Meta’s AI Bill Soars Past $130 Billion as Cash Flow Collapses and Skeptics Multiply

Meta Platforms posted another quarter of solid ad revenue growth. Yet its free cash flow cratered. The gap between promise and payout in its artificial intelligence push grows harder to ignore.

Second-quarter revenue climbed 28% to $60.8 billion. That beat expectations. Earnings per share came in at $6.18, missing forecasts of $7.22. The real shock landed in the cash statement. Free cash flow fell to $784 million from $8.55 billion a year earlier. One number tells the story. Capital spending keeps climbing.

The company now expects 2026 capital expenditures between $130 billion and $145 billion. It lifted the lower end of that range. Mark Zuckerberg, Meta’s chief executive, doubled down. He spoke of bets on personal agents and new businesses. Returns? They remain years away for many projects. But the ad machine hums along, powered by AI recommendations that delivered 5% more conversions on Instagram and 3% on Facebook in recent periods.

Investors didn’t celebrate. Shares slid after the report. Zuckerberg’s fortune dropped nearly $18 billion amid an 11-day losing streak. The Wall Street Journal laid out the case days earlier. Meta stands as the most financially stretched among big tech names. It relies almost entirely on advertising. It lacks a proven record building successful new businesses beyond social media. Higher costs, borrowing limits and investor pushback compound the pressure.

And the spending shows no sign of slowing. Meta plans to double its computing power to 7 gigawatts this year and again to 14 gigawatts next. It operates or builds 32 data centers worldwide. Earlier guidance pointed to as much as $65 billion in 2025 capital expenditures. That figure has ballooned. Big Tech overall eyes more than $700 billion in AI-related outlays for 2026. Meta’s share weighs heavy.

Contrast the results with Microsoft. The software giant showed strong enterprise demand and a bigger pipeline of future revenue. Its remaining performance obligations rose sharply. Meta’s simpler, ad-focused model delivered AI payoffs faster in recommendations and content ranking. Yet Microsoft’s cloud growth eased some spending concerns among its investors. Meta’s cash conversion weakened. Margins faced pressure. The market drew a line.

Earlier this year Meta overshadowed Microsoft in one earnings cycle. Its first-quarter revenue guidance signaled 30% growth to around $55 billion at the midpoint. AI lifted ad clicks. Operating income beat forecasts. The Facebook parent’s results highlighted advantages of its business model over Microsoft’s heavier bet on convincing office workers to adopt chatbots. But that was then.

Now scrutiny intensifies. A Reuters report captured the shift. Free cash flow collapse overshadowed revenue strength. Zuckerberg’s vision for superintelligence and AI agents demands infrastructure at massive scale. The ad business remains the lifeblood. Improvements there prove real. They may not match the scale of outlays.

Analysts voice measured concern. Emarketer senior analyst Minda Smiley noted that AI-driven investments into Meta’s advertising business continue to pay off. But Meta’s exorbitant spending on its AI visions will continue to draw questions and scrutiny from investors eager to see returns. Jeremy Goldman, principal analyst at eMarketer, pointed to the central question heading into recent periods. Whether Zuckerberg’s tens of billions in AI infrastructure bets will pay off.

Meta’s open-source Llama models add another layer. Llama 4 drew lackluster reception. It triggered staff departures. The company responded with a high-stakes talent war. It doled out more than $100 million in pay packages to researchers from rivals. It took a nearly $15 billion stake in Scale AI, valuing the data-labeling startup at $29 billion post-deal. Scale’s CEO Alexandr Wang moved into a top role leading Meta’s AI efforts. Google, once Scale’s largest customer, planned to split.

These moves buy talent and data. They don’t yet deliver clear monetization outside advertising. Meta released Muse Spark in April 2026. Its first major model in over a year. Stock rose on the news. Yet it followed delays in rolling out a flagship “Behemoth” model. Engineers struggled to improve capabilities enough to justify release. Internal questions mounted about the direction of multibillion-dollar investments.

So what does Meta get for all this cash? Better ad targeting. Higher engagement on Reels and other formats. Incremental revenue gains that analysts describe as meaningful but not transformative at current spend levels. Zuckerberg argues the infrastructure builds toward personal AI agents and entirely new product categories. Those could generate fresh revenue streams. Timing remains uncertain. Costs rise faster than visible payoffs.

Comparisons to Amazon surface in investor discussions. Amazon monetizes infrastructure across retail, advertising, AWS and internal use. Every incremental dollar finds a home. Meta asks investors to trust that similar returns will appear. The path looks less clear. X conversations among investors reflect the divide. Some call the quarter ugly yet acknowledge fast growth and AI improvements in ads. Others see pressure points separating winners from the rest among big tech names.

BlackRock-related moves offered one structural response. Meta engineered options to ease cash burn concerns as capex guidance approached $125 billion to $145 billion. Quarterly spend could hit $35 billion to $42 billion in coming periods. Analysts anticipate possible upward revisions if the race for advanced AI accelerates.

Meta isn’t alone. Alphabet faces cash burn alarms. Microsoft eased some worries with cloud beats. The entire sector confronts the question of returns on hundreds of billions. For Meta the stakes appear highest. Its ad dominance funds the experiment. Regulatory pressures, including massive potential penalties from U.S. states on youth safety issues, add to the mix. Legal costs factored into recent results.

Zuckerberg keeps betting big. He shortened timelines on superintelligence in past comments. The company pushes custom AI chips into production. Computing capacity doubles repeatedly. Yet the Wall Street Journal argument gains traction. The case for unchecked spending weakens with each quarterly report that shows cash flow evaporation.

Short-term ad wins buy time. Longer-term bets on agents and new AI products carry the weight of proof. Investors want evidence the infrastructure spend translates into durable competitive advantage and revenue growth that outpaces the burn rate. So far the ad engine delivers. The rest stays speculative.

That tension defines Meta today. Strong top-line momentum. Collapsing cash generation. Soaring capital commitments. And a CEO convinced the payoff justifies the pain. Markets have soured for now. Whether they come around depends on execution in the quarters ahead. The numbers don’t lie. The bill keeps growing.