META's $145B Capex for 2026 Is 2x Prior Guidance Paired With 8,000 Job Cuts

Meta's announcement of a record $145B capex budget for 2026, paired with an 8,000-person workforce reduction, encapsulates a broader theme in Big Tech: the transition from a software-and-services paradigm to an AI-infrastructure paradigm. The workforce cuts are not austerity; they are strategic redeployment. Meta is exiting or consolidating business lines that do not feed the AI training and inference buildout, and freeing up cash and headcount to fund the GPU arms race.

The $145B capex figure is staggering in absolute terms, it is larger than the entire market cap of most S&P 500 companies. For context, Meta's prior guidanceCompany-issued forecasts of future financial performance. was around $40B to $65B annually. The 2x+ ramp reflects an existential bet that the future of Meta's advertising business depends on having world-class AI models to target and personalize content. The capex is directed toward data centers, GPUs (primarily Nvidia H100s and next-gen chips), and networking infrastructure to support training and serving large language models at scale.

The workforce reduction of 8,000 is material but not unprecedented for a company of Meta's size (which has roughly 70,000 employees post-restructuring). The cuts are targeted: business operations, some product teams, and roles that were added during the metaverse spending spree are being eliminated. Importantly, engineering and AI-focused teams are reportedly being spared or expanded. This selective pruning suggests Meta's leadership views the AI capex as non-negotiable, and is willing to absorb near-term earnings headwinds (from capex depreciation and lower headcount productivity) to position for long-term dominance.

Market reaction has been mixed. Meta stock is up on the narrative that capex is being deployed efficiently and that workforce reductions signal discipline. However, the longer-term concern is ROI: does $145B in annual capex actually generate incremental advertising revenue, or does it simply increase Meta's cost of goods sold and compress margins? Skeptics point out that Nvidia and other pure-play infrastructure beneficiaries are the true winners in a capex arms race, not the platforms themselves. Meta's margin compression from depreciation alone could be 3-5% if capex runs at $145B and amortizes over 5-7 years. Other hyperscalers (Google, Microsoft, Amazon) are likely to announce similarly aggressive capex targets in coming weeks, which could accelerate the narrative of a capex plateau and margin-compression risk.