Meta Is a Strong Buy Despite Free Cash Flow Collapsing Under a $145 Billion AI Build

Meta's stock is reacting to cash intensity, not the collapse of the core business

The market is focusing on the wrong variable. After Meta reported 28% Q2 revenue growth, the stock still fell 8.6 percent in the first-quarter follow-through, and after this week's capex update it dropped another about 5% in extended trading. Once investors anchor on weaker free cash flow and the company's $12.4 billion of quarterly free cash flow against $19 billion of capex, every new infrastructure headline gets read as more bad news.

The market is overweighting cash burn

Meta's AI build is undeniably expensive. The company now expects $130 billion to $145 billion of 2026 capital expenditure, and costs and expenses rose 55% year over year in Q2. That pressure is real. But the reaction still looks harsh relative to operating momentum: revenue grew 28%, ad impressions delivered across our Family of Apps increased by 14% year-over-year, and average price per ad increased by 12% year-over-year. For now, the market is treating AI more like a burden than an upgrade to the advertising engine.

Why the bear case still has a real structural argument

The clearest bearish point is also the simplest: unlike Alphabet, Amazon, or Microsoft, Meta lacks a cloud arm to cushion infrastructure spending. That gives investors a reason to be cautious about a bet with an unclear, slow payoff. But that caution is also what creates the opportunity. If Meta can show that AI is improving the core ad business before the market believes it, the stock has room to rerate from here.

AI spending is changing how investors value Meta

What changed is the valuation lens. Meta is no longer being judged only on product cycles and growth rates. It is being priced more like a company in an industrial capex cycle, where duration, input costs, and long-term commitments matter as much as revenue growth.

Locked commitments are making the payback period harder to ignore

The mechanism is straightforward: capex is no longer just about adding capacity. It is also about component pricing and contracted obligations. Meta's $10 billion capex raise, driven by memory costs shows how input inflation can expand the bill before revenue effects are visible. The company also signed about $107 billion in new multi-year infrastructure commitments in a single quarter. That turns AI spending from a flexible software-style budget into something closer to a long-duration capital program.

Once spending is locked in, investors stop asking only whether AI is working. They start asking how long the cash returns will take.

The sector-wide pressure is now part of Meta's valuation

This is where the broader pressure becomes visible. A Reuters analysis of LSEG consensus estimates found hyperscalers could spend more combined on capital expenditures than they generate in free cash flow by 2027. That does not prove the investments will fail. It does show why investors are shifting from a pure growth-multiple framework to a cash-conversion framework.

For Meta, that shift feels more painful because the company still depends on one main monetization engine. The company says At the heart of Meta is its advertising business, and there is still no cloud arm to cushion infrastructure spending. Bears can therefore frame the story around rising hardware costs, long-dated contracts, and a single core cash generator.

Meta's ad engine is still doing the heavy lifting

The cash squeeze is real. But the bull case does not require investors to ignore it. It requires them to recognize that Meta's core advertising business is still strong enough to fund much of the buildout.

Q2 showed that the operating engine is still producing

Q2 revenue reached $60.80 billion. Meta also reported ad impressions delivered across our Family of Apps increased by 14% year-over-year and average price per ad increased by 12% year-over-year. That combination matters. It suggests AI is not sitting outside the business as a pure cost item; it is helping sell more ads, at higher prices, inside a platform that is still growing.

That is different from the future-payoff story bears want to impose. If recommendations, targeting, and ad tools are already improving engagement and pricing, then the AI build is at least partially self-funding through the core business. Meta has also pointed to broader advertiser adoption of its AI tools, with 8 million advertisers using generative AI tools. Bulls do not need proof of a new enterprise platform yet. They need proof that AI is making the existing cash engine more productive, and the quarter still supports that case.

Where the bear case still holds up

The bear rebuttal is also easy to see. Operating margin fell to 31% from 43%, and management said AI powers more precise ad targeting and recommendation systems to boost user engagement while Reels continues to compete with TikTok and YouTube Shorts. Bears will argue that those engagement and pricing gains may still be too slow to justify the size of the infrastructure stack, especially with no cloud arm to cushion infrastructure spending.

The signals that matter most now

• Confirmation: If ad impressions, average price per ad, and advertiser tool adoption remain firm, the market has to keep allowing that this capex is improving the core business.
• Confirmation: If AI keeps showing up in targeting, recommendations, and product competitiveness, the spending starts to look more like moat-building than pure compute cost.
• Break signal: If monetization weakens while margin pressure remains, bears win the central argument: the payoff is still too slow for the multiple the market wants to assign.

The investment setup: own the operating business first

The clearest read is to own Meta for the next few quarters based on the operating business, not the balance-sheet headline. The market is still treating the narrowed capex range largely as another warning rather than evidence management is forcing clarity around an already locked infrastructure base. That creates the opening. Investors are applying a sector-wide cash-conversion scare-where hyperscalers could outspend free cash flow by 2027-to a company whose core ad engine still appears to be working.

What to watch next

• Management continues linking spending to more precise ad targeting and recommendation systems, not just raw compute capacity.
• The ad model stays healthy, with ad impression growth and pricing momentum holding up.
• The company remains exposed to roughly $107 billion in new multi-year infrastructure commitments while monetization softens.

That is the edge here. Fear has pushed Meta into a cash-flow frame before consensus has fully priced in the possibility that AI is improving ad ROI. If that operating evidence holds, the multiple can close faster than the headline numbers change.