Chamath @ 00:55:48Inconclusive
techmarkets
CoreWeave’s long‑term outcome hinges on the actual economic useful life of its Nvidia GPUs: if the GPUs’ useful life is around 10 years as assumed in their debt models, CoreWeave will become a highly successful, "killer" business; if the useful life is closer to 5 years, the company will end up in serious financial trouble and be economically underwater. This will become evident over the coming GPU cycle (by roughly 2030).
As long as that they have that calculated right in their models that they used to borrow all this money to buy all these GPUs from Nvidia. This is going to be a killer business. To the extent that they got that calculation wrong, meaning we thought the useful life was ten years, but it turned out to be five. This business is deeply underwater.View on YouTube
Explanation
The prediction is explicitly about what will become clear over the coming GPU cycle, with a time horizon of “by roughly 2030” for whether CoreWeave’s economics work out given the assumed ~10‑year useful life of its Nvidia GPUs versus a shorter ~5‑year life. As of the current date (November 30, 2025), only about 8–9 months have passed since the podcast on March 8, 2025, and we are still several years away from 2030. There is not yet enough elapsed time or publicly available evidence to determine definitively whether CoreWeave has become a “killer business” or is “deeply underwater” due to GPU useful-life assumptions in its debt models. Therefore, the prediction’s correctness cannot yet be evaluated and remains too early to call.