Meta (NASDAQ: META) delivered a Q1 result that most managers would frame and hang on the wall. Revenue surged 33% year-over-year, passing $56 billion. Gross margins held at nearly 82%, while average revenue per user climbed 28.2% to hit $15.82. Everything pointed to the continuation of the monetization streak that turned Meta into one of the best advertising machines in corporate history.
Yet, investors were unimpressed. The stock remains down around 6.5% year-to-date.
Why? Because Wall Street stopped caring about the present the moment Mark Zuckerberg opened the capital expenditure firehose.
Capex Escalation
Meta has lifted its 2026 capex guidance to between $125 and $145 billion - nearly double last year's spending pace. As recently as 2023, Meta's entire annual capex totaled around $28 billion. Today, that figure isn't far from the spread between the low and high end of its guidance.
That race is creating a massive disconnect between the company's current fundamentals and the market's tolerance for uncertainty. The core ad engine is still humming. Ad impressions rose 19%, average ad prices climbed 12%, and North American ARPU has nearly doubled over the past three years. By traditional metrics, this business looks elite.
But investors increasingly want a cleaner answer to one question- where exactly is the AI revenue?
Zuckerberg's response on the earnings call didn't exactly calm nerves. When asked about milestones he's watching to measure the return on investment over the next 12 to 24 months, he called it "a very technical question" before reverting to Meta's historical playbook - scale the product first, then monetize later.
That strategy worked for Facebook and Instagram, but spending $145 billion before proving monetization is a next-level wager.
Unlike Alphabet and Amazon, which pointed to AI-driven growth in cloud revenue, Meta largely relies on advertising optimization and future promises.
What makes investors more jittery is the operational contradiction in the background. Meta is simultaneously cutting roughly 10% of its workforce - around 8,000 jobs - while pouring unprecedented sums into AI infrastructure. The latest layoffs are the largest restructuring since the "Year of Efficiency" purge in 2023.
The firm is technically laying off humans to buy more machines. And those machines are getting more and more expensive.
Embracing The Chipflation
According to BlackRock's recent outlook, the AI boom will initially be inflationary before eventually turning deflationary. The early phase of AI adoption isn't about productivity gains yet - it's about hyperscalers running a global bidding war for chips, power, cooling systems, and data centers. Meta is now living inside that inflation cycle.
CFO Susan Li cited higher "component pricing" and data center component costs as the main reason for the capex increase. BlackRock's data explains why. DRAM memory prices have skyrocketed 17-fold over the past year. It was an almost absurd reversal from the decades-long decline in semiconductor pricing associated with Moore's Law.
The result is chipflation - a demand shock so intense that even trillion-dollar companies are struggling to secure infrastructure without paying massive premiums.
Meta is fighting back by diversifying its hardware stack. Zuckerberg noted that the company is deploying more than 1 gigawatt of custom silicon co-developed with Broadcom, while also scaling commitments across AMD chips, Nvidia systems, AWS Graviton processors, and other compute providers.
The strategy is partly about performance, but mostly about escaping dependence on the same constrained supply chain everyone else is fighting over.
While AI may eventually deliver the lower costs, higher productivity, and structurally better margins across the economy as BlackRock believes, the deflationary payoff still seems far away.
Until then, Meta must first navigate and survive the most expensive infrastructure arms race the tech sector has ever seen.