Meta raises 2026 AI data center capex forecast to $130-145bn, shares slide

Meta Platforms has lifted its full-year capital expenditure guidance for the second time in 2026, signaling an aggressive push into AI infrastructure as the company races to expand its data center footprint. The updated range of $130bn to $145bn represents a significant escalation from the $115bn to $135bn outlined at the start of the year and the $125bn to $145bn range set in April. The revised outlook underscores the intensifying industry-wide competition to secure compute capacity for large-scale AI workloads.

The company's latest earnings report, released this week, revealed that the increased spending is driven by rising costs associated with AI data center buildouts, including memory components for AI servers. The announcement echoes a similar move by Amazon, which this week raised its capex guidance by $20bn, citing the same cost pressures. While Amazon's increased expenditure was met with investor approval and a share price rally, Meta's update drew a markedly different reaction.

Meta's shares fell approximately ten percent following the announcement, as investors focused on the company's deteriorating free cash flow. For the quarter, free cash flow dropped to $784m, its lowest level in five years, compared with $8.5bn in the same period last year. The sharp decline reflects the heavy upfront costs of building and equipping AI data centers, a trend that has become a defining feature of the current technology investment cycle.

During the earnings call, CEO and founder Mark Zuckerberg defended the scale of the investment. "I get that this is a big investment and it's a big bet," he said. "We see the technology working. We're happy with the trajectory of the lab. I'm excited about the products that are coming. And we believe that this is going to be a big thing." His remarks were aimed at reassuring investors that the company's spending is aligned with a clear strategic vision for AI-driven growth.

Meta also reiterated its interest in becoming a cloud provider, a move that would position the company as a direct competitor to established players like Amazon Web Services and Microsoft Azure. Zuckerberg noted that Meta is "getting a lot of offers for compute at a significant premium over what we paid for it," suggesting that the company's data center capacity could become a lucrative revenue stream. Reports indicate that Meta is in early talks to rent compute capacity to Anthropic in a deal that could be worth as much as $10bn. The company recently hired David Brown, a former AWS cloud lead, to oversee its cloud ambitions.

For the quarter, Meta reported revenues of $61bn, up 28 percent year-over-year, while profits fell 14 percent to $6bn. The company forecast third-quarter revenue of between $61bn and $64bn, with the midpoint falling below Wall Street expectations of $63.1bn. The guidance, combined with the capex increase, has raised questions about the near-term profitability of Meta's AI investments, even as the company positions itself for long-term leadership in the AI infrastructure market.

The broader implications for the data center industry are substantial. Meta's elevated spending is part of a wider trend among hyperscale operators, who are collectively pouring hundreds of billions of dollars into AI-ready infrastructure. This wave of investment is expected to drive demand for advanced cooling systems, high-bandwidth networking, and energy-efficient designs, while also intensifying competition for power and land. For suppliers and developers, the sustained capital flow signals a multi-year expansion cycle, though it also raises concerns about the financial strain on operators as they balance growth with shareholder expectations.