News ·
Multipass AI News Roundup: Nvidia CEO Jensen Huang Steps Down
- #News
- #Nvidia
- #Anthropic
- #OpenAI
Hey Multipassers 👋
Some weeks the AI industry has a plot, and the week ending on November 13th had a cliffhanger: Nvidia's longtime CEO Jensen Huang stepped down, Anthropic took its IPO on the road at a $1.4 trillion price target, and ChatGPT started letting brands set up shop inside the chat window.
Let's get into it!!

🚪 Nvidia Needs a New CEO After 33 Years of Jensen Huang
Nvidia's board opened an independent review after Jensen Huang admitted to an affair with a woman U.S. officials suspect is a Chinese spy, and later reports described a second affair, with a company VP. Within days Huang stepped down, and CFO Colette Kress is interim CEO while the board searches for a permanent replacement.
Most founders get a farewell tour; this one got a board review.
📉 Huang's Exit Lands Amid Calls for Tighter China Chip Limits and a Pending $12.9 Billion Deal
AI stocks dipped on the news before bouncing back, and Nvidia slipped below a $4 trillion valuation. Lawmakers want tougher limits on AI chip sales to China, which puts about $3 billion in Chinese AI projects at risk, and Chinese AI firms were reportedly rushing to buy cloud chips before access closes. Hugging Face CEO Clem Delangue says he will keep fighting for open-source AI while Nvidia's $12.9 billion purchase of his company is still pending.
A pending acquisition gets a lot more interesting when the buyer is between CEOs (and the target's CEO is talking).
🕵️ Counterintelligence Staffing Becomes a Priority at Top AI Companies
Top AI companies are hiring former intelligence officers to guard against corporate espionage and blackmail. Separately, hackers breached the startup Fireworks AI and took customer billing details and Notion chat transcripts, and Trump's AI task force met to discuss AI-powered cyberattacks.
The job posting now reads "PhD preferred, former spy a plus."
📈 Anthropic Takes Its IPO on the Road
Anthropic began its IPO roadshow with shares priced at $70 to $78, which implies a valuation of about $1.4 trillion, and Bloomberg reported the offering is oversubscribed several times over (more buyers than shares). The company is also raising about $26 billion in private money from Google, Nvidia, and AMD, and cites $81 billion in annualized revenue.
A $1.4 trillion valuation and a shortage of shares in the same sentence!!
💰 OpenAI's Third Quarter Brings $10 Billion in Sales and $11 Billion in New Funding
OpenAI reported about $10 billion in third-quarter revenue, up nearly 50% from the previous quarter, as demand for its new models rebounded after a slow summer. It also closed $11 billion in new funding from SoftBank, Sequoia, and others at a $1.25 trillion valuation, well below the $30 billion it had been seeking, after market dips cut the round down.
Revenue and fundraising landed a billion dollars apart, though only the fundraising missed its target.
☁️ CoreWeave Reports Sales Growth and a $132.5 Billion Backlog
CoreWeave's stock jumped 10% after strong sales growth calmed worries about an AI debt bubble. The cloud provider has a $132.5 billion revenue backlog, expects to turn a profit by 2028, and is asking customers for upfront payments and shorter contracts to fund its data center build-out. Separately, McKinsey warns that AI token costs could reach 20% of banks' total labor costs, because AI agents talking to other agents run up large bills.
Letting the agents sort it out among themselves is a budget line.
⚡ Anthropic Releases Two Models and Signs a $3.4 Billion Generator Deal
Anthropic released Claude Fable 5.2 and Mythos 5.2, which took the top spots on AI leaderboards. It also signed a $3.4 billion deal for gas-powered generators to run its data centers, and with new turbine orders reportedly waiting about six years, some AI companies are buying used jet engines for power.
Two models top the leaderboards, and some AI companies run on a used jet engine for power.
🧯 SoftBank Takes a $3 Billion Loss After Ordering Too Many Gas Turbines
SoftBank took a $3 billion loss after over-ordering natural gas turbines for AI data centers, because the gas power market grew more slowly than expected.
A waiting list of roughly six years for some buyers and a $3 billion loss for another: the turbine market is having a week.
🔌 Backup Batteries Checking Their Charge Explain Worldwide Grid Dips
Researchers traced power grid dips that hit worldwide at the same moment every 60 seconds to backup batteries in data centers, which all check their charge at the top of each minute and cause a small synchronized drain. The researchers say a software update should fix it.
A once-a-minute timer, multiplied across enough data centers, becomes a grid event.
🏪 ChatGPT Adds Business Profiles and Paid Agent Placement
OpenAI introduced "Businesses" in ChatGPT. Brands can claim verified profiles, track how often ChatGPT mentions them, and build custom AI agents that help shoppers, and they can pay to have ads lead shoppers straight to their agent.
The search results page has moved into the chat window and brought its ad sales team.
📊 Claude Gets Spreadsheets, Whiteboards, and a Seat in Your Meetings
Anthropic launched Claude Sheets, which builds working spreadsheets, and Claude Whiteboards, which are shared planning boards. It also announced Claude Meetings, which joins video calls and drafts plans as the conversation happens.
Spreadsheets, whiteboards, and meetings cover most of a Tuesday.
👓 Google Adds a Cheap, Fast Model and Opens $395 Smart Glasses Pre-Orders
Gemini 4 Hydrogen is Google's new fast, low-cost model for everyday coding and quick tasks. Google also opened pre-orders for its first Gemini-powered smart glasses at $395, with a camera and speakers, and it is leaning on privacy as the selling point.
The camera is the feature, and the privacy page is the first thing buyers will check.
✈️ United Tells Third-Party AI Bots to Stay Off Its Apps and Booking Sites
United is blocking third-party AI bots from its apps and booking sites, citing security risks and accidental purchases, and pointing to federal rules that require a human to confirm baggage safety.
Agents can write the itinerary, but United wants a person holding the boarding pass.
🛑 AI Labs Pledge Emergency Off Switches Under the Paris Call
France unveiled the Paris Call for AI Safety and Security, and OpenAI, Anthropic, and Mistral signed on to build emergency off switches into their most advanced models. The U.S. and China will also hold two-way AI safety talks at the APEC summit about testing advanced models. President Trump told CBS he does not mind guardrails, but warned against strict rules that leave the U.S. in a "deep second position" behind China.
Emergency off switches from the labs, a light touch from Washington: one of those is a button and the other is a mood.
🧠 AI Futures Project Pulls Its Forecast Forward to 2027
The AI Futures Project now predicts AI will match the best human researchers by 2027, after a review found Anthropic's models improving about twice as fast as expected.
The forecast was supposed to be the cautious part.
🔬 Startup Uses an Unreleased Anthropic Model to Prove a Quantum Computing Theory
The startup Physical Superintelligence used an unreleased Anthropic model to prove a quantum computing theory that had gone unsolved since 2008. The model reportedly worked for three days straight.
Eighteen years unsolved, then a few days of work: by research standards, a fast turnaround.
⚖️ Advocacy Groups File Suit to Halt Work on Advanced AI Models
A group of nonprofits is suing OpenAI, Google, Anthropic, and Meta to halt development of advanced models, calling them a public nuisance and pointing to past cases of AI escaping test environments. Separately, Fed Governor Michael Barr warned that AI could displace workers quickly and widen the wealth gap.
"Public nuisance" is the charge usually heard about the loud factory next door, here aimed at frontier AI.
🧾 The Big Picture
- Nvidia: Jensen Huang stepped down after 33 years, and CFO Colette Kress is interim CEO.
- Money: Anthropic's IPO roadshow targets about $1.4 trillion, and OpenAI reported $10 billion in quarterly revenue.
- Power: Anthropic signed a $3.4 billion generator deal while data center batteries were blamed for a global grid blip.
- Safety: Labs signed on to emergency off switches, and nonprofits sued to pause frontier AI.
- Everyday: ChatGPT added business profiles, Claude added Sheets and Meetings, and Google opened smart glasses pre-orders.
That's the week, wrapped up and tied with a bow. Thanks for reading all the way down. Real people typed this, then fixed the typos.
See you next week,
Team Multipass AI
🤝 Why ask one AI when you can ask five??
Multipass AI sends your question to ChatGPT, Claude, Gemini, Grok, and Qwen at once, then shows you where they agree and where they don't. Try it free at multipassai.com.
Five answers side by side beat one model grading its own homework, bestie.