Is Meta Stock Cheap At 21x, Or Trapped By Endless AI Capex?
Shares have already been marked down, yet the capital commitment behind that decline still has no stated end date.
Meta Platforms (META) trades at $544, down 27% over the past year while the S&P 500 returned about 20%. That reflects substantial market repricing rather than a temporary pullback. The risk still worth arguing about is narrower than that decline implies: the company has committed to a build it will not put an end date on.
$60.8 Billion Of Revenue Left $784 Million Of Free Cash Flow
In the second quarter of 2026, Meta spent $31.1 billion on capital expenditure and generated $784 million of free cash flow, on $60.8 billion of revenue that grew 28% year over year. The growth is real; the cash the business generates is going into servers and data centers rather than accumulating. For now, the balance sheet absorbs that, at $90.3 billion of cash and securities against $83.7 billion of debt.
The Build Has An Address, The Payback Has No Date
The spending has a physical shape: a 1 gigawatt data center in El Paso, Texas, structured as a venture with BlackRock. Capital expenditure for calendar 2026 is guided to $130 billion to $145 billion. The return has no such definition. Management has declined to give a 2027 capital budget, saying its infrastructure planning is still highly dynamic, and has not named which of the new revenue lines scales first, whether that is compute sold directly or the business agents that more than 1 million businesses already use every week. An undated commitment of that size, carried through a single holding, is a different proposition from the Trefis High Quality Portfolio, which seeks to reduce concentration risk by diversifying beyond mega-cap technology equities.
A Multiple This Low Is No Cushion For An Undated Build
The stock is not expensive, at 21.2 times trailing earnings inside a ten-year range of 12.6 to 44.9. But a low multiple protects nothing when the earnings under it are the part at risk. GAAP operating income fell 8% year over year in the second quarter of 2026 even as revenue grew, held down by $2.4 billion of legal charges and $1.2 billion of severance from the May 2026 head count reduction; without those, it would have risen 9%. Those charges are tied to named events, though the legal ones are not finished; depreciation is not tied to any event at all. Infrastructure costs are already climbing on higher depreciation, by the company’s own account, and the assets behind that line are still being built.
The 2027 Capital Number Is What Would Settle This
The market has already priced a good deal of this risk: the shares sit at about 73% of their 52-week high, and the largest peak-to-trough fall over the past year was 32.9%. Options are priced for an unusually large move, with implied volatility at 35, in the 75th percentile of its own trailing year. What nobody can size yet is how long the build runs. A stated 2027 ceiling would turn an open commitment into one a holder can model, and another year without it extends the exposure. A second risk is growing: a $942 million penalty in a New Mexico case over harm to children, with further youth-related trials scheduled in the U.S. that management warns could be material. Yet the primary risk that remains is in the capital plan, not in whether the decline has gone far enough.
How Does Single-Stock Concentration Impact Long-Term Portfolios?
A drawdown of this magnitude highlights the difference between diversified and concentrated holdings. A position that has grown large enough to matter is worth sizing deliberately rather than by accident. What a position that size would do to your net worth is exactly what the Trefis Wealth team computes, with the same rules-based systematic discipline that runs our High Quality Portfolio. Request a free vulnerability audit of your biggest positions.