Meta Platforms reported second-quarter 2026 earnings on July 29, 2026, with revenue growing 28% year-over-year to $60.8 billion — a strong top-line result that came alongside two much less comfortable numbers: earnings per share fell 13% year-over-year to $6.18, and free cash flow collapsed nearly 91% from the same quarter a year earlier, down to $784 million. On the earnings call, CEO Mark Zuckerberg described what he called a “large enterprise opportunity” spanning AI agents, APIs, and compute, and separately told investors he expects billions of people to have their own personal AI agents within five years — a specific, dated prediction about how central AI assistants will become to ordinary consumer life, not just enterprise software.
The spending behind that vision is substantial and growing. Meta raised the lower end of its 2026 capital expenditure guidance to $130 billion, from $125 billion, while keeping the upper end of its range at $145 billion. Part of that build-out includes a newly announced $14 billion data center in El Paso, Texas, developed in partnership with BlackRock — one of several large infrastructure partnerships Meta has struck this year to fund AI compute capacity without carrying the entire capital burden on its own balance sheet.
Selling Compute Instead of Just Building It
In a detail that underscores how tight AI compute supply has become industry-wide, Zuckerberg said Meta is receiving offers to rent out its own computing capacity at a significant premium — and the company is reportedly in early talks to rent AI computing power directly to Anthropic, one of its own competitors in the broader AI race. That a company spending well over $100 billion on its own AI infrastructure is simultaneously fielding demand to lease that same capacity to a rival lab is a concrete sign of just how constrained global AI compute supply remains, even as every major player keeps expanding it.
Investors Aren’t Fully Convinced Yet
The market’s reaction to all of this has been notably cooler than Zuckerberg’s own framing. Meta’s stock extended an 11-day losing streak following the earnings release, and Zuckerberg’s own net worth fell by nearly $18 billion in a single trading day as Wall Street grew more skeptical of the pace of Meta’s AI spending relative to the near-term cash returns it’s actually producing — the 91% free-cash-flow decline being the most direct evidence of that gap between spending and current cash generation.
Zuckerberg’s Public Pitch: AI for Everyone, Not a Few Labs
On July 28, 2026, Zuckerberg published an op-ed in the Wall Street Journal laying out his broader philosophy on AI development, arguing that “superintelligence” — AI systems that meaningfully exceed human capability — should be distributed as widely as possible rather than concentrated inside a small number of institutions, which he described as the greatest risk AI poses to humanity. He framed his preferred approach, which he called “personal superintelligence,” around three stated principles: individual empowerment as the source of prosperity, invention as AI’s primary purpose, and a balance of power as the actual foundation of safety. In the same piece, Zuckerberg said he was surprised that some of the industry’s most prominent leaders — naming Anthropic’s Dario Amodei and OpenAI’s Sam Altman — had offered public views on AI’s future that he described as “so filled with doom,” a direct, attributed characterization of his competitors’ own public statements rather than an independent claim about their character. The op-ed landed the same day more than 1,000 employees across OpenAI, Anthropic, and other frontier labs signed a public letter urging governments to build tools for deliberately slowing AI development — a striking contrast Zuckerberg’s own piece was widely read as responding to directly.
What This Means for Philippine Founders
Zuckerberg’s five-year personal-AI-agent prediction is worth treating as a genuine strategic bet rather than settled fact when a Philippine founder is deciding how much of their own product roadmap to build around consumer-facing AI agents specifically — Meta’s own investors are visibly still deciding whether they believe it enough to tolerate a 91% cash-flow hit to fund it. The Meta-Anthropic compute-leasing talks are a more immediately useful signal: even well-funded AI labs are capacity-constrained enough to consider renting infrastructure from a nominal competitor, which means Philippine startups budgeting for AI compute costs should expect continued tightness and price volatility in the near term, not a steadily falling cost curve they can simply assume will hold. Zuckerberg’s public sparring with Amodei and Altman over AI’s risks versus its distributed benefits is also a reminder that the industry’s most influential leaders are actively, publicly disagreeing about the technology’s own trajectory — a genuine, unresolved debate playing out between the people building the technology, not a settled consensus a founder can simply defer to. Whichever view eventually proves closer to right, the size of the bet Zuckerberg is making on his own — up to $145 billion this year alone — means Meta’s shareholders, and by extension its own workforce and product roadmap, are directly exposed to the outcome of that debate whether or not they share his specific conviction.
Share this article