Meta layoffs 2026, Project OT and the 60 percent smaller-team plan

Meta Layoffs 2026: Secret AI Plan Explored 60% Smaller Teams. The Threat Is Not Over.

Meta explored turning some 10-to-20-person teams into three-to-five-person AI-assisted pods, eliminating management layers and pooling specialists. Zuckerberg canceled a planned November restructuring wave, but the smaller-workforce design remains the danger.

Quick answer

Meta did not plan to lay off 60 percent of its entire workforce. A Reuters investigation found that Meta's Project OT examined scenarios in which some teams could become up to 60 percent smaller through layoffs, redeployments and closed positions. Traditional product teams of 10 to 20 people could be redesigned as three-to-five-person AI-assisted pods with fewer managers and shared specialists. Meta cut about 10 percent of its workforce in May 2026, then stopped planning a broader November restructuring phase before determining its final size. That cancellation does not erase the worker threat. The operating idea remains clear: improve the AI, make teams smaller, remove layers and expect fewer people to cover more work.

Meta Project OT facts workers need to know

The numbers reveal a workforce design built around smaller teams, fewer layers and AI systems that had not yet delivered the productivity management expected.

Up to 60 percent smaller in some scenarios

Meta confirmed that the most aggressive Project OT scenarios examined reducing the size of some teams by up to 60 percent. This was not a plan to fire 60 percent of the entire company.

10 to 20 people could become 3 to 5

An internal AI-native playbook compared traditional product teams of 10 to 20 people with smaller pods containing three to five people and shared specialists.

About 8,000 employees affected in May

Meta's official second-quarter results say approximately 8,000 employees were affected by the May 2026 headcount reduction.

The November phase was canceled

Zuckerberg stopped planning the second Project OT restructuring phase hours before the May layoffs. No final total for the canceled phase had been determined.

Free worker-first intelligence

Get the Corporate Stress Index + Layoff Intelligence Report

Free signals on layoffs, AI job cuts, restructuring, corporate pressure, and workplace survival — before the official story lands.

Free email updates. Unsubscribe anytime.

Watch the episode

META'S Secret AI Layoff Plan Exposed

Watch The Grind Hotline break down Meta Project OT, the up-to-60-percent smaller-team scenarios, the canceled November restructuring wave and why the worker threat remains.

Meta canceled the second wave, not the destination

Meta was not simply handing employees AI tools and hoping they worked faster. It was redesigning the company around a blunt question: how few humans would an AI-native team require?

The Reuters Special Report on Project OT describes an internal plan for smaller human teams supervising virtual workers, fewer management layers, shared specialists and a broad mix of layoffs, redeployments, performance exits and closed openings.

Zuckerberg stopped planning the second restructuring wave scheduled for November. That matters, but it does not make the workforce model disappear. Meta had already defined the destination: three-to-five-person pods covering work previously assigned to teams of 10 to 20.

The immediate purge was reduced. The economic logic survived. If the AI improves, management can return to the same org chart with stronger evidence and less resistance.

No, Meta was not planning to fire 60 percent of the company

Reuters did not report that Meta intended to lay off 60 percent of its entire workforce. Meta confirmed that some individual-team scenarios examined reductions of up to 60 percent and said several major units were outside the exercise.

The scenarios combined different actions. Some employees could be laid off. Others could be redeployed into priority units. Open positions could be closed. Workers labeled poor performers could be pushed out. Project OT leaders canceled the second phase before settling the total number of people who would lose jobs.

That distinction belongs near the top because the accurate version is already severe. A company does not need to remove 60 percent of every employee to create a brutal worker threat. It can shrink selected teams, pool specialists, delete management layers and make the remaining structure the model for future reorganizations.

The headline number therefore describes possible team-level compression, not a confirmed company-wide layoff percentage.

Project OT was a workforce operating system

Project OT stood for Organization Transformation. Meta told Reuters it was a year-long effort involving cost reduction, redesigned teams and transfers into new priority work, including the production of training data for AI models.

The internal playbook replaced traditional product-development roles with fluid builder positions. Engineers, designers and other pod members would perform whatever work was needed to move the product. Fixed role boundaries weakened.

Product designers, user-experience researchers, data scientists, data engineers and machine-learning specialists could be shared across multiple pods instead of holding dedicated seats inside one team. Middle-management layers could disappear while high-level unit leaders gained responsibility for larger populations.

This is the real Meta AI workforce threat. Automation does not need to replace one named worker with one named machine. Management can redesign capacity around smaller pods and then decide which people no longer have a permanent place.

The earlier Meta layoffs story and this investigation answer different questions

The previous Grind Hotline investigation into Meta layoffs and Zuckerberg's AI-agent admission documented the May reduction, the wider headcount history, expensive AI hiring and Zuckerberg's acknowledgment that agent technology had not progressed as expected.

Project OT exposes the planned organization behind that pressure. It shows the pod sizes, the vanished layers, the shared specialists, the two-wave calendar and the most aggressive team-reduction scenarios.

The Meta, Microsoft and Google Corporate Stress Index investigation places Meta inside the wider Big Tech capital-swap pattern. This article stays focused on how Meta intended to translate AI investment into a smaller operating structure.

Why the canceled November phase does not create safety

Reuters could not determine one definitive reason Zuckerberg changed course. The reversal followed several pressures at once: employee revolt, collapsing morale, weak evidence of useful AI productivity, technical disruption and investor questions about the return on Meta's enormous spending.

Meta's favorable employee-sentiment score reportedly fell from 74 percent to 55 percent as the restructuring advanced.

Hours before the May layoffs, Zuckerberg halted planning for the November phase. The next day, Meta cut approximately 10 percent of its workforce and Zuckerberg told employees he did not expect additional company-wide layoffs during 2026.

The words company-wide and expect leave room for uncertainty. Team-specific cuts, performance exits, missing backfills, contractor reductions and reorganizations can continue without becoming another company-wide event.

Zuckerberg also told employees he expected AI agents to improve and begin producing more benefits within three to six months. A delayed technology can postpone a workforce decision. Once performance improves, the same decision becomes easier to defend.

Three-to-five-person pods change who gets a permanent seat

The traditional team in Meta's internal comparison contained 10 to 20 people with specialized roles. The AI-native model used three to four builders, one direction lead and specialists shared across pods.

By June, at least 11 Meta units, including engineering and research teams, had already implemented small pods. The November restructuring wave stopped, but part of the operating model was already running.

That structure concentrates permanent seats around people who can build across boundaries. Dedicated specialists become a shared service. Formal managers lose layers. Product managers may lose exclusive control of planning. Pod leads can inherit daily responsibility without receiving the full authority or tools of a manager.

Shared specialists also face a visibility problem. When one designer, researcher or data scientist supports several pods, management can evaluate that function as a ratio. The question becomes how many pods one specialist can cover rather than how much value one specialist creates.

The guide explaining how companies decide who gets laid off first shows why future structure can outweigh a strong historical review. Your old role can disappear even when the work remains.

Meta employees were helping train systems that could reduce human work

Reuters previously reported that Meta required tracking software on some U.S. employee devices to capture mouse movements and keystrokes for AI training. The goal was to teach agents how humans performed computer-based work.

Employees understood the threat immediately. They were completing the job, producing the behavioral data and watching management build systems intended to perform more of the same work.

Meta later paused the mouse-tracking program as executives tried to repair morale. Pausing one collection method does not erase the broader strategy of training agents on human workflows.

The separate Grind Hotline investigation into the Meta AI layoff lawsuit and productivity scoring examines another risk: employee activity becoming a machine-readable signal inside performance and layoff decisions.

The productivity numbers exposed the difference between motion and value

AI-assisted code activity surged. An internal Meta post reported that code changes on internal platforms and infrastructure rose about 220 percent year over year.

The number reaching users told a weaker story. Changes that produced new or improved features increased only 36 percent. More code was moving through the system, but the growth in useful customer outcomes was far smaller.

Reliability also deteriorated. Reuters reported internal data showing major technical and security incidents up about 40 percent and employee firefighting time up 70 percent. Meta declined to comment on those internal disruption figures.

This is the AI productivity trap in its most expensive form. A dashboard celebrates volume while workers absorb review, correction, security and recovery work. Management sees speed. The surviving employees experience the cleanup.

Meta's public AI message still predicts smaller companies

Meta's public position now emphasizes empowering people rather than replacing them. Zuckerberg's essay, The Future Is for Everyone, argues that AI can create businesses and employment by giving individuals greater capabilities.

The same essay says company sizes may shrink and predicts that small groups using powerful agents will be able to operate at significant scale. That is consistent with the core economics behind Project OT, even though the public framing is more optimistic.

Workers do not need to debate whether AI creates jobs somewhere else in the economy. The immediate question is whether Meta expects the same amount of product work to require fewer dedicated employees inside an existing organization.

Project OT shows that management already explored that answer in concrete team sizes.

Red Flag 1: The plan stopped because execution broke down

Meta did not announce that larger teams were permanently necessary. It encountered employee resistance, weak agent performance and reliability costs while attempting to move too quickly.

That makes the cancellation fragile. A permanent reversal would require management to reject the smaller-team model itself. Reuters found no evidence that Meta abandoned the belief that AI can eventually support fewer layers and leaner teams.

Watch for a renewed pilot described as safer, more gradual or better measured. The language may change while the destination remains intact.

Red Flag 2: Every specialist can become a shared service

Project OT treated designers, researchers, data scientists, data engineers and machine-learning specialists as capabilities that could be pooled across pods.

Pooling changes the employment calculation. Dedicated roles become coverage ratios. Management asks how many teams one person can support and whether an AI tool can absorb part of the remaining demand.

Track whether specialists are removed from product reporting lines, assigned to centralized groups or asked to serve a growing number of teams without matching headcount.

Red Flag 3: Manager elimination becomes capacity reduction

The AI-native playbook removed layers of middle management and placed more employees beneath senior unit leaders. Pod leads handled direction without necessarily receiving formal management authority.

This compresses two kinds of work. Fewer managers retain larger spans while senior individual contributors inherit coordination, prioritization and people problems alongside delivery.

A promotion into informal leadership can become a warning when authority, compensation and staffing stay flat while responsibility expands.

Red Flag 4: AI output can raise expectations before it improves results

The 220 percent increase in code changes creates a new internal benchmark even though customer-facing improvements rose much less and incident response consumed more time.

Once management sees that volume, future plans can assume it continues. Teams may receive higher output targets, shorter delivery windows and less staffing before the technology becomes dependable.

Document the difference between generated activity and finished business value. Shipping, adoption, revenue, stability, security and avoided failures matter more than raw code or task counts.

Red Flag 5: The $130 billion bill keeps demanding a return

Meta's second-quarter Form 10-Q anticipates $130 billion to $145 billion in 2026 capital expenditures supporting AI efforts and the core business.

Its official second-quarter results also show $1.18 billion in severance expense connected to the May reduction. AI investment and workforce restructuring are already appearing together in the financial record.

A spending program this large creates pressure to prove economic value. Revenue growth can satisfy part of that demand. Productivity and lower labor intensity can satisfy another part.

The broader Big Tech AI spending and headcount investigation explains why workers become the balancing item when infrastructure receives the capital.

Which Meta jobs face the clearest Project OT pressure

The exposed categories come directly from the workforce design, not from a claim that every employee in these roles will be cut.

Middle managers face pressure from flatter structures. Product managers can lose planning ownership inside builder-led pods. Designers, user-experience researchers, data scientists, data engineers and machine-learning specialists can be pooled. Engineers may be expected to cover broader builder work with AI assistance.

Employees between priority assignments face another problem. Meta can redeploy people into AI training, small-business initiatives or other strategic units, but a transfer is not permanent protection if the destination remains experimental.

Higher-cost employees, workers with narrow ownership, people supporting duplicated platforms and roles dependent on processes Meta wants to automate should watch the next org charts closely.

Ten Meta warning signs that Project OT is returning

1. Traditional job titles are replaced with builder, direction lead or similarly broad labels.

2. Product teams are reorganized into three-to-five-person pods.

3. Designers, researchers and data specialists move into centralized pools.

4. Managers receive wider spans while vacancies remain unfilled.

5. Pod leads inherit people responsibility without formal manager authority.

6. AI output measures become part of performance conversations.

7. Internal teams resume capturing detailed workflow data to train agents.

8. Technical incidents remain high while delivery expectations continue rising.

9. Meta announces local, function-specific or performance-based cuts while avoiding the phrase company-wide layoffs.

10. Executives say agent performance has improved enough to redesign workflows again.

One signal can reflect an ordinary experiment. Several appearing together suggest the Project OT workforce model is moving again.

Quiet Power moves before the next Meta org chart

Start interviewing while the Meta name still carries maximum market value. Preparation does not mean a layoff is guaranteed. It prevents a future company decision from controlling every option.

Package non-confidential evidence of what you shipped, revenue you supported, costs you reduced, incidents you prevented, systems you stabilized and decisions that depended on your judgment. Raw activity is weaker than completed business outcomes.

Map your role against the Project OT design. Identify whether you are a permanent builder, a direction lead, a pooled specialist, an informal coordinator or a function that can be centralized. Then build the missing evidence or outside options.

Know your vesting dates, equity treatment, severance position, health coverage and lawful personal records. Never remove Meta source code, internal documents, user data or proprietary information. The guide to preparing before a layoff happens provides a clean checklist.

Reconnect with former colleagues, recruiters and trusted managers before urgency changes your tone. Quiet Power means moving while you still have time, leverage and a credible story.

Use three Grind Hotline products for three different decisions

The free Job Threat Check uses seven questions to test company, team, role and manager pressure. Meta employees can use it to examine pod restructuring, shared-specialist risk, AI measurement, missing backfills and whether their work fits the organization being built.

The Layoff Tracker and Corporate Stress Index organize confirmed cuts and public workforce-pressure signals across major employers. Meta belongs in that record because its May reduction is confirmed and Project OT reveals the operating logic behind continued pressure.

The free Weekly Layoff Intelligence Report follows unfinished signals after the headline. It helps workers track new Meta org changes, localized cuts, AI productivity claims, role pooling and whether smaller-team pilots begin spreading again.

Workers already facing termination, severance, a PIP or a difficult exit can review Layoff Career Counselling for private, practical support.

The Grind Hotline Read

Meta's second restructuring wave was stopped. That does not turn Project OT into a harmless brainstorming exercise.

The company had already defined the workforce it wanted: smaller pods, fewer managers, shared specialists, more AI assistance and fewer permanent seats.

Execution failed because the technology produced too much motion, too little usable improvement and too much repair work. Zuckerberg expects the systems to improve. Meta is spending up to $145 billion to help make that happen.

Workers survived this version of the plan. The next version may arrive with better AI, cleaner messaging and the same headcount mathematics.

About The Grind Hotline

The Grind Hotline is an award-winning, worker-first media and workforce intelligence platform covering layoffs, AI job cuts, restructuring, no backfill, outsourcing, performance pressure, severance and corporate strategy in language workers can use. Its business podcast, articles, YouTube reporting and short-form commentary reach audiences in more than 100 countries.

The analysis is built from the perspective of The Host, an ex-banker, former Fortune 100 and Fortune 500 global sales leader, author, entrepreneur and corporate-survival strategist with nearly two decades of experience around large organizations, revenue pressure, management systems and workplace politics. Quiet Power is The Host's practical method for reading corporate signals, protecting leverage and building options without unnecessary confrontation.

The platform's free worker products serve three different decisions. The Job Threat Check measures personal exposure. The Layoff Tracker and Corporate Stress Index organize public employer pressure. The Weekly Layoff Intelligence Report explains what changed and what workers need to watch next. Layoff Career Counselling provides practical support when a layoff, PIP, severance decision or difficult exit becomes personal.

The Host also works directly with companies, technology organizations, financial-services firms and other B2B businesses through three separate commercial products. CallTeam builds and operates outbound calling, appointment-setting, qualification and lead-reactivation systems. The 90-Day Revenue Engine diagnoses and rebuilds targeting, messaging, pipeline, follow-up, CRM discipline and management rhythm. The Sales Execution Lab strengthens calls, discovery, objection handling, follow-up and conversion through hands-on work tied to real execution.

That operating work provides a direct view into what companies do when technology investment, productivity targets, revenue expectations and staffing decisions collide. Reporting and commercial work remain separate and follow The Grind Hotline's published Media and Editorial Standards.

Important Disclaimer

This article is media, commentary, education and career-strategy support based on public reporting, company filings and company statements available on August 26, 2026.

Reuters reported that Meta's Project OT examined scenarios in which some teams could become up to 60 percent smaller through layoffs, redeployments and closed positions. Meta said it never intended to lay off 60 percent of its entire workforce and canceled the second phase before determining a final company-wide job-loss total.

Reuters could not determine one definitive reason Zuckerberg stopped planning the November phase. Discussion of AI readiness, employee resistance, investor pressure, future layoffs, exposed roles and the possible return of smaller-team structures is analysis, not a claim that a particular Meta employee or team will lose a job.

Nothing here is legal, financial, investment, tax, immigration, employment, medical or mental-health advice. Confirm high-stakes decisions through official Meta communications and qualified professionals familiar with the applicable jurisdiction.

More Meta AI workforce pressure signals

AI-assisted code changes rose 220 percent

An internal post reported a 220 percent year-over-year increase in code changes on internal platforms, while changes producing new or improved user features rose only 36 percent.

Incidents rose 40 percent

Reuters reported internal data showing major technical and security incidents up 40 percent, while employee firefighting time increased 70 percent.

Employee sentiment fell from 74 to 55

Meta's internal favorable employee-sentiment measure reportedly declined from 74 percent to 55 percent as the restructuring and AI programs advanced.

$130 billion to $145 billion in 2026 capital spending

Meta's official filing says it anticipates $130 billion to $145 billion in 2026 capital expenditures supporting AI efforts and its core business.

Read next: Meta layoffs, AI spending and smaller-team pressure

These investigations separate Meta's confirmed layoffs, AI-productivity problems, worker-scoring risks and the wider Big Tech capital shift so each page answers a different question.

Meta Layoffs 2026: Zuckerberg Admits the AI Bet Is Not Working

Review the May layoffs, Meta's wider headcount history, AI-agent delay and the spending pressure that preceded the Project OT disclosure.

Meta AI Layoff Lawsuit and Productivity Scores

Understand the allegations involving productivity metrics, medical leave, disability, pregnancy and worker rights.

Meta, Microsoft and Google Corporate Stress Index

See how Meta's capital swap compares with Microsoft and Google's distinct workforce-reduction models.

Big Tech AI Spending and the Headcount Test

Follow how massive infrastructure spending can increase productivity and headcount pressure across major technology companies.

Welcome to the Forever Layoff

See how rolling cuts, no backfill, delayering and selective hiring create a permanent workforce-reset model.

How Companies Decide Who Gets Laid Off First

Understand how cost, overlap, future structure, performance and internal politics shape a layoff list.

Prepare Before a Layoff Happens

Protect lawful records, understand benefits and build outside options before company access disappears.

Am I About to Be Laid Off?

Check the warning signs inside your employer, business unit, team, role and manager relationship.

Questions workers are asking

Did Meta plan to lay off 60 percent of its workforce?

No. Reuters reported that some Project OT scenarios examined making certain teams up to 60 percent smaller. Meta said the scenarios combined layoffs, redeployments and closed positions and that it never intended to lay off 60 percent of the entire company.

What was Meta Project OT?

Project OT, short for Organization Transformation, was a year-long Meta initiative focused on cost reduction, redesigned teams and moving employees into priority work. Internal plans examined smaller AI-assisted pods, fewer management layers and specialists shared across teams.

How small could Meta's AI-native teams become?

An internal comparison reviewed by Reuters showed traditional product teams of 10 to 20 people being redesigned as pods of three to five people, including three or four builders, one direction lead and access to shared specialists.

Why did Zuckerberg cancel the November restructuring wave?

Reuters could not determine one definitive cause. The cancellation followed employee backlash, lower morale, disappointing AI-agent productivity, higher technical disruption and investor scrutiny. Meta canceled the phase before determining its final layoff total.

Will Meta conduct more layoffs in 2026?

Zuckerberg told employees he did not expect additional company-wide layoffs in 2026. That statement does not guarantee there will be no team-specific cuts, performance exits, missing backfills, contractor reductions or other organizational changes.

Which Meta jobs face the most pressure from Project OT?

The structure creates pressure for middle managers, product managers, designers, user-experience researchers, data scientists, data engineers, machine-learning specialists, coordination roles and employees whose work can be pooled across several pods. It does not prove that every worker in these roles will be cut.

Did Meta track employee keystrokes and mouse movements?

Reuters reported that Meta required tracking software on some U.S. employee devices to capture keystrokes and mouse movements for AI training. Meta later paused the mouse-tracking program as it responded to employee concerns.

What do Meta's 220 percent and 36 percent AI productivity figures mean?

An internal post reported AI-assisted code changes on internal platforms up about 220 percent year over year, while changes creating new or improved features for users rose only 36 percent. The comparison suggests activity increased much faster than useful product outcomes.

How much is Meta spending on AI and infrastructure in 2026?

Meta's official second-quarter filing anticipates $130 billion to $145 billion in 2026 capital expenditures supporting AI efforts and the core business.

What should Meta employees do after the Project OT disclosure?

Map your role against the smaller-pod structure, document non-confidential business outcomes, understand vesting and severance, reconnect with recruiters and former colleagues, monitor org changes and build outside options before another decision is announced.

Worker-first signals, not corporate spin

Don’t wait for the company memo.

Get the Corporate Stress Index, layoff intelligence, pressure signals, and workplace survival moves before the official story lands.

Free email updates. Unsubscribe anytime.

Find out whether Meta's smaller-team plan is reaching your job

Use the free Job Threat Check to examine pod restructuring, shared-specialist risk, AI measurement, missing backfills, manager compression and whether your role still fits Meta's next operating model.