META Raises Capex Guidance to $145B While Cutting 8,000 Jobs, 10% of Its Workforce

Meta's strategic recalibration represents one of the most aggressive corporate transformations in recent years, simultaneously cutting labor and doubling down on AI infrastructure investment. The company is eliminating roughly 10% of its workforce (8,000 employees), canceling 6,000 open roles, and shifting 7,000 employees into AI-focused positions. Crucially, the company guided up to $145 billion in capital expenditure, a massive increase from prior guidanceCompany-issued forecasts of future financial performance. and a signal that despite workforce reductions, Meta is committed to owning end-to-end AI infrastructure rather than relying on Nvidia and other vendors.

The narrative is not austerity but workforce reallocation. Meta's leadership has concluded that the company is overstaffed in certain operating divisions, particularly in non-engineering functions, but critically understaffed in AI research, model training, and infrastructure engineering. By cutting lower-productivity headcount and redeploying top talent into AI-intensive roles, Meta is restructuring its cost structure while maintaining or increasing investment in the capabilities it believes will determine competitive advantage in the next era of technology.

The $145 billion capex guidanceCompany-issued forecasts of future financial performance. is striking because it far exceeds traditional cloud capex norms and signals that Meta intends to build significant proprietary AI models and training infrastructure rather than outsource all compute to Nvidia GPUs. This is consistent with themes from Amazon, Google, and Microsoft, all of which are reducing reliance on Nvidia by developing custom silicon and in-house accelerators. Meta's scale, billions of active users across social and messaging platforms, creates a unique use case for building proprietary foundation models and recommendation systems that can leverage vast amounts of internal data.

The risk is execution. Shifting 7,000 employees into new roles requires retraining, organizational change management, and hiring of specialized AI talent in competitive labor markets. If Meta missteps on model development or if custom silicon efforts fail to deliver meaningful efficiency gains, the company will have sacrificed operational flexibility and head count without corresponding upside. Additionally, shareholders may initially interpret workforce cuts as a sign of revenue deceleration, even though management's tone suggests confidence in the business.