Meta Platforms is weighing additional layoffs even as it ramps up spending on artificial intelligence and data center infrastructure, according to recent reporting and company filings. The company has already eliminated 600 positions in its AI divisions, and a March 2026 Reuters report said Meta was planning sweeping cuts that could reduce its workforce by as much as 20%. The tension between cost-cutting and massive capital investment underscores how the company is prioritizing its push toward advanced AI systems.
600 AI Jobs Gone, More Cuts Expected
The first wave of layoffs hit teams that had been central to Meta’s earlier AI efforts. The company confirmed cuts of 600 AI jobs, affecting its Fundamental AI Research (FAIR) group along with AI infrastructure and product teams. Those reductions are notable because FAIR has long been one of the most respected corporate AI labs in the world, responsible for open-source models and foundational research. Yet even as Meta shed those roles, it continued hiring aggressively for its superintelligence lab, signaling that the company is not retreating from AI but rather concentrating its bets on a narrower, more ambitious target.
The 600-job cut, though significant on its own, may prove to be a prelude. Reporting from March 2026 indicated that Meta was planning sweeping layoffs that could reduce its workforce by 20% as AI-related costs continued to climb. Separately, Reuters has reported Meta has considered reducing spending in some non-core areas as it prioritizes AI-related investment. Taken together, the pattern suggests Meta is not simply trimming around the edges but restructuring entire business lines to free up capital for its AI infrastructure buildout.
That restructuring is also reshaping the company’s internal culture. FAIR once symbolized Meta’s commitment to open science and broad-based AI research. The new structure, by contrast, prioritizes a smaller number of high-stakes projects tied directly to superintelligence and large-scale model deployment. Employees who remain are increasingly funneled toward work that serves these core initiatives, while legacy efforts in areas like recommendation systems and applied computer vision face tighter scrutiny over their return on investment.
Capital Spending Surges 73% for Superintelligence
The financial filings tell the story of where those freed-up dollars are going. In late January 2026, Meta disclosed that it had boosted capital spending by 73% for AI development as part of what the company described as a superintelligence push. Reuters reported that shares surged on the announcement, suggesting investors, at least initially, viewed the spending as a credible growth strategy rather than reckless expansion. The company also forecast a sharp increase in 2026 total expenses based on results for its quarter ended December 31, tying much of that rise to data center and AI infrastructure outlays.
The scale of the infrastructure commitment is staggering. The Reuters report on the planned layoffs pointed to rapidly rising AI-related infrastructure costs, including data center spending plans discussed in company guidance and filings. That figure dwarfs the capital expenditure plans of most technology companies and reflects a conviction inside Meta’s leadership that the race toward artificial general intelligence will be won by whoever builds the largest and most capable computing infrastructure first. The company’s Q3 2025 earnings release, filed with the SEC, stated explicitly that 2026 expense growth was expected to be significantly faster than 2025, driven primarily by infrastructure costs including cloud expenses and depreciation.
Much of this money is directed toward custom data centers designed to handle AI training workloads, with dense clusters of accelerators and high-bandwidth networking. Meta’s leadership has framed this buildout as a prerequisite for training frontier-scale models that can power everything from recommendation engines to new generative products. The company is effectively betting that the marginal value of better models will more than offset the enormous capital required to train and serve them.
Accounting Changes Signal a Long-Term Infrastructure Bet
Beyond the headline spending figures, Meta’s regulatory filings reveal quieter but telling adjustments to how the company accounts for its physical assets. The quarterly report for the period ended September 30, 2025, filed with the SEC, documented changes to server and network asset useful lives effective January 1, 2025. Extending the useful life of servers and networking equipment reduces annual depreciation charges, which in turn lowers reported expenses even as the company spends more on new hardware.
This kind of accounting shift matters because it provides a financial cushion. By spreading depreciation over a longer period, Meta can absorb the impact of its massive capital expenditure ramp without immediately cratering operating margins. It may also reflect a view that certain server and network assets can be used longer than previously estimated, though the filing does not attribute the change to any single technical or product strategy. The 10-Q filing detailed the cost structure and drivers tied to data centers and technical infrastructure, underscoring that these are not one-off purchases but part of a sustained buildout that will define Meta’s cost base for years.
Brand positioning reinforces this long-term orientation. Corporate materials featuring the Meta logo and visual identity emphasize a shift from social networking toward a broader technology platform focused on AI, mixed reality, and infrastructure. The messaging aligns with the company’s financial disclosures: Meta wants investors to see it less as an advertising-dependent social media business and more as a foundational player in the next wave of computing.
The Talent Squeeze Behind the Numbers
Most coverage of Meta’s layoffs has focused on the human cost and the strategic contradiction of firing AI workers while chasing AI supremacy. But there is a less obvious consequence that deserves attention: the potential ripple effect on the broader AI labor market. When a company as large as Meta simultaneously lays off hundreds of experienced AI researchers and engineers while aggressively recruiting for a single superintelligence-focused lab, it creates a specific kind of talent distortion.
The workers leaving FAIR and other AI teams are not junior employees. Many hold deep expertise in machine learning, model optimization, and infrastructure engineering. Their sudden availability could benefit smaller AI startups and competing labs in the short term, especially those that cannot match Big Tech salaries but can offer equity and technical autonomy. At the same time, Meta’s concentrated hiring for its superintelligence effort drives up compensation for the most specialized researchers working on frontier model training, alignment, and large-scale compute orchestration.
The net effect could be a widening gap between what top-tier AI talent commands and what mid-tier practitioners earn, making it harder for companies outside the biggest spenders to compete for the people they need most. Universities and public-sector labs may also feel the strain as they struggle to retain faculty and researchers who can now choose between a suddenly deep pool of industry roles. Over time, this imbalance risks concentrating cutting-edge AI research even further inside a handful of firms with the capital to support massive compute budgets.
Investors Approve, but Risks Remain
Wall Street’s initial reaction to Meta’s spending plans was positive. Shares surged after the company announced its 73% capital expenditure increase, a sign that investors are willing to tolerate near-term margin pressure in exchange for the possibility of dominant AI platforms later. The company’s guidance, reflected in its SEC filings, frames the higher expense base as a necessary step toward building durable competitive advantages in infrastructure and models.
Yet the risks are substantial. Meta is committing tens or hundreds of billions of dollars to technologies whose commercial payoff remains uncertain and whose regulatory environment is still evolving. If superintelligence timelines stretch out, or if new rules restrict how large models can be trained and deployed, the company could find itself locked into an expensive infrastructure footprint that is harder to monetize than expected. Meanwhile, deep workforce cuts carry their own costs in morale, institutional knowledge, and public perception.
For now, Meta’s strategy is clear: sacrifice breadth of projects and headcount in order to finance an unprecedented push into AI infrastructure and research. Whether that gamble pays off will depend not only on the company’s technical progress, but also on how the broader ecosystem of regulators, competitors, and talent responds to a world where one firm is willing to spend at a scale few others can match.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
