Meta confirmed on October 22, 2025, that it was eliminating roughly 600 roles within its Superintelligence Labs. The reductions affected longer-standing AI research, infrastructure, and product-related teams, while the newer TBD Lab—focused on next-generation foundation models—was reportedly spared and continued hiring.
That makes this a reorganization of Meta’s AI priorities, not evidence that the company is abandoning artificial intelligence. It also does not necessarily mean 600 people permanently left Meta: affected employees were encouraged to apply for other internal roles, and the company expected many to transfer.
As an Amazon Associate I earn from qualifying purchases.
What Meta changed
The cuts were reported by Axios and subsequently confirmed by Meta to TechCrunch and other outlets. They involved approximately 600 positions in Meta Superintelligence Labs, the organization created to bring the company’s major AI efforts under a more concentrated structure.
Public reporting identified three broad affected areas:
#1 Best Overall
- FAIR, or Fundamental AI Research, Meta’s longer-standing fundamental-research organization;
- AI infrastructure teams; and
- Product-related AI teams.
The available reports do not provide a complete role-by-role breakdown. It would therefore be inaccurate to say that every FAIR employee, infrastructure worker, or product-AI employee was dismissed.
The team Meta protected
The reported exception was TBD Lab, a newer and relatively small group working on Meta’s next generation of foundation models. According to the reporting, Meta continued recruiting for TBD Lab even as it reduced positions elsewhere in Superintelligence Labs.
This distinction is the key to understanding the move. Meta was reducing or reshaping parts of its existing AI organization while concentrating resources on a newer frontier-model effort. The company was cutting AI roles and hiring for AI roles at the same time—but not necessarily for the same teams or types of work.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Rank #2
Why Meta said it was making the cuts
Chief AI Officer Alexandr Wang said in an internal memo that a smaller organization would require fewer conversations to make decisions and would give individual employees more responsibility and impact. The stated goal was a leaner group with more “load-bearing” roles and faster execution.
That is Meta’s management rationale, not independent proof of the cause. The memo indicates that leadership believed the organization had become too difficult to coordinate, but the public reporting does not establish one technical failure or business result that triggered the reductions.
Were 600 people laid off?
Not necessarily. The “roughly 600” figure refers to roles affected or eliminated in the reorganization. Meta reportedly encouraged affected employees to apply for open positions elsewhere in the company and expected most to find internal roles.
Those situations are different:
- Role elimination: a position disappears from the restructured organization.
- Internal transfer: the employee moves to another Meta team.
- External departure: the employee leaves Meta.
The sources available for this report do not establish how many affected employees ultimately left the company. Accordingly, “Meta eliminated about 600 AI roles” is more precise than “600 people lost their jobs.” Details such as the final number of departures, individual job titles, locations, and severance arrangements were not fully disclosed in the reported accounts.
Do these 3 things before closing this tab:
1Repair Windows errors before they cause bigger problems2Scan for outdated or missing drivers - takes under a minute3Clear out junk files and repair common Windows errorsWhy the cuts followed an AI hiring spree
The restructuring came after Meta had made an aggressive effort to build a frontier-AI operation. Reporting described the company recruiting researchers from rival organizations with unusually large compensation packages and placing them in a new superintelligence structure under Wang. TechCrunch’s coverage placed the cuts in that broader reorganization.
The apparent contradiction is therefore the central business story: Meta was not simply reducing its AI ambitions. It was attempting to move people, authority, and investment toward a smaller group focused on next-generation foundation models while reducing layers in other parts of the organization.
Meta continues to compete with companies including OpenAI, Google, and Anthropic for frontier-model talent and computing resources. A reduction in some teams can coexist with higher spending or continued hiring in another AI unit.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.What the move says about Meta’s AI strategy
1. Meta is concentrating frontier-model resources
Protecting TBD Lab while cutting roles in FAIR, infrastructure, and product-related groups suggests that Meta was prioritizing a newer foundation-model effort over a broader, more distributed structure. This does not mean the work in the affected teams was worthless or obsolete. It means leadership was choosing a narrower organizational center of gravity.
2. Speed and accountability were strategic concerns
Wang’s explanation emphasizes decision speed, fewer organizational layers, and individual ownership. In frontier AI, where model-training cycles and research decisions can require substantial coordination, Meta appears to believe that a smaller structure may execute more effectively.
Best Value
3. The restructuring carries trade-offs
A leaner organization can reduce duplication and make responsibility clearer. It can also weaken institutional knowledge, disrupt long-running research programs, and make it harder to maintain infrastructure or translate research into products. The public reporting does not yet show whether Meta’s new structure will improve model development or product performance.
4. The cuts do not prove that Meta’s AI strategy has failed
There is no confirmed evidence in the cited reporting that the reductions were caused by a specific model problem, including alleged issues with Llama 4. Nor do the reports establish that Meta’s AI products were commercially unsuccessful or that the company was primarily seeking cost savings.
The strongest supported conclusion is narrower: Meta believed its AI organization needed to be more concentrated and faster-moving, and it redirected attention toward its newer superintelligence effort.
The Tool Desk
Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →What remains unknown
- The exact roles and locations included in the approximately 600 affected positions.
- How many employees transferred internally and how many left Meta.
- Whether additional reductions or reorganizations were planned.
- How Meta will measure the success of the new structure.
- Whether the smaller organization will produce better models, faster releases, or stronger AI products.
Reports from Reuters, The Associated Press, and The Washington Post broadly matched the account of cuts in parts of Superintelligence Labs alongside continued investment in the newer group.
The bottom line
Meta’s October 2025 announcement was a targeted AI reorganization, not a withdrawal from AI. About 600 roles were affected in FAIR, infrastructure, and product-related teams, while TBD Lab remained protected and continued hiring. For workers, the number of permanent job losses was lower—or could be lower—than the headline figure if internal transfers occurred. For investors and the wider AI industry, the move signals that Meta is betting on a smaller, more tightly controlled foundation-model operation rather than spreading resources evenly across its AI division.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
