AIToday
Large Language ModelsAI Business & IndustryFortune AIPublished: Oct 17, 2026, 22:01 JST

44-minute call led to Meta Superintelligence Labs, $130 billion to $145 billion 2026 capex

44-minute call led to Meta Superintelligence Labs, $130 billion to $145 billion 2026 capex

3 Key Points

  1. What happened

    On John Collison's Cheeky Pint podcast, Zuckerberg and Wang said a 44-minute call and a memo led to Meta Superintelligence Labs, and Meta now expects 2026 capex of $130 billion to $145 billion.

  2. Why it matters

    The partnership behind Meta Superintelligence Labs is tied to a capex scale Zuckerberg himself framed as a $100 billion or $200 billion bet, and investors have flinched at each increase, suggesting spending scrutiny may persist.

WHO IT HITSMeta investors and Wall Street analysts tracking capex will likely focus on whether the Muse app's early traction, including 2.5 million downloads in about two weeks, justifies the 2026 spending range of $130 billion to $145 billion.

Not sure about something? Ask the AI

Questions and answers are published on this page.

Summaries like this, in your inbox every morning.

Context & Analysis

The podcast account fills in how Meta's AI effort was rebuilt after Llama 4 landed to a lukewarm reception in April 2025. Zuckerberg reached out to Wang for advice, and the relationship moved quickly from a 44-minute call to a memo and then to Wang leading Meta Superintelligence Labs. The recruiting push that followed was described as "supernova," with Meta paying $14.3 billion for a 49% stake in Scale AI to bring Wang aboard and reportedly dangling $100 million compensation offers to OpenAI researchers.

The financial stakes were impressed on Wang by Meta's CFO, Susan Li, who warned at an executive gathering in Hawaii that failure would be "very, very bad in financial terms." Meta now expects 2026 capital expenditures of $130 billion to $145 billion, nearly double the $72.2 billion it spent last year. The Muse app launched in the U.S. on Sept. 8 and topped 2.5 million downloads in about two weeks, and Meta followed with Muse Wonder, the lab's first frontier-class model.

FAQ
How did Meta Superintelligence Labs start?
According to Zuckerberg and Wang, it began with a roughly 44-minute call in April 2025, followed two days later by a memo from Wang outlining what he would do in Zuckerberg's position.
How does Muse Wonder compare with rival models?
On Artificial Analysis's Intelligence Index, Muse Wonder Full lands roughly level with top models from OpenAI and Anthropic but a couple of points behind the leader, according to AI Weekly.
How many people are on the core research team?
Wang said the core research team inside Meta Superintelligence Labs numbers fewer than 200 people, closer to 100, at a company with roughly 75,000 employees.

AI news that matters for your work, delivered every morning.

Pick your industry and the AI tools you use, and get news related to your work every day.

Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.

Questions and answers are published on this page.

Related Articles

Next articleOpus 5.5 returns HTTP 400 on ALPHA FORGE's forced tool spec