benefit.

Watch Watch the trial size, published error rates, missed structures, false warnings, and comparisons with surgery without the system. Regulators will also need evidence that the model stays reliable across equipment, hospitals, and patient anatomy.

University College London Hospitals ↗The Guardian ↗
Read article →
06
AI politics, regulation, labor, economics, security, and societyBill Gates says national rules do not cover the AI risks that worry him now.

Bill Gates wants national and international systems to manage AI risk

Bill Gates proposed a global AI organization that borrows parts of nuclear inspections, international aviation rules, and ozone agreements. He told Reuters that he wants to discuss model-release limits and biological-risk monitoring with Chinese President Xi Jinping. Gates also raised job loss, inequality, and a possible tax on AI output when companies use it to replace labor.

Pressure point This is a proposal from a technology investor and philanthropist with financial and institutional ties to AI companies. No government has adopted the plan. U.S.-China cooperation, inspection rights, enforcement, and the technical line for restricting a model are still unresolved.

Watch Watch whether Gates meets Xi, whether the United States proposes a specific forum, and whether any plan defines inspection powers and release thresholds. Labor policy will need observed displacement data, not only forecasts.

Gates Notes ↗NBC News and Reuters ↗
Read article →

The whole AI power map

AI is no longer a tech beat.

HotTea follows where AI moves power, money, labor, security, and state capacity—not only where a new model scores higher.

01

Politics & regulation

Elections, procurement, courts, surveillance, lobbying, and state power.

02

Economics & labor

Productivity, wages, employment, capital spending, concentration, and who captures the gains.

03

War & security

Autonomy, cyber operations, intelligence, targeting, export controls, and escalation risk.

04

AI geopolitics

Chips, energy, alliances, sovereign capability, supply chains, and strategic competition.

05

Markets & companies

Funding, revenue, margins, model economics, enterprise adoption, and infrastructure bets.

06

Science & society

Medicine, education, climate, culture, research, rights, and measurable public outcomes.

AI control, infrastructure, security, medicine, and regulation

AI companies are fighting for control around the models

Nvidia may take over a model hub. Anthropic looked at bringing chips closer to Claude. Governments are restricting grid equipment. Hospitals are testing live AI support. Companies and policy advocates are asking for wider oversight. The through line is control over the systems around AI models.

1

A Hugging Face deal would give Nvidia influence over AI hardware and a major model hub.

2

Anthropic's MatX talks put chip design inside AI strategy. The White House grid order treats power equipment as part of the same control fight.

3

The cyber letter pushes AI defenses toward hospitals, water systems, and other critical infrastructure. The brain-surgery trial puts AI advice inside an operation, with the surgeon still in control.

The watchlist

Signals that could change the read

WatchlistPublic confirmation, deal terms, and neutrality promises for any Nvidia-Hugging Face deal.Tracking
WatchlistFunded cyber-defense programs and measured delivery from the 116 organizations that signed the letter.Tracking
WatchlistThe Energy Department's covered-equipment rules and Anthropic's first proof of a working custom chip.Tracking
WatchlistPublished UCLH trial results and a specific international AI oversight proposal.Tracking

How HotTea works

No optimism quota. No negativity quota. Just the sourced read.

Every reported item links to its source. Company claims remain company claims. High-risk stories require stronger corroboration. Material caveats, conflicts, and unknowns stay in the story. HotTea’s interpretation is visibly separated so readers can disagree without losing the facts.

Edition validated · 6 stories · 11 unique sources

Audit today’s sources →

Tomorrow’s signal, before tomorrow’s noise

Open HotTea. Know what changed.

A new verified edition every morning. If the evidence or release gate fails, the last verified briefing stays live.

Back to today’s top ↑