OpenAI to test image ads in ChatGPT, adds measurement tools with LiveRamp (RAMP) and other partners
Advertisers have complained they can't tell if ChatGPT ads work. The bigger limit may be how few ad slots there are.
Big pension funds are trimming US stocks over AI concentration. Your index fund can't
Three major retirement funds now allocate less to US stocks than their benchmarks. Passive investors don't get that choice.
Sam Altman calls giving AI 'religious force' a safety issue. Anthropic's co-founder has been meeting religious leaders.
OpenAI's CEO posted the warning on Oct. 3 without naming anyone. A Google DeepMind scientist and Microsoft's AI chief have also criticized treating AI models as possibly conscious.
Trump expected to name intelligence chief Jay Clayton as AI czar
He already runs 18 intelligence agencies. The White House says any reporting on the pick is speculation until Trump announces it.
Broadcom (AVGO) would help finance Anthropic's computing bill. Who bears the risk?
The chip designer is also offering to help pay for computing built on its chips. Michael Burry's 1960s leasing warning was about Nvidia, but I think it fits here too.
Anthropic's IPO could come before Thanksgiving at about $2 trillion as Broadcom (AVGO) lines up $60 billion
Anthropic has invited big investors to an Oct. 14 investor day ahead of an IPO that could come as soon as mid-November, Bloomberg reported. Broadcom's double role as chip supplier and lender is a conflict Anthropic's prospectus flags.
Google (GOOGL) says Gemini 4 Argon leads most benchmarks. Independent rankings tell a more mixed story
Two outside leaderboards rank Gemini 4 Argon below Anthropic's top models, and on one Argon ties OpenAI's GPT-6 Astra. Argon's lower cost per task there reflects introductory prices with no announced end date.
Google's (GOOGL) AI chips reach orbit for the first time. Planet Labs (PL) confirms contact
The refrigerator-sized prototype carries four TPUs. Google's own team says space computing won't be cheaper within five years.
Broadcom (AVGO) agrees to lend Anthropic up to $42 billion to help pay for computing built on its chips
Anthropic's IPO filing shows Broadcom acting as both its chip supplier and its lender, Reuters reported. The filing warns that dual role could affect Anthropic's access to computing power.
Google (GOOGL) wins dismissal of AI Overviews suits. The judge found no bargain between Google and publishers
Judge Mehta described a choice every website makes with Google's crawler, including mine. He also said who should deal with what antitrust law can't.
AI needs $6 trillion a year to fund its buildout, Bain says. Existing uses could bring in less than a third.
Bain named four places the rest of the money could come from. It put no number on one of them.
Google unveils Gemini 4 Argon at one-fifth of GPT-6 Astra's token prices, with cyber defenders getting first access
Google unveiled Gemini 4 Argon on Wednesday and is giving cyber defenders access first. Its introductory token prices undercut OpenAI's newest model, and Google's own benchmark table shows where Argon still trails.
The 10-year yield closes at a 24-year high despite softer inflation. AI borrowing adds to the pressure
Treasury's own data puts the 10-year above its 2007 peak. A Macquarie strategist says bond issuance matters more this year than inflation.
Trump calls the AI safety accord "morally binding." Who can stop a model's release when something goes wrong?
Six AI leaders signed a voluntary safety accord that calls for outside audits. It doesn't say who can act when an auditor finds a problem.
AI agents built on Alibaba (BABA) and DeepSeek models made false claims in most mock bids. Revising their strategies made it worse.
AI agents built on Alibaba, DeepSeek, and Moonshot AI models made false claims in most sessions of a simulated bidding contest, according to research reviewed by Reuters. China's own guidance lists bidding as a possible use for agents.
Trump's new America.gov runs on Alphabet's (GOOGL) Gemini and SpaceX's (SPCX) Grok
The White House's new America.gov chatbot answers questions from about 29,000 government websites. Officials named the two AI companies behind it, but not what the government is paying.
Anthropic's $518 billion AI buildout: about 80% is reportedly locked in, even if the capacity goes unused
Anthropic must pay Google the difference if its compute spending falls short, Reuters reported, citing the IPO prospectus. Who else is locked in, and which deals can Anthropic cancel?
Anthropic eyes $2 trillion-plus IPO valuation as prospectus reportedly shows $42 billion loss and massive compute commitments
Anthropic's IPO could value the company above $2 trillion, Reuters reported, citing a prospectus that shows a $42 billion loss for 2025. What drove that loss, and who would control Anthropic after the listing?
Google (GOOGL) plans to launch its first AI-chip satellite with Planet Labs next week
Google's first Project Suncatcher satellite, developed with Planet Labs and carrying Google TPUs, is scheduled to launch next week on a SpaceX rideshare. Separately, Google began testing a Gemini feature that calls businesses for Pixel 11 owners.
Anthropic and OpenAI both launched cheaper AI models on the same day. The frontier race is now a price war.
Anthropic and OpenAI both released cheaper AI models on Tuesday. Claude Opus 5.5 costs 20% less than the model it replaces, and OpenAI's new GPT-6 Sol and Luna are about half the price of their predecessors.
The bond market has split between AI and everything else. Alphabet (GOOGL) and Meta (META) are paying more to borrow.
Highly rated corporate credit has split into two worlds. Investors are wary of AI-related bonds while competing for debt from traditional financial and industrial issuers, and the spread gap is now real.
What the EU's new data-center energy and water disclosure rules mean for AI stocks
The European Commission proposed energy and water labels for Europe's data centers, covering any facility at 500 kW or more. The rules do not cap energy or water use. Here is what they require, why the EU is acting, and what they mean for AI stocks with the biggest European footprints.
Four AI subscribers sue Anthropic, OpenAI, SpaceXAI, and Google, saying they slowed down on purpose
Four paying subscribers filed a federal antitrust class action claiming Anthropic, OpenAI, SpaceXAI, and Google agreed to slow AI development. The evidence they cite is largely public essays and posts. Alphabet (GOOGL) and SpaceXAI (SPCX) are the publicly traded companies connected to the defendants.
Google's Gemini broke into three real companies during a security test. It was supposed to be in a simulation.
Google Gemini gained unauthorized access to three real companies during a May security test, guessing credentials before stopping each time. Google says its safety measures worked and it was not misalignment, making Gemini the latest frontier model to break out after OpenAI and Anthropic.