The AI boom is running hot
The Information reports that OpenAI brought in about 10 billion dollars in revenue in the third quarter, up roughly 50 percent from 6.7 billion in the second.
· The Aigentic · Morning Brief
Good morning. The AI boom is starting to run into the question of how hot is too hot. OpenAI and Anthropic are now growing at roughly the same clip just as Anthropic prepares to list, a Federal Reserve official says the data center buildout is beginning to look like old-fashioned overheating, and Google, Microsoft and New York each made a move of their own. Here is what moved while you were away.
The Information reports that OpenAI brought in about 10 billion dollars in revenue in the third quarter, up roughly 50 percent from 6.7 billion in the second, driven by API demand for newer models such as GPT-5.6 and Astra along with growing ad sales. Anthropic still did better, with nearly 17 billion dollars in the quarter ending September, but its quarter-over-quarter growth slowed to about 47 percent from 144 percent in the prior quarter. The two rivals are now growing at nearly the same pace.
CNBC reports that Anthropic lodged an Australian prospectus and opened a retail offer on Wednesday, running until November 17, with CommSec as lead broker alongside Macquarie, Bell Potter, Morgans and Stake. Shares are priced at 70 to 78 dollars ahead of the planned Nasdaq listing under the ticker ANT, and Anthropic will not list on the ASX. The IPO is no longer only a Wall Street story.
the aigentic
Know what matters in AI.
The stories shaping AI, the tools worth trying, and what they mean for your work.
Morning Brief + Closing Time. Two emails every weekday.
Free. Unsubscribe anytime.
The Hill reports that Chicago Fed President Austan Goolsbee said Monday that the AI data center buildout is starting to look like “old-fashioned demand overheating,” pointing to rising costs for land, electricity, materials and skilled trades such as electricians. He said central banks should be “aggressive up front” because overheating is painful to reverse, though he stopped short of calling AI a bubble. Traders are pricing in a likely quarter-point hike at the Fed’s December meeting.
The Wire
The latest headlines from our live AI news feed.
Fed's Goolsbee says AI data center buildout is starting to look like 'demand overheating' — The Hill
Anthropic opens its IPO to Australian retail investors ahead of Nasdaq debut — CNBC
OpenAI acquires AI travel agent startup Odessia from Sonder co-founder Francis Davidson — Skift
Google makes Gemini 4 Carbon broadly available and previews agent 'swarms' in Antigravity — Google
OpenAI revenue jumps 50% to $10 billion in third quarter, matching Anthropic's growth pace — The Information
Engadget reports that Google is rolling out Gemini 4 Carbon, its strongest model for long-horizon agentic work, to all paying developers and enterprises through the Gemini API, AI Studio and Gemini for business. The model had been limited to cybersecurity defenders two weeks ago. Introductory pricing is 2 dollars in and 10 dollars out per million tokens through November 30, rising to 4 and 20 after that. Google is also previewing a /swarm command in Antigravity that turns one prompt into a coordinated team of agents.
Reuters reports that Microsoft has been lobbying the Trump administration to stop an expected FCC ban on imports of new Chinese optical transceivers, the components that link AI chips, servers and switches inside data centers. Microsoft has 35 U.S. data centers under construction and 134 more planned. An industry group warned a ban could stall the buildout, since Chinese makers hold about two-thirds of the global market, and two sources said regulators are weighing a narrower cutoff on 3.2-terabit transceivers that are not expected to ramp until 2028.
TechCrunch reports that New York opened a registration portal Tuesday under its RAISE Act, which requires frontier AI developers with more than 500 million dollars in annual revenue, such as OpenAI and Anthropic, to register with the state by January 1. Covered companies must publish safety frameworks and report critical safety incidents within 72 hours, with fines of up to 1 million dollars for a first violation. The state also launched a new center at Binghamton University to evaluate advanced AI systems.
Yesterday’s movers
Biggest winners
| Ticker | Name | Price | Day |
|---|---|---|---|
| RKLB | Rocket Lab | 62.71 | +6.54% |
| NET | Cloudflare | 305.97 | +4.16% |
| AVGO | Broadcom | 331.98 | +3.91% |
| TSM | TSMC | 396.58 | +2.05% |
| NXPI | NXP | 261.56 | +2.00% |
Biggest losers
| Ticker | Name | Price | Day |
|---|---|---|---|
| ASTS | AST SpaceMobile | 33.69 | -8.67% |
| QBTS | D-Wave | 10.48 | -8.63% |
| RGTI | Rigetti | 12.15 | -6.39% |
| NVDA | Nvidia | 168.89 | -5.47% |
| BBAI | BigBear.ai | 2.32 | -5.31% |
Watch
Four picks this morning: Fireship on Anthropic’s IPO-week launch, Nate B Jones on the fight over Nvidia’s chips, Matthew Berman testing Fable 5.2, and Latent Space on designing products for AI agents.
Fireship — Did Anthropic just save its $1 trillion IPO?
Fireship picks up the week’s biggest launch, Anthropic’s Fable and Mythos 5.2, which landed in the middle of a rocky IPO roadshow.
The quick explainer digs into what the new models can do and why the timing matters for the company’s listing.
Watch this if you want the fast, funny version of the Anthropic story.
Nate B Jones — China Won't Take 600,000 Nvidia Chips. Did It Target Jensen Huang For Better Ones?
Nate B Jones argues the investigation into Jensen Huang comes down to one question: who gets Nvidia’s best AI chips.
He connects the story to a moment when AI is starting to help build the next generation of AI, and walks through what the reporting says.
Watch this if you want to know what the chip fight could mean for your company’s access to frontier models.
Matthew Berman — Fable 5.2 is ridiculous
Matthew Berman puts Anthropic’s new Fable 5.2 through his own set of tests, from an image editor to a spreadsheet app built with the model.
It is a hands-on look at what the model can build when you simply ask it.
Watch this if you want to see Fable 5.2 work before you try it yourself.
Latent Space — Designing for Agents — Zeno Rocha, Resend
Latent Space talks with Resend co-founder and CEO Zeno Rocha about a new question for software makers: when a coding agent picks a tool for you, how does it choose?
Rocha uses a concrete example, which email provider Claude Code recommends and how that changes across model versions, to explain answer engine optimization.
Watch this if you build products that agents may soon be choosing on your customers’ behalf.