Last updated Nov 29, 2025
Prediction
Chamath
Chamath @ 00:45:34Inconclusive
aimarkets
Over the coming years, there will be a significant downward reset or slowdown in AI-related capital expenditure (e.g., data center and GPU build-out) by large technology companies, as current levels of AI CapEx prove unsustainably high relative to realized returns.
So I think that you're going to have to have some sort of reset in terms of the CapEx that's happened here.View on YouTube
Explanation

Chamath predicted that, over the coming years, there would be a significant downward reset or slowdown in AI‑related capex by large tech firms as current spending proves unsustainably high. As of late 2025, the observable data point the other way: forecasts and guidance still show rapidly rising AI and data‑center capex rather than a reset. Morgan Stanley in August 2025 projected global cloud capex to jump 56 percent year‑on‑year in 2025 to about $445 billion and then rise another 31 percent in 2026, driven largely by AI infrastructure. (deepnewz.com) Dell’Oro Group in 2025 raised its forecast for global data‑center capex to grow at a 21 percent CAGR to $1.2 trillion by 2029, explicitly citing AI GPUs and custom accelerators as the main growth driver. (datacenterdynamics.com) Citigroup now expects AI capex across hyperscalers such as Microsoft, Amazon and Alphabet to reach about $490 billion by 2026 and total roughly $2.8 trillion by 2029, noting that firms are even borrowing to sustain this build‑out. (theoutpost.ai) Bank of America similarly sees AI capex at $414 billion in 2025 (up 44 percent year‑on‑year) and still increasing in 2026, with Google, Microsoft and Meta maintaining or raising their capex guidance rather than cutting it. (fastbull.com) Other analyses suggest AI capex could exceed half a trillion dollars in 2026 and remark that AI infrastructure spending shows no sign of cooling, as hyperscalers continue to expand data‑center capacity into the late 2020s. (fastbull.com) At the same time, some commentators already argue that 2025 might be a near‑term peak and that capex intensity is unlikely to remain far above historical norms indefinitely, implying a possible future pullback but not yet demonstrating one. (forbes.com) Because Chamath’s claim was framed over the coming years and the current evidence shows continued acceleration rather than a clear downward reset, it is too early to determine whether his prediction about an eventual capex reset will ultimately be right or wrong; it has not come true yet, but the time horizon has not fully played out.