The records behind this event
3-
AI Model
DeepSeek-R1
The open reasoning model that erased $600B of NVIDIA's market cap in a day.
R1 (January 20, 2025) matched o1-class reasoning and published how: large-scale reinforcement learning on verifiable rewards, with the R1-Zero ablation showing reasoning emerging from RL alone. Released MIT-licensed with distilled small variants, it topped app stores, triggered the January 27 market convulsion (NVIDIA −~17%, ~$589B in a day), and reset…
Open record → -
Company
DeepSeek
The lab that shook the market: frontier results at a fraction of the compute.
Spun from quant fund High-Flyer by Liang Wenfeng, DeepSeek trained V3 on a 2,048-H800 cluster for a disclosed ~$5.6M final-run cost, then released the R1 reasoning model openly in January 2025, triggering a ~$600B one-day fall in NVIDIA's market value as investors re-priced compute efficiency. Its open-weights MoE releases remain the benchmark for…
Open record → -
Concept
Jevons Paradox
Why efficiency can increase total consumption: the argument of 2025.
Jevons (1865) observed that more efficient steam engines increased total coal use: cheaper output finds more uses. Satya Nadella invoked it the week DeepSeek-R1 crashed NVIDIA's stock: cheaper intelligence, he argued, means vastly more intelligence consumed. The 2025 record (token prices down, token volumes and capex up) has so far behaved exactly as the…
Open record →