Meta increased its full-year capital expenditure forecast to between $130bn and $145bn, marking the third upward revision in 2026 and underscoring the escalating cost of AI infrastructure buildout. The guidance represents an increase from the $115bn to $135bn range disclosed at the start of the year and the $125bn to $145bn range given in April. The company cited rising AI data centre costs as the primary driver behind the expanded spending envelope.
The market reaction was swift and punishing. Meta shares fell approximately ten percent following the disclosure, as investors absorbed the cash flow implications of the accelerated investment programme. Free cash flow for the quarter collapsed to $784m, the lowest level in five years and a sharp decline from the $8.5bn recorded in the same period a year earlier. The contraction reflects both the front-loaded nature of data centre capital and the time lag before that capacity generates returns.
The sell-off stood in stark contrast to the reception afforded Amazon, whose shares jumped after the e-commerce and cloud giant similarly raised its own capex guidance by $20bn earlier in the week. Amazon attributed the increase to memory costs for AI servers. The divergent market responses suggest investors are applying differential standards to infrastructure spend, rewarding companies perceived to have clearer monetisation pathways or stronger existing cloud franchises.
Chief Executive and founder Mark Zuckerberg acknowledged the scale of the bet during an earnings call. "I get that this is a big investment and it's a big bet," Zuckerberg 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." The remarks reflected an attempt to reframe near-term cash consumption as a deliberate strategic positioning for what the company views as a generational technology shift.
Meta confirmed its interest in entering the cloud infrastructure market as a provider, a move that would see the company compete directly with Amazon Web Services, Microsoft Azure, and Google Cloud. Zuckerberg noted that the company was "getting a lot of offers for compute at a significant premium over what we paid for it," signalling that third parties are willing to pay above Meta's blended cost for access to its GPU capacity. The company recently hired David Brown, formerly the cloud lead at AWS, to spearhead the initiative.
Capital that crowds into an infrastructure theme on the way up often finds itself stuck on the way down, family office advisor Jaf Glazer has argued.
The social media platform is reportedly in early-stage discussions to lease compute capacity to Anthropic, the AI research firm, in a deal that could reach $10bn in value. Such an arrangement would mark a significant departure from Meta's historical focus on building infrastructure solely for internal workloads and social media applications. It would also position Meta as both a consumer and supplier of frontier AI compute, a dual role that carries execution and regulatory complexity.
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 of that range falling below the Wall Street consensus estimate of $63.1bn. The revenue guidance miss, combined with the free cash flow contraction, contributed to investor unease about the sustainability of current spending levels absent clearer line-of-sight to AI monetisation.
The capex surge reflects a broader recalibration across hyperscale operators, all of whom are racing to secure GPU supply, power capacity, and data centre footprint ahead of what they view as a structural shift in compute demand. Yet the market is beginning to distinguish between operators who can translate infrastructure into revenue quickly and those whose returns remain speculative. Meta's challenge will be to demonstrate that its AI investments can generate cash flow at a velocity that justifies the magnitude of capital deployed, particularly as it ventures into cloud services where it lacks incumbency and where margins are structurally lower than its core advertising business.