What is Klarna co-founder Sebastian Siemiatkowski's concern about AI investments?

Siemiatkowski, an investor in prominent AI companies, is concerned about the tech industry's multibillion-dollar spending on data centers to power AI models, believing the huge sums may not be worth it.

How much have tech giants invested in AI infrastructure?

According to the article, just four tech groups - Alphabet, Amazon, Meta, and Microsoft - announced a combined $112 billion in capital expenditure for the third quarter of 2025 to fund their AI expansion.

Why does Siemiatkowski think less computing power is needed for advanced AI?

Siemiatkowski argues that AI models themselves are the "most effective compression technology ever invented" and could run more efficiently over time, requiring less computing power than tech companies are planning to spend.

Home / Business and Economy / AI Investor Warns of Risky Data Center Spending Spree

AI Investor Warns of Risky Data Center Spending Spree

17 Nov, 2025

•

Summary

Klarna co-founder Sebastian Siemiatkowski warns of excessive AI infrastructure investments
Concerns over $1tn being spent on data centers to power AI models
Believes less computing power is needed as AI models are "most effective compression technology"

AI Investor Warns of Risky Data Center Spending Spree

As of November 17th, 2025, Klarna co-founder Sebastian Siemiatkowski, who holds shares in leading AI companies, has expressed concerns over the tech industry's massive investments in data centers to power AI models. Siemiatkowski, an AI evangelist, believes the industry is pouring too much money into computing infrastructure, with just four tech giants - Alphabet, Amazon, Meta, and Microsoft - announcing a combined $112 billion in capital expenditure for the third quarter.

Siemiatkowski argues that the industry may be overestimating the computing power needed to run advanced AI, as the models themselves are the "most effective compression technology ever invented" and could become more efficient over time. He points to the popularity of ChatGPT as evidence of widespread AI adoption, but questions whether spending $1 trillion on servers is truly worth it.