The Minutes of Dylan Patel on cost https://minutesof.com/dylan-patel/on/cost/ Everything Dylan Patel has said on cost: 4 verbatim quotes between December 2024 and July 2026, each with a timestamp and a link to the recording it came… en Mon, 31 Aug 2026 13:10:04 +0000 Patel reports that tokens now represent 30% of his 90-person company's costs versus 70% for employees. https://minutesof.com/q/289f5368-a657-42fd-9826-8e45417ec2ab/ https://minutesof.com/q/289f5368-a657-42fd-9826-8e45417ec2ab/ “My my own company of 90 people, 30% of my cost now is tokens versus 70% employee costs.” — RAISE Summit Wed, 22 Jul 2026 16:43:49 +0000 cost Patel states each gigawatt AI deployment costs tens of billions of dollars end-to-end. https://minutesof.com/q/5246d720-be43-4fb0-a34a-d42ebce98657/ https://minutesof.com/q/5246d720-be43-4fb0-a34a-d42ebce98657/ “Each of these gigawatt deployments once you take it from all the way from the power to the GPUs and the cloud contract, those are tens of billions of dollars each gigawatt.” — Nebius Tue, 02 Sep 2025 17:00:38 +0000 cost Patel says DeepSeek's model is 400 to 600x cheaper than original GPT-4. https://minutesof.com/q/b503a431-8e3a-43e6-b85f-702276952f7a/ https://minutesof.com/q/b503a431-8e3a-43e6-b85f-702276952f7a/ “And so now DeepSeek's model is something like 400 to 600x cheaper than the original GPT-four.” — MedBricks Webcast Thu, 27 Mar 2025 13:55:08 +0000 cost Patel calculates o1 models increase inference costs by roughly 50x compared to standard models per query. https://minutesof.com/q/ba6dec46-b29b-45c6-8b5b-c55b085a966f/ https://minutesof.com/q/ba6dec46-b29b-45c6-8b5b-c55b085a966f/ “Cost increase for a single token to be generated is four to five x, but then I'm generating 10 x as many tokens.” — BG2 Pod Mon, 23 Dec 2024 20:12:56 +0000 cost