Loading Thinker Node
正在读取方法论、关键决策和影响关系。
The independent analyst who traces the semiconductor supply chain down to its raw chemicals
Dylan Patel is the most influential independent semiconductor analyst of the AI-compute era and the Founder, CEO, and Chief Analyst of the research firm SemiAnalysis. He holds no engineering degree: rejected by MIT and Stanford, he graduated from the University of Georgia's Terry College of Business in data analysis, risk management, and legal studies. As a child working night shifts at his immigrant parents' rural Georgia motel, he taught himself semiconductor process technology and supply chains on chip-geek forums. On May 22, 2020 — his 24th birthday — he launched SemiAnalysis as a solo Substack post, growing it from a one-person venture into a multi-country research firm serving hedge funds, venture capital, hyperscalers, and chip manufacturers. His core method is tracing the supply chain from Austrian chemical suppliers all the way to the GPU in a datacenter, using satellite imagery to track datacenter construction and TCO (total cost of ownership) models to force executives to confront math they would rather avoid. In May 2023 he published the leaked internal Google memo 'We Have No Moat' and introduced the now-standard 'GPU-rich vs GPU-poor' framework. He appears frequently on Lex Fridman, BG2 (Bill Gurley & Brad Gerstner), Dwarkesh, Latent Space, and Stratechery. His business model centers on deep reports and data products sold to institutional clients, though the blurred line between research and investing has drawn controversy.
This page shows the indexable preview. The complete profile includes methodology steps, anti-patterns, decision context, reasoning, outcomes, and relationship graph data for API or MCP workflows.
Use Dylan Patel's mental models to analyze this product decision.
Which Dylan Patel methodology fits a strategy tradeoff?
Compare Dylan Patel's approach with another thinker for my case.
Having no engineering degree and being rejected by elite schools is no barrier to becoming the most authoritative voice in a field. Real authority comes from understanding something more deeply than the people who are supposed to — reading primary sources, building models by hand, talking to engineers at every layer of the supply chain, and writing about it relentlessly until the whole industry has to take notice.
In the AI race, outcomes are decided less by small algorithmic differences than by who has more and more-efficient compute. His 'GPU-rich vs GPU-poor' lens splits the entire industry: the acquisition, deployment, and unit economics of compute predict an organization's fate more reliably than the model itself.
Everyone watches the GPU, but what makes GPUs possible — and what will stop them — sits upstream: TSMC's advanced logic capacity, HBM high-bandwidth memory, CoWoS packaging, grid power, and at the very bottom ASML's EUV lithography machines. To understand the limits of AI-compute scaling you must walk the supply chain to find the single most-constrained 'lowest rung.'
Classify the entire AI industry into the haves and have-nots by the scale of compute they can command, then infer each player's strategic options from it.
In the 'Gemini Eats The World' report he used the framework to separate 'GPU-rich' players like Google and OpenAI from the many compute-starved startups and labs; the vocabulary has since become industry-standard.
Walk down the supply chain: what constrains the whole system's capacity is always the most-limited rung, not the most visible one.
He frames three bottlenecks to AI-compute scaling — logic (TSMC capacity), memory (HBM), and power — and makes a contrarian call that by 2028-2029 the binding constraint falls to the lowest rung: ASML's EUV lithography output.
Break a chip's or a datacenter's real economics down to the last cent, forcing everyone past the marketing claims to confront the hard math.
After his TCO and performance analysis of AMD's MI300X, AMD CEO Lisa Su reportedly arranged a one-on-one meeting within 24 hours — a blog's cost math forced a chip giant to respond directly.
With no engineering degree, rejected by MIT and Stanford, raised in a rural Georgia motel, he became a chip authority that Jensen Huang, Lisa Su, and Sam Altman all pay attention to. The very condition of 'not being handed the door' pushed him to build the door — and analyze its supply chain — himself.
SemiAnalysis's research is prized for independence and bite, yet Patel personally invests in roughly 20 startups and led a $50M SPV in Fluidstack, with the firm exploring a formal VC fund. The line between research and investing makes 'independence' both his signature and his sharpest controversy.
约2011-2019
Motel night shifts, self-teaching on chip forums, a business education, and a short finance stint
Raised rotating through shifts at his parents' rural motel, the night-shift computer led him to forums on semiconductor process technology and supply chains, sparking an obsession. Rejected by MIT and Stanford, he attended the University of Georgia's Terry College (data analysis, risk management, legal studies), graduated in 2017, briefly worked at a finance firm, and left over how bonuses were distributed.
2020-2022
Building first-hand depth and an audience through the SemiAnalysis blog
Launched SemiAnalysis with a single Substack post on May 22, 2020 (his 24th birthday), with no investors and no team. Through obsessive supply-chain tracing and hyper-specific writing, he gradually built a widely cited reputation across 'Silicon Twitter' and the semiconductor world.
2023-2024
Rising to authority via the 'No Moat' memo, the GPU-rich/poor framework, and public critiques
In May 2023 he published the leaked Google 'We Have No Moat' memo, sparking industry-wide debate; that year his 'GPU-rich vs GPU-poor' framework became standard vocabulary; his sharp critique of AMD's MI300X reportedly led CEO Lisa Su to meet him within 24 hours. He became a frequent guest on high-impact shows like BG2 and Lex Fridman.
Lesson: Real expertise can be born outside institutions, given enough curiosity and first-hand digging.
Lesson: Rejection does not mean the path is blocked; an unconventional skill mix can be rarer and more valuable.
Lesson: Public, free, hyper-specific content is the fastest path to building authority from zero.
The deep, deeply technical hardware-review tradition embodied by the AnandTech founder is a spiritual source for Patel's generation of chip writers: winning professional readers' trust through extreme detail.
Stratechery pioneered the 'independent analyst plus subscription' model that served as a template for SemiAnalysis; Ben Thompson has also publicly called SemiAnalysis one of his most-cited resources.
Author of Fabricated Knowledge and a long-time collaborator at SemiAnalysis, deeply complementary in semiconductor investing and industry analysis.
VCs, hedge funds, and hyperscalers use SemiAnalysis's supply-chain, TCO, and bottleneck analysis as decision inputs, with frameworks like 'GPU-rich/poor' becoming their shared language.
He pushed 'analyst-as-primary-source' (satellite imagery, channel checks, end-to-end tracing) into the mainstream, shaping the methodology of a new generation of independent researchers.
Executives at companies like Nvidia and AMD reportedly respond to and even cite his analysis, making his research an external reference point for vendors' performance claims.
A research scientist at the Allen Institute for AI and author of Interconnects, he joined Patel on shows like Lex Fridman, exchanging deeply at the intersection of models and compute.
Host of the Dwarkesh Podcast (no relation), he has interviewed Patel at length on AI-compute bottlenecks, an important platform for spreading his views.
A well-known venture investor and BG2 co-host, he engages Patel on compute build-out, Nvidia, and capital cycles on the podcast.
It's one of my most important and most-cited resources on Stratechery.
Content I wish I could have written.
