The Briefing

Thanks for reading The Briefing, our nightly column where we break down the day’s news. If you like what you see, I encourage you to subscribe to our reporting here.

Greetings!

The AI laggards are going on the offensive! Meta Platforms on Thursday unveiled its latest AI model, Muse Spark 1.1, a day after SpaceX’s AI unit introduced a new model of its own. Like SpaceX, Meta claimed its new model delivers performance comparable to that of established models from AI labs such as OpenAI, Anthropic and Google. We’ll have to wait a few days to see whether developers agree. (Meanwhile, OpenAI unveiled its latest models, GPT 5.6).

What may be most significant, however, is Meta’s apparent intention to spend a fortune subsidizing usage of its new model in hopes of pulling market share from its rivals. In an interview with Bloomberg, Meta CEO Mark Zuckerberg said Meta’s pricing for what users would pay to use the models would be “very aggressive and attractive.” Bloomberg’s report said Meta’s prices would be about a quarter of those advertised by top models from Anthropic and OpenAI—in other words, dirt cheap.

That’s not a way to make money in AI, but that clearly isn’t Meta’s strategy. It knows that the cost of AI has become a paramount issue for many businesses, so much so that they’re trying various methods to reduce it, including the use of open-source models or routing some AI work to older and more cost-effective models. In that environment, Meta presumably thinks by undercutting everyone else on price, it can persuade users to at least try out its new model—and potentially hook them. If that approach works, Meta could jack up the price later. It’s a similar approach to that pursued by firms in other industries—such as video-streaming company Netflix or ride-hailing firms such as Uber. It may well work but competition is fierce. (Check out The Information‘s TITV today for an analysis of OpenAI’s latest models).

In the meantime, though, Meta’s investment in AI seems likely to mount. Already Meta had projected laying out between $125 billion and $145 billion on capital expenditures for its AI data centers this year, in addition to significantly higher operating expenses, also thanks to AI, on outside cloud services and AI talent. How much money exactly is Zuckerberg willing to fork over on this quest?

We shouldn’t be surprised. Since 2020, Meta’s Reality Lab division has lost $87 billion, in large part due to Zuckerberg’s metaverse strategy. Despite that heavy spending, Meta has has backed away from the metaverse over the past couple of years as it has focused on AI.

Investors didn‘t seem to mind Meta’s plans. While Meta stock has been an underperformer so far this year, it was up 4.7% on Thursday.

Top Coinbase Lawyer Looks for New Battlefield

About a year ago, Coinbase Chief Legal Officer Paul Grewal said in a LinkedIn post that he had “the best job in law.” Apparently he no longer feels that way. Coinbase said on Thursday Grewal would step down, after nearly six years at the firm.

“Now is my time for new adventures,” Grewal said in a post on X Thursday, using what is now the standard phrase used by people changing jobs.

Grewal had a busy time at Coinbase. He led the effort to defend the company from the Securities and Exchange Commission’s 2023 lawsuit, which alleged that Coinbase was running an unregistered securities exchange and broker. Last year, under a new presidential administration, the SEC dropped the lawsuit. More recently, Grewal led negotiations on the crypto market structure bill, the Clarity Act, which is still stalled in Congress.

Crypto advocates are hoping to pass the bill before the midterms. It will provide a comprehensive framework on how to regulate tokens and platforms. Coinbase said Grewal will continue to serve as an adviser to the company until the end of October. In the meantime, we’re wondering where he will go next. What’s the bet it’s an AI firm?—Yueqi Yang

In Other News

• Fidji Simo, OpenAI’s CEO of AGI Deployment, is stepping down from that role to become a part-time advisor, she revealed Thursday. Simo had been on medical leave since April after the worsening of a chronic illness she has long suffered from.

• OpenAI, as part of its effort to attract more business customers, announced a new agent—ChatGPT Work—which taps into corporate data to automate the creation of spreadsheets and presentations, and can also handle more complex tasks like updating financial forecasts and conducting research.

• Anthropic on Thursday reiterated its policy that access to its Claude models and coding tools is not permitted in China, responding to the Chinese government’s recent warning about a security backdoor in Claude Code.

• Anthropic’s Long-Term Benefit Trust, the group charged with appointing new board members for the company, named Ben Bernanke, former chair of the Federal Reserve, as its fourth member on Thursday.

Today on The Information’s TITV

Check out today's episode of TITV in which we speak with the CEO of Daydream about the current state of AI and e-commerce.

Recommended Newsletter

Start your day with Applied AI, the newsletter from The Information that uncovers how leading businesses are leveraging AI to automate tasks across the board. Subscribe now for free to get it delivered straight to your inbox twice a week.