Klarna CEO Sebastian Siemiatkowski says he is “nervous about the size of these investments in these data centers”; he owns stakes in OpenAI, Perplexity, and xAI
OpenAI backer Sebastian Siemiatkowski says tech industry is committing too much money to huge computing infrastructure X: @quinnypig X: Corey Quinn / @quinnypig : Can someone give a coherent explanation how building this astounding amount of compute isn't simply building a bridge to nowhere? What's it supposed to unlock past “more AI?”
Context & Ripple Effects
The warning comes as related coverage has documented a global data-center buildout facing capacity limits and ROI questions. It is notable because Siemiatkowski is not a detached critic: he holds stakes in OpenAI, Perplexity, and xAI.
It also complicates the growth narrative around AI providers. OpenAI has said it expects to finance infrastructure from rapidly growing revenues rather than seek government guarantees for data centers.
First-order effects
- Siemiatkowski’s comments add a skeptical investor perspective to the infrastructure push of companies in which he has a financial interest, sharpening attention on whether compute commitments can earn adequate returns.
- OpenAI, Perplexity, and xAI face more visible scrutiny of the commercial rationale behind ever-larger compute deployments, even though the comments announce no change in their plans.
Second-order effects
- Data-center developers, chip suppliers, and infrastructure financiers may face tougher questions from investors about utilization, customer demand, and the pace at which committed capacity becomes productive revenue.
- AI companies seeking large compute contracts will have stronger incentives to demonstrate revenue-backed demand, reinforcing the debate prompted by concerns over returns on the buildout.
Third-order effects
- If investor skepticism broadens, AI infrastructure competition could shift from securing the most capacity to proving that capacity can be financed and monetized sustainably.
- The sector may increasingly separate firms able to fund compute from operating cash flow from those dependent on continued external capital; that outcome remains contingent on actual AI demand and utilization.
The trend: AI’s infrastructure race is entering a more financially disciplined phase, in which the central question is whether compute investment converts into durable revenue before capital commitments outpace demand.