Meta’s $31.1B AI Splurge Leaves Free Cash Flow at $784M
Meta Platforms’ second-quarter results delivered a stark contrast: its core advertising business continued to grow, but capital spending on artificial intelligence infrastructure soared to $31.1 billion, nearly double the prior-year quarter and consuming virtually every dollar of operating cash. Free cash flow collapsed to just $784 million, down from an average of roughly $12 billion in the previous eight quarters, pushing the stock down as much as 10% in after-hours trading before recovering slightly to a 7% decline.
The company’s Family of Apps segment—Facebook, Instagram, WhatsApp, and Messenger—saw operating income slip to $23.4 billion from $25.0 billion a year ago, as rising infrastructure costs offset revenue gains. Meta issued $24.9 billion in long-term debt and suspended share buybacks; CFO Susan Li described a deliberate shift toward debt financing for long-lived infrastructure. Depreciation and amortization charges rose 46% to $6.4 billion, reflecting the rapid obsolescence of AI hardware.
On the earnings call, CEO Mark Zuckerberg confirmed that Meta is exploring a cloud business, revealing it has received “a lot of offers for compute at a significant premium.” However, he cautioned against selling raw computing for short-term profit, arguing that selling “intelligence”—AI-powered services layered on that infrastructure—would yield far higher margins. He promised an update soon but offered no specifics.
Why Zuckerberg Is Choosing AI Over Stock Buybacks and a Cloud Pivot
From Cash Machine to Capex Heavyweight
Meta’s advertising juggernaut historically generated enormous free cash flow, allowing the company to stockpile billions. The AI arms race has upended that model. Second-quarter capex of $31.1 billion nearly matched operating cash flow of $31.9 billion. Full-year capex is now guided at $130–$145 billion, meaning the remaining $79–$94 billion will be spent in the second half against quarterly operating cash of roughly $32 billion—implying negative free cash flow for the rest of the year. The company has turned to debt markets, issuing $24.9 billion in the quarter, and cancelled share buybacks after repurchasing over $10 billion a year ago.
The Cloud Temptation vs. Zuckerberg’s Vision
Zuckerberg’s revelation that Meta is fielding offers for its compute at a premium underscores a dilemma: Amazon, Google, and Microsoft already monetize their AI infrastructure through cloud services, while Meta has so far kept its capacity to itself. Zuckerberg characterized a cloud offering as a “side quest” and stressed that selling “intelligence”—such as the AI-powered ad system that already improved conversion by 15.7%, business messaging agents, and a mysterious 24/7 personal assistant—would generate significantly higher margins than renting bare-metal compute. By declining to sell raw compute for quick profits, he is betting that Meta’s unique trove of user data and social graph gives it an advantage in building AI services that competitors cannot replicate. Still, the cloud announcement lacked any timeline, pricing, or revenue model, leaving investors with little to value.
Debt-Fueled Buildout Raises Investor Questions
The shift to debt financing marks a historic pivot. CFO Susan Li said Meta is deliberately moving toward a greater mix of debt to match long-lived infrastructure. While this avoids large equity dilution, it increases financial risk if AI returns disappoint. Critically, when asked about 2027 capex, Li declined to provide a number, saying only that Meta remains “demand-constrained”—it has more profitable uses for compute than compute supply. That framing suggests spending could stay elevated indefinitely, but without clarity, investors are left guessing about the path back to sustained free cash flow.
What This Means for the AI Arms Race
Meta’s spending is table stakes in the hyperscaler club. The company must build massive data centers, secure power, and replace hardware every few years just to stay competitive with rivals who have cloud revenue to offset costs. The promise of AI agents for a million businesses and a personal assistant that builds a holistic profile of user health, finances, and relationships represents a vision that, if realized, could create deep new revenue streams. However, those products are not yet live, and Zuckerberg acknowledged the limits of what he could share. Until concrete service launches materialize, the market will view the capex as a high-stakes bet with no near-term payoff.
What Meta’s Strategy Means for Investors, Advertisers, and Rivals
Investors: Meta has guided full-year capex of $130–$145 billion, with the remaining $79–$94 billion concentrated in the second half. Quarterly operating cash flow of roughly $32 billion means free cash flow will be deeply negative through year-end. The $24.9 billion debt raise and halt of buybacks signal management is prioritizing infrastructure over near-term shareholder returns. The promised cloud business update will be the next critical catalyst—look for any revenue timeline, margin targets, or specifics on compute rental versus intelligence services. Absent that, the stock may remain under pressure.
Advertisers: Zuckerberg cited a 15.7% improvement in ad conversion driven by AI. If sustained, that performance should justify continued ad spending on Meta’s platforms. However, watch for any pass-through of rising infrastructure costs into ad pricing or changes in the Family of Apps operating margin. The AI agents for business messaging could eventually offer new, high-intent customer engagement tools, but no launch date was given.
Competitors: Meta’s entry into cloud services—even as a side endeavor—could disrupt pricing if it offers compute at a premium to what it paid. The bigger long-term threat, however, is the full AI stack: an ad system with superior conversion, business messaging agents, and a personal assistant that leverages Meta’s unmatched user data. Rivals should factor Meta’s continued demand-constrained posture as a signal that its AI capacity buildout will not let up.
Risk & Opportunity Assessment
| Commercial Risk | High | Capex consumes almost all operating cash, leaving free cash flow at just $784M and turning negative in coming quarters; the company has taken on $24.9B in new debt and halted buybacks, increasing financial leverage without clear near-term monetization. |
| Competitive Risk | Medium | Amazon, Google, and Microsoft already have massive cloud businesses that offset AI capex. Meta’s delay in launching a cloud service and lack of a concrete AI services timeline could allow rivals to pull further ahead in enterprise AI and compute rental, potentially eroding the advantage Meta’s data could confer. |
| Regulatory Risk | Low | No immediate regulatory actions are mentioned; however, large-scale AI infrastructure buildout and data usage for personal assistants could attract future scrutiny. |
| Reputation Risk | Medium | A 10% stock plunge and investor skepticism about runaway capex, combined with vague cloud and AI service plans, could undermine confidence in management’s capital allocation discipline and the credibility of its long-term AI narrative. |
| Technology Disruption | High | AI hardware obsolescence is rapid (depreciation up 46% to $6.4B in one quarter). A shift in AI model architectures or computing paradigms could render a portion of Meta’s massive infrastructure investments less valuable, a risk especially acute given the company is spending billions without immediate offsetting cloud revenue. |
| Commercial Opportunity | High | The AI-driven ad system already shows a 15.7% conversion improvement, and the pipeline includes paid business messaging agents and a personal assistant. If the cloud business materializes with premium compute pricing, even as a side stream, it could provide a multi-billion-dollar revenue cushion that transforms the investment case. |
Comments 0