More than fourteen months have passed since Mark Zuckerberg put $14.3 billion into Scale AI in June 2025. The gap between that opening bet and the current shape of Meta Superintelligence Labs tells a sharper story than the headline did. Zuckerberg bought two things that week: a 49% stake in Scale AI, and its 28-year-old founder, Alexandr Wang, installed at once as Meta's first chief AI officer1.
Meta said so plainly in its own statement. "Meta has finalized our strategic partnership and investment in Scale AI. As part of this, we will deepen the work we do together producing data for AI models and Alexandr Wang will join Meta to work on our superintelligence efforts."11 Wang arrived with a brief: build a research group that could match OpenAI and Google DeepMind. He built it fast, at high cost, around a handful of names.
The Summer of Nine Figures
Wang built it with two men who came from different places. He had founded Scale AI in 2016 after leaving MIT, growing it into the data supplier many frontier labs used for training and testing1. Nat Friedman, the former GitHub chief executive who ran the AI fund NFDG for years, joined Meta in July 2025 to lead product work. Daniel Gross, Friedman's NFDG partner and the founding chief executive of Safe Superintelligence, followed that same month into a parallel products role5. Read the pair as one purchase. Zuckerberg took a whole venture partnership, and the founder relationships it carried, in one move.
Shengjia Zhao supplied the technical core. OpenAI's co-creator of ChatGPT took the new role of MSL chief scientist on July 25, 202523. CNBC and TechCrunch both read it as Zuckerberg reaching past Meta's own veterans for outside authority. Ruoming Pang had arrived weeks earlier from Apple4. He spent fifteen years at Google, then joined Apple in 2021 to lead the roughly 100-person team behind Apple Intelligence. Meta's offer reportedly ran to tens of millions a year.
Numbers that size traveled fast. Reports of nine-figure signing packages for single researchers spread within weeks of the Scale AI deal6. DeepLearning.AI's The Batch recorded the knock-on effect: engineering pay across the sector rose as rival labs matched Meta to keep their own people. Wang, Friedman, Gross, Zhao and Pang were the visible part of a campaign that pulled researchers from OpenAI, Google DeepMind and Apple for months.
Four Groups, One Org Chart
August 2025 brought structure to the spending. Zuckerberg split MSL into four groups1. TBD Lab, the frontier-model unit Wang runs himself. FAIR, the fundamental-research arm inherited from Meta's original AI organization. Products and Applied Research, the consumer-facing group under Friedman. MSL Infra, the compute and systems layer led by Aparna Ramani. Four boxes on a chart rarely explain themselves, and this one does. Each names a job Meta needed done at once. Ship a model that competes with GPT and Gemini. Hold academic credibility. Put superintelligence work into Meta AI and Instagram. Keep the infrastructure under all of it standing.
The design also settled a question every frontier lab faces. Should research and product share a reporting line, or run in parallel? Meta chose parallel. A frontier-model delay inside TBD Lab stays there, and Meta AI features on Friedman's team keep shipping on their own schedule. Product trouble stays contained too, and the scientists Zhao recruited keep working clear of it.
Wang's two roles, chief AI officer and TBD Lab lead, put authority in one place, and that mattered later. Executives who run both a company-wide function and its most visible unit tend to outlast reorganizations aimed at everyone below them. His position by autumn 2025 fit the pattern exactly.
The Freeze After the Rush
Meta paused AI hiring on Aug. 21, 2025, TechCrunch reported, barely a month after the Pang and Zhao announcements dominated coverage7. The freeze followed a summer in which Meta's offers reset pay across the industry. A pause signals either satisfaction with the roster, or, more likely, that folding in dozens of senior hires takes time.
October brought the harder correction. Meta cut about 600 positions from the AI division on Oct. 22, 2025, in a move CNBC described as trimming a "bloated" unit while Wang consolidated control of what remained8. Axios, reporting the same day, read the cuts as proof the four-group design worked as intended9. TBD Lab absorbed most of the frontier-model headcount, and adjacent teams doing overlapping work saw researchers reassigned or released.
Meta paused AI hiring on Aug. 21, 2025, barely a month after hiring Zhao, then cut 600 AI jobs on Oct. 22
Dated steps at Meta Superintelligence Labs, from the Scale AI deal in June 2025 to LeCun's exit on Nov. 20
Each pin is one dated step on a rail spaced by real time: blue pins are the Scale AI bet and the summer's marquee hires, and magenta pins are the freeze, the cuts and LeCun's exit.
Sources: Wikipedia [1]; CNBC [2]; Entrepreneur [4]; TechCrunch [7]; CNBC [8]; TechCrunch [11]. Chart by Ryan Elliott Dennis.
The numbers behind this chart
| Date | Event |
|---|---|
| June 2025 | $14.3B into Scale AI. Meta puts $14.3 billion into Scale AI; A 49% stake, with Alexandr Wang as chief AI officer |
| July 7, 2025 | Pang joins Meta. Ruoming Pang leaves Apple for Meta; Reported pay of tens of millions a year |
| July 25, 2025 | Zhao chief scientist. Shengjia Zhao becomes MSL chief scientist; OpenAI's co-creator of ChatGPT |
| Aug. 21, 2025 | Meta pauses AI hiring |
| Oct. 22, 2025 | Meta cuts ~600 AI jobs. Meta cuts about 600 AI positions; Wang consolidates control of what remains |
| Nov. 20, 2025 | Yann LeCun leaves Meta. Meta's chief AI scientist since 2013 |
LeCun's Long Exit
Yann LeCun's exit carried more weight than any layoff number. Meta's chief AI scientist since 2013, the year the company built its research arm around him, he left on Nov. 20, 20251. Zhao's arrival had made an awkward pair: two scientists of serious standing on overlapping ground. LeCun had argued for years that large language models are a limited path to general intelligence. MSL's LLM-first structure under Wang left him alone inside the group he helped define. He won the Turing Award in 2018 for work on convolutional neural networks, and spent more than a decade as the loudest doubter of the scaling idea driving his own employer's plan.
The Decoder tied his exit in January 2026 to the founding of Advanced Machine Intelligence Labs, built around V-JEPA-style world models that train on video and spatial data10. Alex LeBrun took the chief executive job, and LeCun serves as executive chair, setting research direction while handing off daily work.
LeCun explained the break in his own terms. "You don't tell a researcher what to do. You certainly don't tell a researcher like me what to do."10 He named the cost of saying so out loud. "I'm sure there's a lot of people at Meta, including perhaps Alex, who would like me to not tell the world that LLMs basically are a dead end." Then he refused to soften it. "I'm not gonna change my mind because some dude thinks I'm wrong. I'm not wrong. My integrity as a scientist cannot allow me to do this." He also described what the churn looked like from inside. "A lot of people have left, a lot of people who haven't yet left will leave."
Reading the Org Chart as Strategy
Muse Spark shipped April 8, 2026, the first release in Meta's new Muse model family, and the clearest evidence the restructured chart produces output as reliably as it produces turnover1. Read the whole run as a strategy document rather than a personnel drama. A $14.3 billion bet bought Meta a seat among frontier labs. Nine-figure packages bought speed. The four-group restructuring bought accountability. Losing LeCun and others bought something Meta may have wanted anyway: a research organization narrower and more aligned around Wang's LLM bet than the one Zuckerberg first assembled.
Wang's tenure now lands in an industry that expects this pattern. Heavy capital, famous hires, quiet attrition, and a shipped product that mostly justifies the spending afterward. Talent was the early signal. Muse was the late one. Whether the pattern holds past one product cycle depends on retention numbers Meta rarely shares, and a Muse roadmap that, as of September 2026, is its clearest answer to the money spent.
By the numbers
- Meta paid $14.3 billion for 49% of Scale AI in June 2025, and got Alexandr Wang as chief AI officer111.
- Tens of millions a year: Ruoming Pang's reported pay after leaving Apple's 100-person model team on July 7, 20254.
- July 25, 2025: the day Shengjia Zhao, OpenAI's co-creator of ChatGPT, became MSL chief scientist2.
- Four groups came out of the August 2025 restructuring: TBD Lab, FAIR, Products and Applied Research, and MSL Infra1.
- Six hundred positions went in Meta's Oct. 22, 2025 AI cut, as Wang consolidated control of the rest8.
- Nov. 20, 2025 was Yann LeCun's last day as Meta's chief AI scientist1.
- April 8, 2026 brought Muse Spark, the first model shipped under the new chart1.
What to watch
MSL's second Muse release will show whether TBD Lab ships as fast as it hired1. Retention data, if Meta ever shares it, would show whether the four-group structure settled the churn of 202589. AMI Labs' first published results give the clearest test of LeCun's world-model bet against the architecture his former employer funded at scale10. Rival labs will keep testing whether Meta's pay has a ceiling, and a new hiring surge would confirm Wang still has Zuckerberg's full backing into 2027.
Sources
- Wikipedia contributors, "Meta Superintelligence Labs," Wikipedia, accessed Sept. 4, 2026, https://en.wikipedia.org/wiki/Meta_Superintelligence_Labs
- CNBC Staff, "Meta names OpenAI's Shengjia Zhao as chief scientist of AI Superintelligence Lab," CNBC, July 25, 2025, https://www.cnbc.com/2025/07/25/zuckerberg-shengjia-zhao-meta-ai-lab-chief-scientist-openai.html
- TechCrunch Staff, "Meta names Shengjia Zhao as chief scientist of AI superintelligence unit," TechCrunch, July 25, 2025, https://techcrunch.com/2025/07/25/meta-names-shengjia-zhao-as-chief-scientist-of-ai-superintelligence-unit
- Entrepreneur Staff, "Meta Poaches Top Apple Executive With Compensation Offer Reportedly in the Tens of Millions," Entrepreneur, July 7, 2025, https://www.entrepreneur.com/business-news/meta-offers-apple-manager-tens-of-millions-to-join-ai-team/494354
- Entrepreneur Staff, "Meta Poaches Safe Superintelligence CEO for New AI Team," Entrepreneur, July 17, 2025, https://www.entrepreneur.com/business-news/meta-poaches-safe-superintelligence-ceo-for-new-ai-team/493604
- The Batch Staff, "Meta's Hiring Spree Pushes Up Salaries for AI Engineers Across the Industry," DeepLearning.AI, The Batch, July 25, 2025, https://www.deeplearning.ai/the-batch/metas-hiring-spree-pushes-up-salaries-for-ai-engineers-across-the-industry
- TechCrunch Staff, "Report: Meta is hitting pause on AI hiring after its poaching spree," TechCrunch, Aug. 21, 2025, https://www.techcrunch.com/2025/08/21/report-meta-is-hitting-pause-on-ai-hiring-after-its-poaching-spree/
- CNBC Staff, "Meta lays off 600 from 'bloated' AI unit as Wang cements leadership," CNBC, Oct. 22, 2025, https://www.cnbc.com/2025/10/22/meta-layoffs-ai.html
- Axios Staff, "Meta's Alexandr Wang reorgs superintelligence lab," Axios, Oct. 22, 2025, https://www.axios.com/2025/10/22/meta-superintelligence-tbd-ai-reorg
- The Decoder Staff, "LeCun Exits Meta for His Own Startup, Defends His Independence as a Researcher," The Decoder, Jan. 3, 2026, https://the-decoder.com/you-certainly-dont-tell-a-researcher-like-me-what-to-do-says-lecun-as-he-exits-meta-for-his-own-startup/
- TechCrunch, "Scale AI confirms 'significant' investment from Meta, says CEO Alexandr Wang is leaving," TechCrunch, June 13, 2025, https://techcrunch.com/2025/06/13/scale-ai-confirms-significant-investment-from-meta-says-ceo-alexandr-wang-is-leaving/