AI Stocks Are Selling Off, but Meta Plans to Spend Up to $145B: Will Wall Street Keep Funding the AI

The AI trade is facing one of its biggest tests yet.

Last week, the Philadelphia Semiconductor Index fell roughly 11%, as investors grew increasingly concerned about AI infrastructure financing, elevated valuations and whether massive capital spending could eventually generate sufficient returns.

Yet just as Wall Street begins questioning the economics of AI, Meta is preparing to report its Q3 earnings on October 21.

And the company has already outlined an extraordinary spending plan: $130 billion to $145 billion in capital expenditures for 2026.

That creates an interesting contradiction.

AI infrastructure stocks are getting punished for spending too much, while Meta is preparing to defend one of the largest AI investment programs in the industry.

So the real question going into earnings is no longer simply whether Meta can deliver another strong quarter.

It is whether investors are still willing to pay for a business model built around spending heavily today to generate larger returns tomorrow.

Meta is certainly not an AI startup without an established revenue stream.

In Q2, the company generated $60.8 billion in revenue, up 28% year over year. Advertising impressions increased 14%, while the average price per ad rose 12%. These numbers suggest Meta's core advertising business remains strong, with AI-powered recommendations and advertising optimization helping improve monetization.

But another set of numbers tells a more complicated story.

Meta generated approximately $31.86 billion in operating cash flow during Q2. However, capital expenditures, including principal payments on finance leases, reached $31.08 billion.

That left just $784 million in free cash flow, compared with approximately $8.55 billion in the same quarter last year.

This does not mean Meta has suddenly stopped making money. The company still generated roughly $18.78 billion in quarterly operating income.

Instead, it shows just how aggressively Meta is reinvesting its cash into AI infrastructure.

And that exposes the central tension in Meta's investment story:

AI is already helping its advertising business generate more revenue, but building the infrastructure needed for the next stage of AI is consuming an enormous amount of cash.

For investors, this is no longer just about whether AI demand is real.

It is about whether the incremental revenue created by AI can eventually justify spending hundreds of billions of dollars on GPUs, servers, data centers, networking and electricity.

There is another important distinction between Meta and some of the AI infrastructure companies currently under pressure.

Companies such as CoreWeave depend heavily on infrastructure financing and long-term compute contracts to support expansion. Rising borrowing costs, equipment depreciation and customer concentration can therefore become major concerns.

Meta faces many of the same infrastructure cost pressures, but it has something different: an established advertising business generating substantial operating cash flow.

That gives Meta more flexibility to finance AI investment from existing operations, although the scale of its spending still creates meaningful financial risk.

The next question is whether Meta can build new AI revenue streams beyond advertising optimization.

AI already plays an important role in improving recommendations across Facebook and Instagram. But investors also want to know whether new AI products, assistants and enterprise tools can become independent sources of revenue.

Those are two very different investment propositions.

Using AI to improve an existing $60 billion-plus quarterly business is one thing. Building entirely new AI businesses large enough to justify another $100 billion-plus in annual infrastructure spending is another.

And that is why this week's earnings report matters well beyond Meta itself.

If Meta can demonstrate that stronger advertising monetization is supporting continued AI investment, it may help reassure investors across the semiconductor and infrastructure supply chain.

But if spending continues growing faster than revenue and free cash flow, Wall Street may become less tolerant of aggressive AI investment plans.

Tiger View

Tiger thinks Meta's upcoming earnings could become an important test of whether the AI trade is shifting from rewarding capital spending to demanding measurable returns.

Earlier in the AI cycle, the market focused heavily on GPU demand, HBM shortages and hyperscaler CapEx.

Now investors are asking tougher questions.

Who is actually generating cash flow from AI? Which companies can finance expansion internally? And how much future growth has already been priced into valuations?

Meta occupies a particularly interesting position because it is both a major buyer of AI infrastructure and an established business already monetizing AI through advertising.

That makes it a useful test case for the broader industry.

Tiger would watch three signals in the October 21 report:

First: Can advertising revenue and AI-driven monetization continue growing strongly?

Second: Are operating margins and free cash flow stabilizing despite rising infrastructure costs?

Third: Will Meta maintain or raise its 2026–2027 capital spending expectations, and what evidence will management provide about future returns?

If the next phase looks like:

Stronger advertising monetization → better AI efficiency → improving profits and cash flow → clearer CapEx returns,

then Meta could help rebuild confidence in the economics of large-scale AI investment.

But if the pattern becomes:

Revenue keeps growing → CapEx grows even faster → margins and free cash flow remain pressured,

investors may become increasingly reluctant to reward high-spending AI companies with premium valuations.

This matters not only for META, but also for the companies supplying its GPUs, custom chips, advanced manufacturing and memory.

So Tiger's key question heading into earnings is not:

"Can Meta beat revenue expectations again?"

It is:

"When a company plans to spend up to $145 billion on AI infrastructure in a single year, how long will Wall Street wait for the investment to pay off?"

Related Stocks

AI & Advertising: $Meta Platforms(META)$Meta Platforms, Inc.(META)$

Watch: advertising growth, AI-driven monetization, operating margins, free cash flow and future CapEx guidance.

AI Compute: $NVIDIA(NVDA)$

Watch: whether Meta and other hyperscalers continue expanding GPU purchases despite growing scrutiny of infrastructure returns.

Custom AI Chips: $Broadcom(AVGO)$

Watch: hyperscaler ASIC demand and how customers allocate spending between general-purpose GPUs and custom accelerators.

Advanced Manufacturing: $Taiwan Semiconductor Manufacturing(TSM)$

Watch: whether sustained AI infrastructure investment continues translating into advanced-node and packaging demand.

AI Memory: $Micron Technology(MU)$

Watch: HBM and server DRAM demand, as well as how rising memory costs affect data-center economics.

AI Infrastructure: $CoreWeave(CRWV)$

Watch: long-term compute contracts, financing costs, debt obligations and free cash flow generation.

Today's Poll

Meta plans to spend up to $145B on AI infrastructure in 2026. How should investors view it?

① AI demand remains strong — investing early is the right strategy

② Meta's advertising cash flow gives it an advantage over highly leveraged AI companies

③ CapEx is rising too fast — margins and free cash flow are becoming major concerns

④ Wait for earnings — what matters most is whether AI revenue can justify the spending

For market discussion only. This is not investment advice. Markets involve risk, and investment decisions should be made carefully.

Sources: Meta Platforms Q2 2026 official earnings release (July 29), Meta Investor Relations (October 12), Reuters (October 16–19), and Financial Times (October 19). This is a pre-earnings analysis. Q3 results have not yet been announced, and forward-looking figures are not actual reported results.