Three Risks Every Meta Shareholder Should Underwrite: Capex, Reality Labs, and a $1.4 Trillion Lawsuit

Even a great business becomes a bad investment at the wrong price. And to judge whether a price is right, you have to price in what can go wrong first.

I still think Meta's underlying business is strong: revenue growing above 30%, operating margin above 40%, more than 3.5 billion daily users. But the stock fell 11% over the past year for reasons, and markets rarely do that for nothing.

The three items below are risks I believe you have to consciously accept before owning this stock. They're ordered by how much weight I give them.

1. $145 billion is going out, and nobody knows when it comes back

This is the part that makes even the bulls uncomfortable.

Meta raised 2026 capital expenditure guidance to $125–145 billion. The sums going into data centers, chips and compute infrastructure are, literally, astronomical.

And the moment money is spent, it shows up in the financials. Here's what that already looks like.

• Free cash flow is under pressure. It was roughly $50 billion last year, and that number has nowhere to go but down.
• Meta has not repurchased shares since the third quarter of 2025.
• There are reports the company is even considering issuing new stock to fund the buildout.

That last one matters most. If you believe the stock is undervalued, share issuance is the last thing you want, because your ownership gets diluted rather than concentrated. Arguing a stock is cheap while printing shares at that price is logically awkward.

The bear framing is that this is the classic tech trap: a company falls in love with a new technology, builds expensive infrastructure, and then demand doesn't show up at the scale assumed. Whether it's compute leasing, ad tools or enterprise products, if AI monetization doesn't generate revenue that justifies the spend, Meta is left holding a mountain of expensive infrastructure nobody wants.

One more layer: however good it is, the core advertising business is still tied to the macro cycle. In a recession or slowdown, ad budgets get cut very quickly. And Meta competes for attention against TikTok, YouTube and every other short-form platform. The ad machine is powerful, but it isn't immune to the real world.

I covered the backdrop of simultaneous headcount cuts and AI investment in Meta Cuts 20% of Staff While Investing $135 Billion in AI — What's the Strategy?.

2. Reality Labs is still burning $4 billion a quarter

This has been a thorn in Meta's side for years.

I'd heard the division had been shut down entirely, so I checked. It hasn't closed — it has been scaled back sharply.

In a single quarter it posted more than $4 billion in operating losses, and the revenue coming out of Reality Labs is trivial against its operating costs. In plain terms, Meta is betting billions on the future of computing, headsets, glasses and virtual environments, and right now that bet is deeply underwater.

Bulls will say Zuckerberg is playing a long game and it eventually pays off. Bears point out that 'eventually' has been the answer for several years running. I think the bears have earned that point by now. Enough time has passed.

The real reason I classify this as a risk isn't the size of the loss. It's that every dollar going into Reality Labs is a dollar that could have gone to shareholders. With up to $145 billion already committed to AI infrastructure, the company is fighting two expensive wars at once.

I'd add a counterpoint, though. The scaling back is a good sign, not a bad one — it's evidence the company can reassess its own bets. Back in 2022 when the stock was at $88, the thing the market feared most was the metaverse. The metaverse push failed, and the business got better anyway. If the lesson learned is that a big investment isn't always the right choice, that lesson is an asset.

3. Legal and regulatory exposure is enormous, and the stock barely reflects it

This is the item I think most investors spend the least time on.

Four states — California, Colorado, Kentucky and New Jersey — have sued Meta over youth addiction and safety issues, seeking roughly $1.4 trillion. That's trillion with a T.

Will Meta actually pay that? I don't think so. This is how state and federal governments typically operate. They lead with an enormous number to create fear, use it as the anchor, and make a far smaller figure look reasonable by comparison. Settlements usually land somewhere in the middle.

That doesn't mean the risk isn't real. Trials are coming, and even a fraction of that number is a meaningful hit given the current capex load. I wrote about algorithms themselves becoming the legal target in Section 230 Lawsuits: The Moment Algorithms Became Legally Liable. It isn't only Meta's problem, but Meta is the biggest target.

And it isn't only a US problem.

• EU Digital Markets Act: continuing pressure on Meta's data-driven advertising model in Europe, with direct implications for how ads are targeted and what they can be priced at.
• Former-employee class action: an ongoing suit alleging discrimination in AI-driven layoff decisions.

None of these alone is likely to be catastrophic. Combined, they form a cloud of uncertainty over the stock. The core bear argument is that the market behaves as though these risks don't exist, and I largely agree with that. Regulatory risk tends to hit the share price at the worst possible moment, precisely when it turns out worse than expected.

So what do you do with this?

Having worked through all three, I haven't changed my view on the quality of Meta's business. Last quarter's EPS of $7.31 beat expectations, gross margin runs at 82%, and the five-year average return on invested capital is 18.6%. The business is not deteriorating.

What changes isn't the verdict — it's the price. Recognizing risk doesn't mean avoiding the stock; it means demanding a lower entry price in proportion to that risk. Underwrite these three and your required return naturally rises, and when required return rises, your buy price falls.

The assumptions I actually used to calculate fair value are laid out in What Is Meta Actually Worth?, and its relative standing within big tech is in Why Meta Is the Only Real Opportunity in the Magnificent 7.

FAQ

Q: Is there any real chance Meta pays the $1.4 trillion the four states are seeking? A: I'd put it very low. Leading with a huge number is closer to a negotiating tactic, and these cases typically settle far lower. But whatever the settlement figure, cash does leave the company, and during a period of heavy capex that impact lands harder than usual.

Q: Is the pause in buybacks really that big a deal? A: It matters more as a signal than in itself. A company that believes its own stock is cheap and has spare cash normally buys it. No repurchases since Q3 2025 means cash is being allocated to infrastructure first, and combined with reports of possible share issuance, the direction is backwards from a shareholder's perspective.

Q: Would fully shutting Reality Labs help the stock? A: In the near term, yes — removing $4 billion-plus of quarterly operating losses improves earnings metrics immediately. But the division has already been cut back sharply, so the remaining benefit is smaller than it once would have been. AI capex is now the larger swing factor on the income statement.

Q: How exposed is the advertising business to a recession? A: Ad budgets are among the first things companies cut. That said, a large share of Meta's advertising is direct-response with measurable ROI, which tends to survive further into a cutting cycle than brand-oriented media. That's not immunity — it's relative resilience.