Meta CEO Mark Zuckerberg has spent much of 2026 pushing the company toward an ambitious goal: building a workforce designed around artificial intelligence.
The strategy was supposed to transform how Meta operates, with AI agents handling more routine tasks, smaller teams moving faster and employees using artificial intelligence throughout the product-development process.
But the plan has run into unexpected obstacles.
An internal restructuring programme known as Project OT, short for Organization Transformation, envisioned a major redesign of Meta’s workforce. Some internal scenarios considered reducing the size of individual teams by as much as 60%, while shifting thousands of employees into areas considered strategically important to the company’s AI ambitions.
Meta ultimately carried out a major workforce reduction in May, cutting roughly 10% of its employees. But plans for a second wave of layoffs later in the year were abandoned shortly before they were due to move forward.
The reversal highlights a difficult reality facing Meta and other large technology companies: replacing traditional workflows with AI may be far more complicated than simply deploying powerful new models.
Zuckerberg’s Vision for an AI-Native Meta
Meta’s AI strategy goes beyond using chatbots to help employees write emails or generate code.
Zuckerberg has been pursuing a much broader transformation in which artificial intelligence becomes part of the company’s basic operating structure.
The idea behind Project OT was to create what executives described as an “AI-native” company.
Under this model, AI agents would perform many repetitive tasks while smaller groups of highly skilled employees would supervise the systems, develop products and make higher-level decisions.
The approach could potentially reduce the need for large teams and layers of management.
Internal planning documents reportedly explored scenarios in which some Meta teams could be reduced by as much as 60%. However, those figures represented planning exercises rather than a company-wide target.
Meta has stressed that it did not intend to eliminate 60% of its overall workforce.
Instead, the company says the exercise considered different combinations of employee redeployments, unfilled positions and potential reductions as executives evaluated how the organisation might change.
Meta Still Carried Out Major Layoffs
Despite abandoning the second planned round of cuts, Meta did not completely reverse its restructuring programme.
The company eliminated approximately 10% of its workforce in May.
Thousands of other employees were moved into newly created teams or reassigned to work considered more important to Meta’s future AI strategy.
The difference is significant.
Rather than simply removing employees, Meta attempted to reshape the workforce around artificial intelligence.
Employees who previously worked in traditional engineering, product or management structures could find themselves working in smaller groups designed to move projects from concept to prototype more quickly.
The company has also been looking more closely at which employees possess skills that could become especially valuable as AI changes how software and products are developed.
The Rise of AI “Tech Pods”
One of the most distinctive parts of Meta’s transformation has been the development of smaller teams known internally as tech pods.
Instead of large groups working on lengthy product-development cycles, these teams are designed to operate with only a handful of people supported by AI tools.
One early experiment involved five small pods, with each group consisting of only a few engineers and a designer.
The teams were expected to build prototypes in short cycles rather than spending months planning products before development began.
The philosophy was straightforward: if AI can reduce the amount of time required to write code, analyse information and perform other routine tasks, smaller groups should theoretically be able to accomplish more.
Meta executives compared the approach to a fast break in basketball.
The idea was that AI would allow employees to experiment with more ideas, build prototypes faster and eliminate some of the traditional costs associated with product development.
By the middle of the year, versions of the smaller-team structure had reportedly spread across numerous Meta divisions.
The Company Wanted Fewer Management Layers
Meta’s AI transformation was not limited to engineers.
The company was also examining its management structure.
One internal strategy involved reducing layers of management and giving employees broader responsibilities.
Traditional roles such as product designer and engineer could potentially become less rigid as AI tools made it easier for employees to perform tasks outside their original specialisations.
That could lead to a broader “builder” role, where a smaller number of employees handle several parts of the product-development process.
The concept reflects a wider trend across Silicon Valley.
As generative AI becomes increasingly capable of writing code, producing designs, analysing information and generating content, technology companies are questioning whether traditional job boundaries still make sense.
For Meta, the goal was to turn that technological change into a new organisational model.
AI Productivity Did Not Meet Expectations
The biggest problem was that the technology did not immediately deliver everything executives had hoped for.
According to the Reuters investigation, internal data showed that autonomous AI agents were not producing the expected productivity improvements.
That became a significant problem for a restructuring strategy built partly around the assumption that AI could enable substantially smaller teams to accomplish the same amount of work.
AI systems can perform certain tasks remarkably quickly, but that does not automatically translate into higher productivity across an entire organisation.
Employees still have to monitor AI-generated work, correct mistakes, integrate systems into existing infrastructure and deal with unexpected technical problems.
In some cases, those additional responsibilities can offset the time saved by automation.
Meta’s experience illustrates the difference between demonstrating that an AI system can perform a task and proving that the system can reliably replace an established human workflow.
Employee Resistance Complicated the Transformation
The restructuring also faced resistance from inside Meta.
Employees became increasingly concerned that the company’s AI strategy was not simply about making workers more productive but about reducing the number of people needed to perform certain jobs.
Those concerns intensified as reports emerged that Meta was considering much larger workforce reductions.
The uncertainty created another challenge for management.
A company attempting to transform itself around AI needs employees to experiment with new tools and adopt unfamiliar workflows. At the same time, employees who believe those same technologies could eventually eliminate their jobs may be less enthusiastic about the transformation.
That tension is becoming increasingly common across the technology industry.
AI can be presented as a productivity tool that allows employees to accomplish more. But when companies simultaneously announce cost-cutting programmes and workforce reductions, employees may interpret the technology as a replacement strategy.
Meta’s experience demonstrates how difficult it can be to separate those two narratives.
Meta Says Humans Still Make Key Decisions
Meta has pushed back against the idea that artificial intelligence is taking over decisions about individual employees.
The company has said that performance ratings and promotion decisions remain the responsibility of human managers.
That distinction is important as companies introduce AI into increasingly sensitive areas of their operations.
Using AI to generate code or summarise documents is very different from allowing an algorithm to determine who receives a promotion or who loses their job.
Meta’s position is that artificial intelligence can assist employees and managers without becoming the final authority over personnel decisions.
The company’s approach also includes efforts to identify and retain employees whose skills are considered particularly difficult to replace.
Why Meta Abandoned the Second Layoff Wave
The precise reason Zuckerberg decided to cancel the second planned wave of layoffs remains unclear.
Reuters reported that the decision was made just hours before the first round of cuts in May, with a second round originally expected in November.
By that point, however, employee resistance had grown and internal AI productivity results had reportedly fallen short of expectations.
Meta has not publicly provided a single explanation for why the second round was abandoned.
The decision could therefore reflect several factors, including employee morale, the performance of AI systems, the need to retain specialised talent and the practical difficulties of reorganising a company as large as Meta.
What is clear is that the company proceeded with a substantial restructuring while stopping short of the more aggressive scenarios considered during its internal planning.
The AI Spending Challenge Continues
Abandoning additional layoffs does not mean Meta is backing away from artificial intelligence.
The opposite appears to be true.
Meta continues to invest heavily in AI infrastructure, custom chips, data centres and AI models.
The company has also been expanding its AI products and working toward increasingly capable AI agents.
That creates an unusual situation.
Meta is spending enormous amounts of money to develop AI while simultaneously discovering that integrating AI into a workforce of tens of thousands of employees is not straightforward.
The company therefore faces two separate challenges: building the technology and figuring out how to organise people around it.
The second challenge may ultimately prove just as important as the first.
A Warning for Other Technology Companies
Meta’s experience could provide an important lesson for other businesses pursuing AI-driven restructuring.
Artificial intelligence can automate specific tasks, but transforming an entire organisation requires much more than deploying new software.
Companies have to redesign workflows, retrain employees, change management structures and determine which responsibilities should remain with humans.
They also need reliable ways to measure whether AI is actually improving productivity.
That last point is particularly important.
An AI system can appear impressive during demonstrations while producing far smaller benefits when introduced into complex real-world operations.
The gap between AI’s technical capabilities and its practical business value can become especially visible inside large organisations with complicated systems and established processes.
Meta’s AI Revolution Is Still Underway
Project OT may not have delivered the sweeping transformation originally envisioned, but Meta’s broader AI strategy is far from over.
The company continues to build AI infrastructure and develop increasingly autonomous systems. It is also experimenting with smaller teams, new job structures and different ways of integrating artificial intelligence into everyday work.
What has changed is the pace and scale of the workforce transformation.
Instead of assuming that AI can immediately replace large numbers of employees, Meta appears to be taking a more cautious approach after the difficulties surrounding its initial restructuring plans.
For Zuckerberg, the challenge is now bigger than simply developing powerful AI.
He must demonstrate that the technology can reliably improve the productivity of one of the world’s largest technology companies without creating organisational instability in the process.
The story of Meta’s AI workforce transformation shows that replacing traditional corporate structures with artificial intelligence is not as simple as replacing people with machines.
AI may be capable of writing code, analysing data and performing increasingly complex tasks. But turning those capabilities into measurable gains across a huge organisation requires experimentation, trust and a carefully managed transition.
For Meta, that transition is still underway.
