Big Tech's Numbers Aren't Adding Up Anymore

Photo: TechCrunch

Article

Big Tech's Numbers Aren't Adding Up Anymore

Three recent stories show the same pattern: the headline figures Big Tech reports and the reality underneath them are drifting apart.

NagiOctober 8, 20264 min read

The gap between what Big Tech says about itself and what can be verified is widening. Three stories logged on this beat in the past week point the same direction: a reported revenue figure that turns out to be far smaller, a profit surge that reflects scarcity rather than durable demand, and a safety design that regulators describe as theater. The common thread is that the metrics these companies put in front of investors, regulators, and users no longer reliably describe the underlying business.

A Projection That Didn't Survive Contact

The clearest case is OpenAI. As TechCrunch reported, the AI lab's annualized revenue had previously been put at roughly $70 billion, and a new report claims it is a whole lot less than that, with the shortfall described as $20 billion against earlier projections. The precise arithmetic matters less than the mechanism. Annualized revenue is a run-rate extrapolation, and for a company whose growth curve is steep and recent, small changes in the most recent month produce large changes in the annualized figure. When that number circulates as if it were settled, investors, partners, and the press anchor on it.

The correction is not evidence that the AI business is failing. It is evidence that the headline number was doing work it could not support. For US technology companies broadly, this matters because a significant share of the current investment cycle in data centers, chips, and power contracts is justified by AI revenue projections of exactly this kind. If the projections move by tens of billions of dollars, the capital commitments built on top of them deserve a second look.

Profit Records Built on Scarcity

Samsung's preliminary third-quarter report, covered by SiliconANGLE, tells a superficially opposite story. The company said it expects an operating profit of about 107.4 trillion won, roughly $80.17 billion, exceeding 100 trillion won for the first time and described as a world record-breaking quarterly profit for the technology industry, up more than ninefold from a year earlier.

The figure is real. The question is what it measures. Record memory pricing is the product of constrained supply meeting intense demand, and that is a condition that invites its own correction: capacity gets built, prices fall, and margins compress. For US technology companies, this cuts both ways. American cloud providers and device makers are buyers in that market, so Samsung's record profit is, in part, their rising input cost. A record on Samsung's income statement is not a record on everyone else's.

The Safety Feature That Wasn't

The third story concerns product claims rather than financial ones. New York alleges that TikTok gave teens and children a placebo safety feature instead of a real one, according to TechCrunch. The state's lawsuit is one of more than two dozen cases brought by states accusing the social media giant of designing its platform to encourage addictive use among children.

The allegation is specific: that a feature presented as protective did not actually protect. This is the same divergence seen in the revenue and profit stories, moved from the income statement to the product surface. A company can put a number, or a toggle, in front of the public that does not correspond to the underlying reality. Litigation will now test whether that characterization holds, and TikTok has not conceded it here. But the pattern in the claim is what matters for the beat: disclosure and design have become the contested ground.

Why These Three Belong Together

On their face, a revenue shortfall, a memory-driven profit record, and a state lawsuit share nothing. The thread is that each involves a public-facing figure or feature that overstated the thing it purported to describe. OpenAI's annualized revenue overstated the run-rate. Samsung's record overstated the durability and the sector-wide meaning of its quarter. The TikTok feature allegedly overstated the protection it offered.

That is a problem of measurement and disclosure, and it is a Big Tech problem in particular because these companies operate at a scale where their reported metrics become inputs for other decisions. Cloud pricing, chip procurement, content moderation policy, state attorneys general, and ordinary users all calibrate against what these firms say. When those statements drift from underlying conditions, the miscalibration spreads.

What It Means in the United States

For US investors, the practical consequence is a higher burden of proof. A run-rate number that moves by $20 billion invites skepticism toward the next one, and companies seeking capital for AI infrastructure will face questions about how their figures are constructed rather than how large they are. For US consumers, the TikTok case is the more direct concern: if a platform's safety controls are alleged to be cosmetic, parents and regulators have little choice but to treat product claims as marketing until independent evidence says otherwise. For US technology firms competing with Samsung's memory business, the record profit is simultaneously a signal of tight supply and a warning that today's pricing supports tomorrow's capacity.

None of this argues that Big Tech's numbers are fabricated. It argues that the numbers have become less informative than their precision suggests, and that the burden of interpretation has shifted to the people reading them.

What to Watch

Watch how OpenAI's revenue figure is restated in subsequent reporting, and whether partners who cited the earlier number adjust anything. Watch whether Samsung's record quarter is followed by capacity announcements that would loosen memory pricing, which would show up in US cloud and device costs. Watch the consolidated state litigation against TikTok, including whether the placebo-feature allegation is substantiated or narrowed in court filings. And watch whether regulators or exchanges respond to any of this with new expectations about how run-rate and safety claims are disclosed.

The test ahead is simple: whether the gap between the reported number and the underlying reality narrows on its own, or whether it takes a court, a correction, or a price cycle to close it.

More on this beat: Companies on TechManNews.

#Big Tech#OpenAI#Samsung#TikTok#Earnings#Regulation

Newsletter

Get Tech News in Your Inbox

The latest AI, gadgets, software and startup stories from TechManNews, delivered every morning - free.