corpus.blog
41554 blogs · [ { "id": "01a07df5-1955-7061-924c-00a80756ab04", "title": "xAI’s Colossus 2 – First Gigawatt Datacenter In The World, Unique RL Methodology, Capital Raise", "url": "https://semianalysis.com/2025/09/16/xais-colossus-2-first-gigawatt-datacenter/", "published_at": "2025-09-16T17:38:01+00:00" }, { "id": "01a07df5-1955-7061-924c-00a807a5e387", "title": "Another Giant Leap: The Rubin CPX Specialized Accelerator & Rack", "url": "https://semianalysis.com/2025/09/10/another-giant-leap-the-rubin-cpx-specialized-accelerator-rack/", "published_at": "2025-09-10T19:57:18+00:00" }, { "id": "01a07df5-1955-7061-924c-00a808903175", "title": "Huawei Ascend Production Ramp: Die Banks, TSMC Continued Production, HBM is The Bottleneck", "url": "https://semianalysis.com/2025/09/08/huawei-ascend-production-ramp/", "published_at": "2025-09-08T09:54:57+00:00" }, { "id": "01a07df5-1955-7061-924c-00a80973bdc4", "title": "Amazon’s AI Resurgence: AWS & Anthropic’s Multi-Gigawatt Trainium Expansion", "url": "https://semianalysis.com/2025/09/03/amazons-ai-resurgence-aws-anthropics-multi-gigawatt-trainium-expansion/", "published_at": "2025-09-03T20:55:46+00:00" }, { "id": "01a07df5-1955-7061-924c-00a8099a492a", "title": "H100 vs GB200 NVL72 Training Benchmarks – Power, TCO, and Reliability Analysis, Software Improvement Over Time", "url": "https://semianalysis.com/2025/08/20/h100-vs-gb200-nvl72-training-benchmarks/", "published_at": "2025-08-20T04:56:35+00:00" }, { "id": "01a07df5-1955-7061-924c-00a80a36a031", "title": "GPT-5 Set the Stage for Ad Monetization and the SuperApp", "url": "https://semianalysis.com/2025/08/13/gpt-5-ad-monetization-and-the-superapp/", "published_at": "2025-08-13T00:27:14+00:00" }, { "id": "01a07df5-1955-7061-924c-00a80acc0f75", "title": "Scaling the Memory Wall: The Rise and Roadmap of HBM", "url": "https://semianalysis.com/2025/08/12/scaling-the-memory-wall-the-rise-and-roadmap-of-hbm/", "published_at": "2025-08-12T01:16:06+00:00" }, { "id": "01a07df5-1955-7061-924c-00a80b58164f", "title": "Robotics Levels of Autonomy", "url": "https://semianalysis.com/2025/07/30/robotics-levels-of-autonomy/", "published_at": "2025-07-30T17:02:25+00:00" }, { "id": "01a07df5-1955-7061-924c-00a80bbecf85", "title": "Intel 18A Details & Cost, Future of DRAM 4F2 vs 3D, Backside Power Adoption (or Not), China’s FlipFET, Digital Twins from Atoms to Fabs, and More", "url": "https://semianalysis.com/2025/07/21/vlsi2025/", "published_at": "2025-07-21T14:23:37+00:00" }, { "id": "01a07df5-1955-7061-924c-00a80c0744f5", "title": "Meta Superintelligence – Leadership Compute, Talent, and Data", "url": "https://semianalysis.com/2025/07/11/meta-superintelligence-leadership-compute-talent-and-data/", "published_at": "2025-07-11T20:12:19+00:00" }, { "id": "01a0cabd-c43f-73da-9b91-5e01bef513d3", "title": "Introduction to AI Networking Model", "url": "https://semianalysis.com/2025/07/11/introduction-to-ai-networking-model/", "published_at": "2025-07-11T16:18:14+00:00" }, { "id": "01a0cabd-c43f-73da-9b91-5e01bf57cc59", "title": "DeepSeek Debrief: >128 Days Later ", "url": "https://semianalysis.com/2025/07/03/deepseek-debrief-128-days-later/", "published_at": "2025-07-03T18:31:49+00:00" }, { "id": "01a0cabd-c43f-73da-9b91-5e01bfb561d3", "title": "How Oracle Is Winning the AI Compute Market", "url": "https://semianalysis.com/2025/06/30/how-oracle-is-winning-the-ai-compute-market/", "published_at": "2025-06-30T13:24:25+00:00" }, { "id": "01a0cabd-c43f-73da-9b91-5e01c0041756", "title": "AI Training Load Fluctuations at Gigawatt-scale – Risk of Power Grid Blackout?", "url": "https://semianalysis.com/2025/06/25/ai-training-load-fluctuations-at-gigawatt-scale-risk-of-power-grid-blackout/", "published_at": "2025-06-25T18:15:55+00:00" }, { "id": "01a0cabd-c43f-73da-9b91-5e01c0e6defd", "title": "NVIDIA Tensor Core Evolution: From Volta To Blackwell", "url": "https://semianalysis.com/2025/06/23/nvidia-tensor-core-evolution-from-volta-to-blackwell/", "published_at": "2025-06-23T06:03:17+00:00" }, { "id": "01a0cabd-c43f-73da-9b91-5e01c0e9c33a", "title": "AMD Advancing AI: MI350X and MI400 UALoE72, MI500 UAL256", "url": "https://semianalysis.com/2025/06/13/amd-advancing-ai-mi350x-and-mi400-ualoe72-mi500-ual256/", "published_at": "2025-06-13T15:38:21+00:00" }, { "id": "01a0cabd-c43f-73da-9b91-5e01c143fc57", "title": "The New AI Networks | Ultra Ethernet UEC | UALink vs Broadcom Scale Up Ethernet SUE", "url": "https://semianalysis.com/2025/06/11/the-new-ai-networks-ultra-ethernet-uec-ualink-vs-broadcom-scale-up-ethernet-sue/", "published_at": "2025-06-11T14:48:38+00:00" }, { "id": "01a0cabd-c43f-73da-9b91-5e01c1d81040", "title": "Scaling Reinforcement Learning: Environments, Reward Hacking, Agents, Scaling Data", "url": "https://semianalysis.com/2025/06/08/scaling-reinforcement-learning-environments-reward-hacking-agents-scaling-data/", "published_at": "2025-06-08T23:08:18+00:00" }, { "id": "01a0cabd-c43f-73da-9b91-5e01c2d16599", "title": "AMD vs NVIDIA Inference Benchmark: Who Wins? – Performance & Cost Per Million Tokens", "url": "https://semianalysis.com/2025/05/23/amd-vs-nvidia-inference-benchmark-who-wins-performance-cost-per-million-tokens/", "published_at": "2025-05-23T05:41:13+00:00" }, { "id": "01a0cabd-c43f-73da-9b91-5e01c2d57289", "title": "AI Arrives In The Middle East: US Strikes A Deal with UAE and KSA", "url": "https://semianalysis.com/2025/05/16/ai-arrives-in-the-middle-east-us-strikes-a-deal-with-uae-and-ksa/", "published_at": "2025-05-16T01:57:23+00:00" }, { "id": "01a0cabe-6b2e-73f1-8ee8-edf76a987dfc", "title": "Microsoft’s Datacenter Freeze – 1.5GW Self-Build Slowdown & Lease Cancellation Misconceptions", "url": "https://semianalysis.com/2025/04/28/microsofts-datacenter-freeze/", "published_at": "2025-04-28T10:31:36+00:00" }, { "id": "01a0cabe-6b2e-73f1-8ee8-edf76ad8823c", "title": "AMD 2.0 – New Sense of Urgency | MI450X Chance to Beat Nvidia | Nvidia’s New Moat", "url": "https://semianalysis.com/2025/04/23/amd-2-0-new-sense-of-urgency-mi450x-chance-to-beat-nvidia-nvidias-new-moat/", "published_at": "2025-04-23T12:22:49+00:00" }, { "id": "01a0cabe-6b2e-73f1-8ee8-edf76b616767", "title": "Huawei AI CloudMatrix 384 – China’s Answer to Nvidia GB200 NVL72", "url": "https://semianalysis.com/2025/04/16/huawei-ai-cloudmatrix-384-chinas-answer-to-nvidia-gb200-nvl72/", "published_at": "2025-04-16T04:18:18+00:00" }, { "id": "01a0cabe-6b2e-73f1-8ee8-edf76c00ed2d", "title": "High-NA is Here (for R&D), EUV Cost, Pattern Shaping Gaining Share, 6×12″ Mask, Metal Oxide & Dry Resist, Hyper-NA", "url": "https://semianalysis.com/2025/04/14/spie2025/", "published_at": "2025-04-14T08:26:39+00:00" }, { "id": "01a0cabe-6b2e-73f1-8ee8-edf76cfe0aa6", "title": "Tariff Armageddon? | GPU Loopholes, Mexico Supply Chain Shift, Wafer Fab Equipment Vulnerabilities, Optical Module Pricing Surge, Datacenter Equipment", "url": "https://semianalysis.com/2025/04/10/tariff-armageddon-gpu-loopholes/", "published_at": "2025-04-10T08:47:50+00:00" }, { "id": "01a0cabe-6b2e-73f1-8ee8-edf76d43b60c", "title": "The GPU Cloud ClusterMAX™ Rating System | How to Rent GPUs", "url": "https://semianalysis.com/2025/03/26/the-gpu-cloud-clustermax-rating-system-how-to-rent-gpus/", "published_at": "2025-03-26T09:11:39+00:00" }, { "id": "01a0cabe-6b2e-73f1-8ee8-edf76db3f736", "title": "NVIDIA GTC 2025 – Built For Reasoning, Vera Rubin, Kyber, CPO, Dynamo Inference, Jensen Math, Feynman", "url": "https://semianalysis.com/2025/03/19/nvidia-gtc-2025-built-for-reasoning-vera-rubin-kyber-cpo-dynamo-inference-jensen-math-feynman/", "published_at": "2025-03-19T02:08:11+00:00" }, { "id": "01a0cabe-6b2e-73f1-8ee8-edf76e561bb4", "title": "America Is Missing The New Labor Economy – Robotics Part 1", "url": "https://semianalysis.com/2025/03/11/america-is-missing-the-new-labor-economy-robotics-part-1/", "published_at": "2025-03-11T07:03:18+00:00" }, { "id": "01a0cabe-6b2e-73f1-8ee8-edf76ecb4651", "title": "Datacenter Anatomy Part 2 – Cooling Systems", "url": "https://semianalysis.com/2025/02/13/datacenter-anatomy-part-2-cooling-systems/", "published_at": "2025-02-13T16:09:36+00:00" }, { "id": "01a0cabe-6b2e-73f1-8ee8-edf76fca22c0", "title": "TSMC N2 + Next-Gen SoIC, Intel EMIB-T, Meta 3D Stacked Memory, CFET, 2D Materials, and More", "url": "https://semianalysis.com/2025/02/05/iedm2024/", "published_at": "2025-02-05T07:07:10+00:00" }, { "id": "01a0cabe-f4d5-71c5-bcf7-abe7f1f91ae0", "title": "DeepSeek Debates: Chinese Leadership On Cost, True Training Cost, Closed Model Margin Impacts", "url": "https://semianalysis.com/2025/01/31/deepseek-debates/", "published_at": "2025-01-31T04:56:46+00:00" }, { "id": "01a0cabe-f4d5-71c5-bcf7-abe7f2747e3a", "title": "OpenAI Stargate Joint Venture Demystified | Microsoft Sore Loser, Does Softbank Have The Capital?, Texas GigaCampus, Winners & Losers", "url": "https://semianalysis.com/2025/01/23/openai-stargate-joint-venture-demystified/", "published_at": "2025-01-23T06:57:24+00:00" }, { "id": "01a0cabe-f4d5-71c5-bcf7-abe7f366bcf2", "title": "2025 AI Diffusion Export Controls – Microsoft Regulatory Capture, Oracle Tears, Impacts Quantified, Model Restrictions", "url": "https://semianalysis.com/2025/01/15/2025-ai-diffusion-export-controls-microsoft-regulatory-capture-oracle-tears/", "published_at": "2025-01-15T08:43:45+00:00" }, { "id": "01a0cabe-f4d5-71c5-bcf7-abe7f3c7cd61", "title": "Nvidia’s Christmas Present: GB300 & B300 – Reasoning Inference, Amazon, Memory, Supply Chain", "url": "https://semianalysis.com/2024/12/25/nvidias-christmas-present-gb300-b300-reasoning-inference-amazon-memory-supply-chain/", "published_at": "2024-12-25T11:06:19+00:00" }, { "id": "01a0cabe-f4d5-71c5-bcf7-abe7f476ee3e", "title": "MI300X vs H100 vs H200 Benchmark Part 1: Training – CUDA Moat Still Alive", "url": "https://semianalysis.com/2024/12/22/mi300x-vs-h100-vs-h200-benchmark-part-1-training/", "published_at": "2024-12-22T22:03:18+00:00" }, { "id": "01a0cabe-f4d5-71c5-bcf7-abe7f4aa6e1c", "title": "Scaling Laws – O1 Pro Architecture, Reasoning Training Infrastructure, Orion and Claude 3.5 Opus “Failures”", "url": "https://semianalysis.com/2024/12/11/scaling-laws-o1-pro-architecture-reasoning-training-infrastructure-orion-and-claude-3-5-opus-failures/", "published_at": "2024-12-11T03:38:44+00:00" }, { "id": "01a0cabe-f4d5-71c5-bcf7-abe7f59cab73", "title": "Intel on the Brink of Death | Culture Rot, Product Focus Flawed, Foundry Must Survive", "url": "https://semianalysis.com/2024/12/09/intel-on-the-brink-of-death/", "published_at": "2024-12-09T06:31:47+00:00" }, { "id": "01a0cabe-f4d5-71c5-bcf7-abe7f68daf6f", "title": "Amazon’s AI Self Sufficiency | Trainium2 Architecture & Networking", "url": "https://semianalysis.com/2024/12/03/amazons-ai-self-sufficiency-trainium2-architecture-networking/", "published_at": "2024-12-03T17:29:30+00:00" }, { "id": "01a0cabe-f4d5-71c5-bcf7-abe7f77f6c75", "title": "Fab Whack-A-Mole: Chinese Companies are Evading U.S. Sanctions", "url": "https://semianalysis.com/2024/10/28/fab-whack-a-mole-chinese-companies/", "published_at": "2024-10-28T08:38:43+00:00" }, { "id": "01a0cabe-f4d5-71c5-bcf7-abe7f85ab9f5", "title": "Datacenter Anatomy Part 1: Electrical Systems", "url": "https://semianalysis.com/2024/10/14/datacenter-anatomy-part-1-electrical/", "published_at": "2024-10-14T14:10:06+00:00" }, { "id": "01a0cabf-6537-71c3-9407-3bf4c4df3f6b", "title": "AI Neocloud Playbook and Anatomy", "url": "https://semianalysis.com/2024/10/03/ai-neocloud-playbook-and-anatomy/", "published_at": "2024-10-03T13:54:17+00:00" }, { "id": "01a0cabf-6537-71c3-9407-3bf4c5ce7fc8", "title": "Clash of the Foundries: Gate All Around + Backside Power at 2nm", "url": "https://semianalysis.com/2024/10/01/clash-of-the-foundries/", "published_at": "2024-10-01T22:20:57+00:00" }, { "id": "01a0cabf-6537-71c3-9407-3bf4c5f45f6c", "title": "Multi-Datacenter Training: OpenAI’s Ambitious Plan To Beat Google’s Infrastructure", "url": "https://semianalysis.com/2024/09/04/multi-datacenter-training-openais/", "published_at": "2024-09-04T11:01:57+00:00" }, { "id": "01a0cabf-6537-71c3-9407-3bf4c62eedaa", "title": "The Memory Wall: Past, Present, and Future of DRAM", "url": "https://semianalysis.com/2024/09/03/the-memory-wall/", "published_at": "2024-09-03T05:13:23+00:00" }, { "id": "01a0cabf-6537-71c3-9407-3bf4c6c4703a", "title": "Nvidia’s Blackwell Reworked – Shipment Delays & GB200A Reworked Platforms", "url": "https://semianalysis.com/2024/08/04/nvidias-blackwell-reworked-shipment/", "published_at": "2024-08-04T20:48:12+00:00" }, { "id": "01a0cabf-6537-71c3-9407-3bf4c7617297", "title": "GB200 Hardware Architecture – Component Supply Chain & BOM", "url": "https://semianalysis.com/2024/07/17/gb200-hardware-architecture-and-component/", "published_at": "2024-07-17T05:07:55+00:00" }, { "id": "01a0cabf-6537-71c3-9407-3bf4c839685c", "title": "GB200 & GB200A Component & Supply Chain Model", "url": "https://semianalysis.com/2024/07/16/semianalysis-gb200-component-and/", "published_at": "2024-07-16T04:14:15+00:00" }, { "id": "01a0cabf-6537-71c3-9407-3bf4c8886737", "title": "100,000 H100 Clusters: Power, Network Topology, Ethernet vs InfiniBand, Reliability, Failures, Checkpointing", "url": "https://semianalysis.com/2024/06/17/100000-h100-clusters-power-network/", "published_at": "2024-06-17T20:06:42+00:00" }, { "id": "01a0cabf-6537-71c3-9407-3bf4c9408c78", "title": "OpenAI Chip Team Is Now Serious", "url": "https://semianalysis.com/2024/06/03/openai-chip-team-is-now-serious/", "published_at": "2024-06-03T15:00:24+00:00" }, { "id": "01a0cabf-6537-71c3-9407-3bf4ca21899c", "title": "How Dell Is Beating Supermicro", "url": "https://semianalysis.com/2024/05/30/how-dell-is-beating-supermicro/", "published_at": "2024-05-30T07:25:13+00:00" }, { "id": "01a0cabf-c4e7-712f-b8ac-f8b6b7776d7f", "title": "Apple’s AI Strategy: Apple Datacenters, On-device, Cloud, And More", "url": "https://semianalysis.com/2024/05/27/apples-ai-strategy-apple-datacenters/", "published_at": "2024-05-27T03:31:15+00:00" }, { "id": "01a0cabf-c4e7-712f-b8ac-f8b6b7eb9b9a", "title": "OpenAI Is Doomed? – Et tu, Microsoft?", "url": "https://semianalysis.com/2024/05/07/openai-is-doomed-et-tu-microsoft/", "published_at": "2024-05-07T13:52:28+00:00" }, { "id": "01a0cabf-c4e7-712f-b8ac-f8b6b8536c21", "title": "Intel’s 14A Magic Bullet: Directed Self-Assembly (DSA)", "url": "https://semianalysis.com/2024/04/18/intels-14a-magic-bullet-directed/", "published_at": "2024-04-18T14:02:18+00:00" }, { "id": "01a0cabf-c4e7-712f-b8ac-f8b6b8d34646", "title": "Nvidia Blackwell Perf TCO Analysis – B100 vs B200 vs GB200 NVL72", "url": "https://semianalysis.com/2024/04/10/nvidia-blackwell-perf-tco-analysis/", "published_at": "2024-04-10T06:52:20+00:00" }, { "id": "01a0cabf-c4e7-712f-b8ac-f8b6b8da0b97", "title": "Is Intel Back? Foundry & Product Resurgence Measured", "url": "https://semianalysis.com/2024/04/02/is-intel-back-foundry-and-product/", "published_at": "2024-04-02T22:34:55+00:00" }, { "id": "01a0cabf-c4e7-712f-b8ac-f8b6b90ce713", "title": "Nvidia’s Optical Boogeyman – NVL72, Infiniband Scale Out, 800G & 1.6T Ramp", "url": "https://semianalysis.com/2024/03/25/nvidias-optical-boogeyman-nvl72-infiniband/", "published_at": "2024-03-25T11:11:55+00:00" }, { "id": "01a0cabf-c4e7-712f-b8ac-f8b6b91bf2be", "title": "Nvidia B100, B200, GB200 – COGS, Pricing, Margins, Ramp – Oberon, Umbriel, Miranda", "url": "https://semianalysis.com/2024/03/18/nvidia-b100-b200-gb200-cogs-pricing/", "published_at": "2024-03-18T12:12:39+00:00" }, { "id": "01a0cabf-c4e7-712f-b8ac-f8b6b9953459", "title": "Astera Labs IPO – The Next Connectivity Superhero or Steamrolled By Competition?", "url": "https://semianalysis.com/2024/03/17/astera-labs-ipo-the-next-connectivity/", "published_at": "2024-03-17T15:19:19+00:00" }, { "id": "01a0cabf-c4e7-712f-b8ac-f8b6b9c64d59", "title": "CXL Is Dead In The AI Era", "url": "https://semianalysis.com/2024/03/16/cxl-is-dead-in-the-ai-era/", "published_at": "2024-03-16T04:26:44+00:00" }, { "id": "01a0cabf-c4e7-712f-b8ac-f8b6b9ca6a52", "title": "AI Datacenter Energy Dilemma – Race for AI Datacenter Space", "url": "https://semianalysis.com/2024/03/13/ai-datacenter-energy-dilemma-race/", "published_at": "2024-03-13T17:35:36+00:00" }, { "id": "01a0cac0-0dd8-73e8-9d9c-1f66cda0d210", "title": "Groq Inference Tokenomics: Speed, But At What Cost?", "url": "https://semianalysis.com/2024/02/21/groq-inference-tokenomics-speed-but/", "published_at": "2024-02-21T10:44:53+00:00" }, { "id": "01a0cac0-0dd8-73e8-9d9c-1f66ce90a335", "title": "Hybrid Bonding Process Flow – Advanced Packaging Part 5", "url": "https://semianalysis.com/2024/02/09/hybrid-bonding-process-flow-advanced/", "published_at": "2024-02-09T06:24:36+00:00" }, { "id": "01a0cac0-0dd8-73e8-9d9c-1f66cf6958e4", "title": "Compliance Policies", "url": "https://semianalysis.com/2024/02/03/semianalysis-compliance-policies/", "published_at": "2024-02-03T20:39:44+00:00" }, { "id": "01a0cac0-0dd8-73e8-9d9c-1f66cf738d2e", "title": "Neural Network Quantization & Number Formats From First Principles", "url": "https://semianalysis.com/2024/01/11/neural-network-quantization-and-number/", "published_at": "2024-01-11T23:34:26+00:00" }, { "id": "01a0cac0-0dd8-73e8-9d9c-1f66d0533217", "title": "Intel GenAI For Yield, TSMC CFET & 3D Stacking, AMD 3D Device Modeling, Applied Materials Material Innovation, SK Hynix HBM4, Micron 3D DRAM & FeRAM, Hybrid Bonding vs TCB – IEDM 2023", "url": "https://semianalysis.com/2024/01/03/intel-genai-for-yield-tsmc-cfet-and/", "published_at": "2024-01-03T20:55:37+00:00" }, { "id": "01a0cac0-0dd8-73e8-9d9c-1f66d0a87e65", "title": "Inference Race To The Bottom – Make It Up On Volume?", "url": "https://semianalysis.com/2023/12/18/inference-race-to-the-bottom-make/", "published_at": "2023-12-18T21:05:28+00:00" }, { "id": "01a0cac0-0dd8-73e8-9d9c-1f66d0bd06fb", "title": "ASML Dilemma: High-NA EUV is Worse vs Low-NA EUV Multi-Patterning", "url": "https://semianalysis.com/2023/12/11/asml-dilemma-high-na-euv-is-worse/", "published_at": "2023-12-11T15:37:47+00:00" }, { "id": "01a0cac0-0dd8-73e8-9d9c-1f66d1307a1a", "title": "AMD MI300 Performance – Faster Than H100, But How Much?", "url": "https://semianalysis.com/2023/12/06/amd-mi300-performance-faster-than/", "published_at": "2023-12-06T20:01:29+00:00" }, { "id": "01a0cac0-0dd8-73e8-9d9c-1f66d1652c6e", "title": "GPU Cloud Economics Explained – The Neocloud Hidden Truth", "url": "https://semianalysis.com/2023/12/04/gpu-cloud-economics-explained-the/", "published_at": "2023-12-04T12:22:09+00:00" }, { "id": "01a0cac0-0dd8-73e8-9d9c-1f66d1f70da0", "title": "Microsoft Swallows OpenAI’s Core Team – GPU Capacity, Incentive Structure, Intellectual Property, OpenAI Rump State", "url": "https://semianalysis.com/2023/11/20/microsoft-swallows-openais-core-team/", "published_at": "2023-11-20T10:57:11+00:00" }, { "id": "01a0cac0-615b-7115-9a73-f887a0eadb84", "title": "Microsoft Infrastructure – AI & CPU Custom Silicon Maia 100, Athena, Cobalt 100", "url": "https://semianalysis.com/2023/11/15/microsoft-infrastructure-ai-and-cpu/", "published_at": "2023-11-15T16:09:34+00:00" }, { "id": "01a0cac0-615b-7115-9a73-f887a19a46e4", "title": "Nvidia’s New China AI Chips Circumvent US Restrictions | H20 Faster Than H100 | Huawei Ascend 910B", "url": "https://semianalysis.com/2023/11/09/nvidias-new-china-ai-chips-circumvent/", "published_at": "2023-11-09T10:18:29+00:00" }, { "id": "01a0cac0-615b-7115-9a73-f887a2592e76", "title": "AMD MI300 Ramp, GPT-4 Performance, ASP & Volumes", "url": "https://semianalysis.com/2023/11/01/amd-mi300-ramp-gpt-4-performance/", "published_at": "2023-11-01T07:19:00+00:00" }, { "id": "01a0cac0-615b-7115-9a73-f887a311e541", "title": "Wafer Wars: Deciphering Latest Restrictions On AI And Semiconductor Manufacturing", "url": "https://semianalysis.com/2023/10/24/wafer-wars-deciphering-latest-restrictions/", "published_at": "2023-10-24T10:18:12+00:00" }, { "id": "01a0cac0-615b-7115-9a73-f887a31a992c", "title": "Going Vertical: Gate All Around, 3D DRAM, 3D NAND – Kokusai Electric IPO", "url": "https://semianalysis.com/2023/10/15/going-vertical-gate-all-around-3d/", "published_at": "2023-10-15T20:32:35+00:00" }, { "id": "01a0cac0-615b-7115-9a73-f887a4045196", "title": "Nvidia’s Plans To Crush Competition – B100, “X100”, H200, 224G SerDes, OCS, CPO, PCIe 7.0, HBM3E", "url": "https://semianalysis.com/2023/10/10/nvidias-plans-to-crush-competition/", "published_at": "2023-10-10T04:33:44+00:00" }, { "id": "01a0cac0-615b-7115-9a73-f887a43f1d1b", "title": "AI Doomer vs Techno-Optimist: Social Fabric, AI Girlfriends, Mass Media", "url": "https://semianalysis.com/2023/10/05/ai-doomer-vs-techno-optimist-social/", "published_at": "2023-10-05T00:54:42+00:00" }, { "id": "01a0cac0-615b-7115-9a73-f887a48a23fc", "title": "Amazon Anthropic: Poison Pill or Empire Strikes Back", "url": "https://semianalysis.com/2023/10/02/amazon-anthropic-poison-pill-or-empire/", "published_at": "2023-10-02T00:44:55+00:00" }, { "id": "01a0cac0-615b-7115-9a73-f887a4dc3b40", "title": "Nvidia Buys Illumina? The AI Foundry for Healthcare – The Hardware of Life", "url": "https://semianalysis.com/2023/09/27/nvidia-buys-illumina-the-ai-foundry/", "published_at": "2023-09-27T08:02:18+00:00" }, { "id": "01a0cac0-615b-7115-9a73-f887a4f24d07", "title": "Arm and a Leg: Arm’s Quest To Extract Their True Value", "url": "https://semianalysis.com/2023/09/14/arm-and-a-leg-arms-quest-to-extract/", "published_at": "2023-09-14T06:08:33+00:00" }, { "id": "01a0cac0-b000-7364-ab84-cb706275ee9c", "title": "China AI & Semiconductors Rise: US Sanctions Have Failed", "url": "https://semianalysis.com/2023/09/12/china-ai-and-semiconductors-rise/", "published_at": "2023-09-12T07:14:09+00:00" }, { "id": "01a0cac0-b000-7364-ab84-cb706336840b", "title": "TPUv5e: The New Benchmark in Cost-Efficient Inference and Training for", "url": "https://semianalysis.com/2023/09/01/tpuv5e-the-new-benchmark-in-cost/", "published_at": "2023-09-01T07:38:03+00:00" }, { "id": "01a0cac0-b000-7364-ab84-cb7063697f1d", "title": "Broadcom’s Google TPU Revenue Explosion, Networking Boom, VMWare Integration", "url": "https://semianalysis.com/2023/08/30/broadcoms-google-tpu-revenue-explosion/", "published_at": "2023-08-30T21:16:07+00:00" }, { "id": "01a0cac0-b000-7364-ab84-cb7063ce914b", "title": "Google Gemini Eats The World – Gemini Smashes GPT-4 By 5X, The GPU-Poors", "url": "https://semianalysis.com/2023/08/28/google-gemini-eats-the-world-gemini/", "published_at": "2023-08-28T00:02:15+00:00" }, { "id": "01a0cac0-b000-7364-ab84-cb70642e0600", "title": "Nvidia’s Optical Ascent: >$1B Revenue; The Missing 800G Ramp; AI Head-Fakes", "url": "https://semianalysis.com/2023/08/23/nvidias-optical-ascent-1b-revenue/", "published_at": "2023-08-23T18:41:45+00:00" }, { "id": "01a0cac0-b000-7364-ab84-cb70648d441e", "title": "Nvidia’s Ramp – Volume, ASP, Cloud Pricing, Margins, EPS, Cashflow, China, Competition", "url": "https://semianalysis.com/2023/08/20/nvidias-ramp-volume-asp-cloud-pricing/", "published_at": "2023-08-20T18:12:59+00:00" }, { "id": "01a0cac0-b000-7364-ab84-cb70657d5ba4", "title": "Energizing AI: Power Delivery Competition Heats Up Vicor, MPS, Delta, ADI, Renesas, Infineon", "url": "https://semianalysis.com/2023/08/01/energizing-ai-power-delivery-competition/", "published_at": "2023-08-01T06:51:00+00:00" }, { "id": "01a0cac0-b000-7364-ab84-cb706654f5b0", "title": "AI Expansion – Supply Chain Analysis For CoWoS And HBM", "url": "https://semianalysis.com/2023/07/26/ai-expansion-supply-chain-analysis/", "published_at": "2023-07-26T05:09:55+00:00" }, { "id": "01a0cac0-b000-7364-ab84-cb70671d5fd8", "title": "NAND Flash Monopoly Broken? Tokyo Electron Moly Dep + Cryo Etch Takes On Lam Research For The Future Of NAND", "url": "https://semianalysis.com/2023/07/16/nand-flash-monopoly-broken-tokyo/", "published_at": "2023-07-16T16:01:53+00:00" }, { "id": "01a0cac0-b000-7364-ab84-cb7067899a4d", "title": "GPT-4 Architecture, Infrastructure, Training Dataset, Costs, Vision, MoE", "url": "https://semianalysis.com/2023/07/10/gpt-4-architecture-infrastructure/", "published_at": "2023-07-10T19:14:31+00:00" }, { "id": "01a0cac1-09f7-71e7-8156-fb99fd377de6", "title": "AI Capacity Constraints – CoWoS and HBM Supply Chain", "url": "https://semianalysis.com/2023/07/05/ai-capacity-constraints-cowos-and/", "published_at": "2023-07-05T20:29:23+00:00" }, { "id": "01a0cac1-09f7-71e7-8156-fb99fdebd634", "title": "Tesla AI Capacity Expansion – H100, Dojo D1, D2, HW 4.0, X.AI, Cloud Service Provider", "url": "https://semianalysis.com/2023/06/27/tesla-ai-capacity-expansion-h100/", "published_at": "2023-06-27T17:36:12+00:00" }, { "id": "01a0cac1-09f7-71e7-8156-fb99fe8051cc", "title": "Rebuilding Intel – Foundry vs IDM Decades of Inefficiencies Unraveled", "url": "https://semianalysis.com/2023/06/22/rebuilding-intel-foundry-vs-idm-decades/", "published_at": "2023-06-22T08:07:03+00:00" }, { "id": "01a0cac1-09f7-71e7-8156-fb99ff39b238", "title": "AMD MI300 – Taming The Hype – AI Performance, Volume Ramp, Customers, Cost, IO, Networking, Software", "url": "https://semianalysis.com/2023/06/12/amd-mi300-taming-the-hype-ai-performance/", "published_at": "2023-06-12T16:58:00+00:00" }, { "id": "01a0cac1-09f7-71e7-8156-fb99ff792a2a", "title": "Zen 4c: AMD’s Response to Hyperscale ARM & Intel Atom", "url": "https://semianalysis.com/2023/06/05/zen-4c-amds-response-to-hyperscale/", "published_at": "2023-06-05T10:23:00+00:00" }, { "id": "01a0cac1-09f7-71e7-8156-fb99fffe6430", "title": "AI Server Cost Analysis – Memory Is The Biggest Loser", "url": "https://semianalysis.com/2023/05/29/ai-server-cost-analysis-memory-is/", "published_at": "2023-05-29T12:42:40+00:00" }, { "id": "01a0cac1-09f7-71e7-8156-fb9a0040b526", "title": "Nvidia’s InfiniBand Problem – Spectrum-X AI Fabric, Tomahawk-5, Jericho-3AI, Quantum-2", "url": "https://semianalysis.com/2023/05/28/nvidias-infiniband-problem-qmx-ai/", "published_at": "2023-05-28T16:33:42+00:00" }, { "id": "01a0cac1-09f7-71e7-8156-fb9a008eff0c", "title": "Meta Custom Silicon: What’s Old Is New", "url": "https://semianalysis.com/2023/05/19/meta-custom-silicon-whats-old-is/", "published_at": "2023-05-19T20:57:51+00:00" }, { "id": "01a0cac1-09f7-71e7-8156-fb9a01262f68", "title": "Sound The Siryn: AmpereOne 192-Core CPU", "url": "https://semianalysis.com/2023/05/18/sound-the-siryn-ampereone-192-core/", "published_at": "2023-05-18T15:50:44+00:00" }, { "id": "01a0cac1-09f7-71e7-8156-fb9a01b0acbb", "title": "On Device AI – Double-Edged Sword", "url": "https://semianalysis.com/2023/05/13/on-device-ai-double-edged-sword/", "published_at": "2023-05-13T10:05:21+00:00" } ] posts Claim your blog
Back to SemiAnalysis
Blog · corpus.blog/blogs/semianalysis.com/posts

SemiAnalysis

semianalysis.com

2025

xAI’s Colossus 2 – First Gigawatt Datacenter In The World, Unique RL Methodology, Capital Raiseoriginal ↗
16 Sept 2025
Another Giant Leap: The Rubin CPX Specialized Accelerator & Rackoriginal ↗
10 Sept 2025
Huawei Ascend Production Ramp: Die Banks, TSMC Continued Production, HBM is The Bottleneckoriginal ↗
8 Sept 2025
Amazon’s AI Resurgence: AWS & Anthropic’s Multi-Gigawatt Trainium Expansionoriginal ↗
3 Sept 2025
H100 vs GB200 NVL72 Training Benchmarks – Power, TCO, and Reliability Analysis, Software Improvement Over Timeoriginal ↗
20 Aug 2025
GPT-5 Set the Stage for Ad Monetization and the SuperApporiginal ↗
13 Aug 2025
Scaling the Memory Wall: The Rise and Roadmap of HBMoriginal ↗
12 Aug 2025
Robotics Levels of Autonomyoriginal ↗
30 Jul 2025
Intel 18A Details & Cost, Future of DRAM 4F2 vs 3D, Backside Power Adoption (or Not), China’s FlipFET, Digital Twins from Atoms to Fabs, and Moreoriginal ↗
21 Jul 2025
Meta Superintelligence – Leadership Compute, Talent, and Dataoriginal ↗
11 Jul 2025
Introduction to AI Networking Modeloriginal ↗
11 Jul 2025
DeepSeek Debrief: >128 Days Later original ↗
3 Jul 2025
How Oracle Is Winning the AI Compute Marketoriginal ↗
30 Jun 2025
AI Training Load Fluctuations at Gigawatt-scale – Risk of Power Grid Blackout?original ↗
25 Jun 2025
NVIDIA Tensor Core Evolution: From Volta To Blackwelloriginal ↗
23 Jun 2025
AMD Advancing AI: MI350X and MI400 UALoE72, MI500 UAL256original ↗
13 Jun 2025
The New AI Networks | Ultra Ethernet UEC | UALink vs Broadcom Scale Up Ethernet SUEoriginal ↗
11 Jun 2025
Scaling Reinforcement Learning: Environments, Reward Hacking, Agents, Scaling Dataoriginal ↗
8 Jun 2025
AMD vs NVIDIA Inference Benchmark: Who Wins? – Performance & Cost Per Million Tokensoriginal ↗
23 May 2025
AI Arrives In The Middle East: US Strikes A Deal with UAE and KSAoriginal ↗
16 May 2025
Microsoft’s Datacenter Freeze – 1.5GW Self-Build Slowdown & Lease Cancellation Misconceptionsoriginal ↗
28 Apr 2025
AMD 2.0 – New Sense of Urgency | MI450X Chance to Beat Nvidia | Nvidia’s New Moatoriginal ↗
23 Apr 2025
Huawei AI CloudMatrix 384 – China’s Answer to Nvidia GB200 NVL72original ↗
16 Apr 2025
High-NA is Here (for R&D), EUV Cost, Pattern Shaping Gaining Share, 6×12″ Mask, Metal Oxide & Dry Resist, Hyper-NAoriginal ↗
14 Apr 2025
Tariff Armageddon? | GPU Loopholes, Mexico Supply Chain Shift, Wafer Fab Equipment Vulnerabilities, Optical Module Pricing Surge, Datacenter Equipmentoriginal ↗
10 Apr 2025
The GPU Cloud ClusterMAX™ Rating System | How to Rent GPUsoriginal ↗
26 Mar 2025
NVIDIA GTC 2025 – Built For Reasoning, Vera Rubin, Kyber, CPO, Dynamo Inference, Jensen Math, Feynmanoriginal ↗
19 Mar 2025
America Is Missing The New Labor Economy – Robotics Part 1original ↗
11 Mar 2025
Datacenter Anatomy Part 2 – Cooling Systemsoriginal ↗
13 Feb 2025
TSMC N2 + Next-Gen SoIC, Intel EMIB-T, Meta 3D Stacked Memory, CFET, 2D Materials, and Moreoriginal ↗
5 Feb 2025
DeepSeek Debates: Chinese Leadership On Cost, True Training Cost, Closed Model Margin Impactsoriginal ↗
31 Jan 2025
OpenAI Stargate Joint Venture Demystified | Microsoft Sore Loser, Does Softbank Have The Capital?, Texas GigaCampus, Winners & Losersoriginal ↗
23 Jan 2025
2025 AI Diffusion Export Controls – Microsoft Regulatory Capture, Oracle Tears, Impacts Quantified, Model Restrictionsoriginal ↗
15 Jan 2025

2024

Nvidia’s Christmas Present: GB300 & B300 – Reasoning Inference, Amazon, Memory, Supply Chainoriginal ↗
25 Dec 2024
MI300X vs H100 vs H200 Benchmark Part 1: Training – CUDA Moat Still Aliveoriginal ↗
22 Dec 2024
Scaling Laws – O1 Pro Architecture, Reasoning Training Infrastructure, Orion and Claude 3.5 Opus “Failures”original ↗
11 Dec 2024
Intel on the Brink of Death | Culture Rot, Product Focus Flawed, Foundry Must Surviveoriginal ↗
9 Dec 2024
Amazon’s AI Self Sufficiency | Trainium2 Architecture & Networkingoriginal ↗
3 Dec 2024
Fab Whack-A-Mole: Chinese Companies are Evading U.S. Sanctionsoriginal ↗
28 Oct 2024
Datacenter Anatomy Part 1: Electrical Systemsoriginal ↗
14 Oct 2024
AI Neocloud Playbook and Anatomyoriginal ↗
3 Oct 2024
Clash of the Foundries: Gate All Around + Backside Power at 2nmoriginal ↗
1 Oct 2024
Multi-Datacenter Training: OpenAI’s Ambitious Plan To Beat Google’s Infrastructureoriginal ↗
4 Sept 2024
The Memory Wall: Past, Present, and Future of DRAMoriginal ↗
3 Sept 2024
Nvidia’s Blackwell Reworked – Shipment Delays & GB200A Reworked Platformsoriginal ↗
4 Aug 2024
GB200 Hardware Architecture – Component Supply Chain & BOMoriginal ↗
17 Jul 2024
GB200 & GB200A Component & Supply Chain Modeloriginal ↗
16 Jul 2024
100,000 H100 Clusters: Power, Network Topology, Ethernet vs InfiniBand, Reliability, Failures, Checkpointingoriginal ↗
17 Jun 2024
OpenAI Chip Team Is Now Seriousoriginal ↗
3 Jun 2024
How Dell Is Beating Supermicrooriginal ↗
30 May 2024
Apple’s AI Strategy: Apple Datacenters, On-device, Cloud, And Moreoriginal ↗
27 May 2024
OpenAI Is Doomed? – Et tu, Microsoft?original ↗
7 May 2024
Intel’s 14A Magic Bullet: Directed Self-Assembly (DSA)original ↗
18 Apr 2024
Nvidia Blackwell Perf TCO Analysis – B100 vs B200 vs GB200 NVL72original ↗
10 Apr 2024
Is Intel Back? Foundry & Product Resurgence Measuredoriginal ↗
2 Apr 2024
Nvidia’s Optical Boogeyman – NVL72, Infiniband Scale Out, 800G & 1.6T Ramporiginal ↗
25 Mar 2024
Nvidia B100, B200, GB200 – COGS, Pricing, Margins, Ramp – Oberon, Umbriel, Mirandaoriginal ↗
18 Mar 2024
Astera Labs IPO – The Next Connectivity Superhero or Steamrolled By Competition?original ↗
17 Mar 2024
CXL Is Dead In The AI Eraoriginal ↗
16 Mar 2024
AI Datacenter Energy Dilemma – Race for AI Datacenter Spaceoriginal ↗
13 Mar 2024
Groq Inference Tokenomics: Speed, But At What Cost?original ↗
21 Feb 2024
Hybrid Bonding Process Flow – Advanced Packaging Part 5original ↗
9 Feb 2024
Compliance Policiesoriginal ↗
3 Feb 2024
Neural Network Quantization & Number Formats From First Principlesoriginal ↗
11 Jan 2024
Intel GenAI For Yield, TSMC CFET & 3D Stacking, AMD 3D Device Modeling, Applied Materials Material Innovation, SK Hynix HBM4, Micron 3D DRAM & FeRAM, Hybrid Bonding vs TCB – IEDM 2023original ↗
3 Jan 2024

2023

Inference Race To The Bottom – Make It Up On Volume?original ↗
18 Dec 2023
ASML Dilemma: High-NA EUV is Worse vs Low-NA EUV Multi-Patterningoriginal ↗
11 Dec 2023
AMD MI300 Performance – Faster Than H100, But How Much?original ↗
6 Dec 2023
GPU Cloud Economics Explained – The Neocloud Hidden Truthoriginal ↗
4 Dec 2023
Microsoft Swallows OpenAI’s Core Team – GPU Capacity, Incentive Structure, Intellectual Property, OpenAI Rump Stateoriginal ↗
20 Nov 2023
Microsoft Infrastructure – AI & CPU Custom Silicon Maia 100, Athena, Cobalt 100original ↗
15 Nov 2023
Nvidia’s New China AI Chips Circumvent US Restrictions | H20 Faster Than H100 | Huawei Ascend 910Boriginal ↗
9 Nov 2023
AMD MI300 Ramp, GPT-4 Performance, ASP & Volumesoriginal ↗
1 Nov 2023
Wafer Wars: Deciphering Latest Restrictions On AI And Semiconductor Manufacturingoriginal ↗
24 Oct 2023
Going Vertical: Gate All Around, 3D DRAM, 3D NAND – Kokusai Electric IPOoriginal ↗
15 Oct 2023
Nvidia’s Plans To Crush Competition – B100, “X100”, H200, 224G SerDes, OCS, CPO, PCIe 7.0, HBM3Eoriginal ↗
10 Oct 2023
AI Doomer vs Techno-Optimist: Social Fabric, AI Girlfriends, Mass Mediaoriginal ↗
5 Oct 2023
Amazon Anthropic: Poison Pill or Empire Strikes Backoriginal ↗
2 Oct 2023
Nvidia Buys Illumina? The AI Foundry for Healthcare – The Hardware of Lifeoriginal ↗
27 Sept 2023
Arm and a Leg: Arm’s Quest To Extract Their True Valueoriginal ↗
14 Sept 2023
China AI & Semiconductors Rise: US Sanctions Have Failedoriginal ↗
12 Sept 2023
TPUv5e: The New Benchmark in Cost-Efficient Inference and Training fororiginal ↗
1 Sept 2023
Broadcom’s Google TPU Revenue Explosion, Networking Boom, VMWare Integrationoriginal ↗
30 Aug 2023
Google Gemini Eats The World – Gemini Smashes GPT-4 By 5X, The GPU-Poorsoriginal ↗
28 Aug 2023
Nvidia’s Optical Ascent: >$1B Revenue; The Missing 800G Ramp; AI Head-Fakesoriginal ↗
23 Aug 2023
Nvidia’s Ramp – Volume, ASP, Cloud Pricing, Margins, EPS, Cashflow, China, Competitionoriginal ↗
20 Aug 2023
Energizing AI: Power Delivery Competition Heats Up Vicor, MPS, Delta, ADI, Renesas, Infineonoriginal ↗
1 Aug 2023
AI Expansion – Supply Chain Analysis For CoWoS And HBMoriginal ↗
26 Jul 2023
NAND Flash Monopoly Broken? Tokyo Electron Moly Dep + Cryo Etch Takes On Lam Research For The Future Of NANDoriginal ↗
16 Jul 2023
GPT-4 Architecture, Infrastructure, Training Dataset, Costs, Vision, MoEoriginal ↗
10 Jul 2023
AI Capacity Constraints – CoWoS and HBM Supply Chainoriginal ↗
5 Jul 2023
Tesla AI Capacity Expansion – H100, Dojo D1, D2, HW 4.0, X.AI, Cloud Service Provideroriginal ↗
27 Jun 2023
Rebuilding Intel – Foundry vs IDM Decades of Inefficiencies Unraveledoriginal ↗
22 Jun 2023
AMD MI300 – Taming The Hype – AI Performance, Volume Ramp, Customers, Cost, IO, Networking, Softwareoriginal ↗
12 Jun 2023
Zen 4c: AMD’s Response to Hyperscale ARM & Intel Atomoriginal ↗
5 Jun 2023
AI Server Cost Analysis – Memory Is The Biggest Loseroriginal ↗
29 May 2023
Nvidia’s InfiniBand Problem – Spectrum-X AI Fabric, Tomahawk-5, Jericho-3AI, Quantum-2original ↗
28 May 2023
Meta Custom Silicon: What’s Old Is Neworiginal ↗
19 May 2023
Sound The Siryn: AmpereOne 192-Core CPUoriginal ↗
18 May 2023
On Device AI – Double-Edged Swordoriginal ↗
13 May 2023
Next