The Minutes of Dylan Patel on gpu supply chain https://minutesof.com/dylan-patel/on/gpu-supply-chain/ Everything Dylan Patel has said on GPU supply chain: 24 verbatim quotes between November 2024 and August 2026, each with a timestamp and a link to the… en Sun, 30 Aug 2026 16:22:17 +0000 Patel says NVIDIA has locked up over 60% of supply chain capacity this year in long-term contracts. https://minutesof.com/q/7f7c89c2-e221-4652-adbe-c84494b67064/ https://minutesof.com/q/7f7c89c2-e221-4652-adbe-c84494b67064/ “The NVIDIA has locked up over 60% of the capacity this year in long term contracts alone, and they're buying more on top of that.” — CNBC Television Mon, 16 Mar 2026 21:58:55 +0000 gpu supply chain Patel reports NVIDIA is negotiating over $250 billion in supply contracts across components this year. https://minutesof.com/q/99be2869-e51d-4edf-b958-4a0eb31001f1/ https://minutesof.com/q/99be2869-e51d-4edf-b958-4a0eb31001f1/ “And as you look at what they're negotiating in the market today, there's over 250,000,000,000 of wafers, of memory, of substrates, of PCBs, of networking equipment that they're going to sign this year,” — CNBC Television Mon, 16 Mar 2026 21:58:55 +0000 gpu supply chain Patel argues NVIDIA is setting up supply chain to manufacture tens of millions of AI chips. https://minutesof.com/q/6de9e7cc-be2c-4b14-aefa-a55d944bc979/ https://minutesof.com/q/6de9e7cc-be2c-4b14-aefa-a55d944bc979/ “I can't buy millions, tens of millions, which is what Jensen's, setting his supply chain up for.” — CNBC Television Mon, 16 Mar 2026 21:58:55 +0000 gpu supply chain Patel details that producing one gigawatt of Rubin chips requires specific wafer volumes across multiple manuf https://minutesof.com/q/ff6866ea-ea84-4f63-9d38-dd6b9d9f502b/ https://minutesof.com/q/ff6866ea-ea84-4f63-9d38-dd6b9d9f502b/ “a gigawatt of, you know, NVIDIA's Rubin chips. Right? So Rubin is announced at GTC, I believe, the week this podcast goes live.” — Dwarkesh Patel Fri, 13 Mar 2026 16:26:57 +0000 gpu supply chain Patel calculates Anthropic added $1.5 billion in compute infrastructure in one month to serve new revenue grow https://minutesof.com/q/b1ba2e20-500e-4a5d-8e13-3d8ecec9f6a2/ https://minutesof.com/q/b1ba2e20-500e-4a5d-8e13-3d8ecec9f6a2/ “If Anthropic added 2,500,000,000 of revenue, and their gross margin is 40%, they added like $1,500,000,000 of compute in one month. Mhmm. Right?” — Latent Space Thu, 26 Feb 2026 21:15:18 +0000 gpu supply chain Patel predicts Google will have zero cash flow profit in 2027, spending all revenue on AI infrastructure. https://minutesof.com/q/1a954518-d905-4ab6-b6df-6800d4e18a40/ https://minutesof.com/q/1a954518-d905-4ab6-b6df-6800d4e18a40/ “There's no reason why Google will have any profit in '27 at all, right, in terms of cash flow. They will just spend every dollar they make on on AI infrastructure.” — Latent Space Thu, 26 Feb 2026 21:15:18 +0000 gpu supply chain Patel states OpenAI has over 500,000 GPUs for R&D but GPT-5 pre-training uses under 100,000 GPUs. https://minutesof.com/q/68335aae-f02e-47ac-b961-50c4701d12af/ https://minutesof.com/q/68335aae-f02e-47ac-b961-50c4701d12af/ “OpenAI has over 500,000 GPUs working on r and d. Right? Let's call it r and d.” — SAIL Media Thu, 15 Jan 2026 16:46:15 +0000 gpu supply chain Patel says the standard inference deployment unit has shifted from single nodes to hundreds of GPUs. https://minutesof.com/q/ee180f12-a142-4bbb-8e53-e032a9458215/ https://minutesof.com/q/ee180f12-a142-4bbb-8e53-e032a9458215/ “the standard unit for an inference deployment being hundreds of GPUs instead of a single node. And then there's all these different things about traffic.” — Open Compute Project Thu, 23 Oct 2025 06:01:19 +0000 gpu supply chain Patel notes NVIDIA Blackwell performance improved dramatically from launch to present, as did AMD hardware. https://minutesof.com/q/8a048f38-2c8d-4690-a1d2-66e094aaa6fb/ https://minutesof.com/q/8a048f38-2c8d-4690-a1d2-66e094aaa6fb/ “If you tried to use NVIDIA's Blackwell six months ago, the numbers were not amazing, right? But now they're actually amazing. So how did that progress over time? Same with AMD, right?” — Open Compute Project Thu, 23 Oct 2025 06:01:19 +0000 gpu supply chain Patel identifies GB200's power efficiency advantage but notes deployment challenges with backplane and liquid https://minutesof.com/q/d7dccc2a-7013-4972-b5c9-372d97cae003/ https://minutesof.com/q/d7dccc2a-7013-4972-b5c9-372d97cae003/ “GB200 has a huge power efficiency advantage, right? Everyone here understands the challenges of running and deploying GB200. There's a lot of challenges with the backplane.” — Open Compute Project Thu, 23 Oct 2025 06:01:19 +0000 gpu supply chain Patel reports GB200 is 10x more power efficient than H200 at certain interactivity rates. https://minutesof.com/q/70c03321-ad13-4824-91b8-9ec2ebd3b7be/ https://minutesof.com/q/70c03321-ad13-4824-91b8-9ec2ebd3b7be/ “But it turns out at certain interactivity rates, I. E. Tokens per second per user, it's 10x more efficient per watt, right, compared to H200.” — Open Compute Project Thu, 23 Oct 2025 06:01:19 +0000 gpu supply chain Patel finds AMD MI355 beats NVIDIA B200 on performance TCO in certain publicly usable configurations. https://minutesof.com/q/b9346a92-e689-4ff1-a8aa-a519c318c01e/ https://minutesof.com/q/b9346a92-e689-4ff1-a8aa-a519c318c01e/ “So we do different scenarios. We do document processing, which is 8,000 context in, 1,000 out. We do chat, which is 1,000 in, 1,000 out.” — Open Compute Project Thu, 23 Oct 2025 06:01:19 +0000 gpu supply chain Patel shows B200 has 15x raw performance advantage over H100 but only 10x performance per TCO. https://minutesof.com/q/101225bc-685d-4ce2-a64b-7e2df33b1f98/ https://minutesof.com/q/101225bc-685d-4ce2-a64b-7e2df33b1f98/ “If we don't divide by TCO, then it looks like the performance of B200 is actually 15x that of H100, versus the performance TCO is only 10x,” — Open Compute Project Thu, 23 Oct 2025 06:01:19 +0000 gpu supply chain Patel lists NVIDIA's structural advantages across networking, HBM, process nodes, speed to market, and supplie https://minutesof.com/q/af9d9adc-9470-4ff6-aba9-8433597b2e9f/ https://minutesof.com/q/af9d9adc-9470-4ff6-aba9-8433597b2e9f/ “NVIDIA's gonna have better networking than you. They're gonna have better, HBM. They're gonna have better process node. They're gonna come to market faster.” — a16z Mon, 18 Aug 2025 18:26:43 +0000 gpu supply chain Patel argues competing with NVIDIA requires leaping forward beyond supply chain advantages across networking, https://minutesof.com/q/1bef91d6-c75f-476c-a33f-1cec565d5243/ https://minutesof.com/q/1bef91d6-c75f-476c-a33f-1cec565d5243/ “They're gonna have better negotiations with whether it's TSMC or SK Hynix and the memory and silicon side or all the rack people or, like, copper cables, everything, they're gonna have better cost efficiency.” — a16z Mon, 18 Aug 2025 18:26:43 +0000 gpu supply chain Patel says OpenAI and Anthropic are receiving 30% of all GPU chips being produced this year. https://minutesof.com/q/c914f140-c01e-4389-9fe1-0e490d20dbdc/ https://minutesof.com/q/c914f140-c01e-4389-9fe1-0e490d20dbdc/ “30% of the chips are going to them, just those two companies. But that's actually like, okay, well, 70% of the stuff, who's making off well, one third of it is ads, whether it be ByteDance or Meta or many of the other people who are doing ads.” — a16z Mon, 18 Aug 2025 18:26:43 +0000 gpu supply chain Patel explains competitors need 5x hardware advantage over NVIDIA but risk failure if AI workloads shift befor https://minutesof.com/q/f8382350-ebd6-4af4-ac16-1dfdcd3fde66/ https://minutesof.com/q/f8382350-ebd6-4af4-ac16-1dfdcd3fde66/ “So now you need to do something, you know, that will give you five x advantage, right, in hardware efficiency for a certain type of workload, and then pray the workload doesn't shift.” — a16z Mon, 18 Aug 2025 18:26:43 +0000 gpu supply chain Patel says GPU hardware and networking represent 80% of data center costs, with power and cooling only 20%. https://minutesof.com/q/3a4b0c19-d93c-478e-bced-c99f4ea7645a/ https://minutesof.com/q/3a4b0c19-d93c-478e-bced-c99f4ea7645a/ “It's the GPU purchases. It's the networking. It's the it's the physical data center conversion power conversion equipment. All of this stuff is, like, 80% of the cost.” — a16z Mon, 18 Aug 2025 18:26:43 +0000 gpu supply chain Patel notes CoreWeave got H100 services online six months before every hyperscaler in many cases. https://minutesof.com/q/c2e290f0-d29d-46af-b1bb-46bc871b30f2/ https://minutesof.com/q/c2e290f0-d29d-46af-b1bb-46bc871b30f2/ “Get services online with h one hundreds before every hyperscaler by a factor of, like, as much as, like, six months in many cases.” — CoreWeave Fri, 23 May 2025 17:27:49 +0000 gpu supply chain Patel says CoreWeave can build $10 billion clusters from start to completion much faster than competitors. https://minutesof.com/q/373f08c1-885e-4cb0-8a0c-139ecb2143a4/ https://minutesof.com/q/373f08c1-885e-4cb0-8a0c-139ecb2143a4/ “it lets you get to these, you know, call it $10,000,000,000 clusters, right, in in time scales, from project start to completion that are much much shorter than anyone else.” — CoreWeave Fri, 23 May 2025 17:27:49 +0000 gpu supply chain Patel says GPT-4 used 24,000 GPUs, GPT-5 uses 100,000, and Microsoft is building toward a million GPUs. https://minutesof.com/q/49107836-7a8a-4175-851d-c6bc483f45a1/ https://minutesof.com/q/49107836-7a8a-4175-851d-c6bc483f45a1/ “GPT four was trained with 24,000 GPUs roughly, and GPT five is on the order of a 100,000. And then they're trying to build this data center over the next few years. That's a million.” — Scaling Intelligence Tue, 12 Nov 2024 04:17:00 +0000 gpu supply chain Patel reveals OpenAI signed a deal with Oracle for 200 megawatts because Microsoft cannot provide enough compu https://minutesof.com/q/77bb9fd8-374c-4dd6-9378-89d774c7068b/ https://minutesof.com/q/77bb9fd8-374c-4dd6-9378-89d774c7068b/ “Again, a gigawatt is like 500,000 plus GPUs. That's billions, tens of billions of dollars. So OpenAI signed a deal for just 200 megawatts.” — Scaling Intelligence Tue, 12 Nov 2024 04:17:00 +0000 gpu supply chain