SemiAnalysis
SemiAnalysis is an independent research and analysis firm covering the semiconductor and AI supply chain end to end, from capital equipment, fabs, foundries, memory and chip design through networking, datacenters, software, and AI training/inference. It bridges business decisions and technical industry realities through primary research, institutional models, open benchmarks, a newsletter, and consulting/advisory work.
At a Glance
- Institutional investors, hedge funds, private equity, and venture investors
- Semiconductor manufacturers, chip designers, foundries, memory and equipment companies
- Hyperscalers, AI labs, GPU clouds, neoclouds, and datacenter operators
- Enterprises deploying AI infrastructure
- +2 more
AI Tools by SemiAnalysis
(1)
InferenceX
Open Source LLM Inference Benchmarks
Discussions
No discussions yet
Be the first to start a discussion about SemiAnalysis
Latest News
Beijing Will Not Pace the Frontier: China’s Speed-First AI Safety Regime
Anthropic Subscriptions Offer 5x+ More Value Than OpenAI
How GLM5.3 Sparse Attention Affects HBM Memory Usage
Intel Panther Lake Teardown
Products & Services
Free and paid long-form research publication on semiconductors, AI infrastructure, and the broader supply chain, published approximately every two weeks.
Weekly and ad hoc institutional intelligence connecting semiconductor and AI developments to company, supply-chain, and investment implications.
Institutional model forecasting accelerator and HBM shipments, ASPs, revenue, wafer starts, packaging, yields, and installed base by SKU, company, and customer.
Total-cost-of-ownership toolkit for GPU clouds covering server $/hour, training $/PFLOP, inference $/million tokens, deployment scenarios, financing, and financial returns.
Market Position
SemiAnalysis positions itself as independent, product-first, and bottoms-up, differentiating from sell-side and traditional industry research by reconciling primary supply-chain research with quantitative models and frequent updates. Its competitive set includes specialist semiconductor and technology research providers such as TechInsights, Omdia, TrendForce, and The Linley Group, as well as independent technology/AI research publications and benchmarking projects; InferenceX competes specifically with static or vendor-controlled AI inference benchmarks by emphasizing continuous, multi-vendor, reproducible, real-world measurements.
Leadership
Founders
Dylan Patel
Founder and CEO. He started building SemiAnalysis in 2020 as a solo publication and grew it into a research and advisory firm; Crunchbase lists him as the founder, CEO, and chief analyst and identifies the University of Georgia as his alma mater.
Executive Team
Dylan Patel
Founder and CEO
Founded SemiAnalysis in 2020 and built it from a solo publication into a team covering semiconductors, AI infrastructure, cloud economics, and frontier models.
Founding Story
Dylan Patel began publishing SemiAnalysis in 2020 to close the information gap between what semiconductor and AI companies say and what the underlying supply chain shows. The initial vision was product-first, bottoms-up analysis grounded in primary research, field work, teardowns, and quantitative models rather than vendor guidance or conventional sell-side reporting.
Business Model
Revenue Model
A mixed model of free and paid newsletter subscriptions; separately scoped institutional subscriptions and licenses for Core Research and industry models; paid data/API access; and retained advisory, consulting, and technical due-diligence engagements.
Pricing Tiers
Public/free newsletter access.
Subscriber access to the full newsletter article archive and paid research; group and corporate subscriptions are available.
Core Research, industry models, dashboards, and API access are scoped and priced by the sales team for the customer use case and licensing requirements.
Retained advisory, bespoke projects, and hourly or project-based technical engagements, scoped through sales.
Target Markets
- Institutional investors, hedge funds, private equity, and venture investors
- Semiconductor manufacturers, chip designers, foundries, memory and equipment companies
- Hyperscalers, AI labs, GPU clouds, neoclouds, and datacenter operators
- Enterprises deploying AI infrastructure
- Government and policy organizations
- Engineers, researchers, and industry professionals following AI and semiconductors
- Investment research and diligence for funds and institutional investors
- AI infrastructure and GPU-cloud build, pricing, procurement, and deployment decisions
- Semiconductor company strategy, product planning, supply-chain forecasting, and competitive analysis
- Hyperscaler, AI-lab, enterprise, neocloud, and sovereign-AI capacity planning
- Benchmarking inference performance, cost, power, and interactivity across chips and software stacks
- Government and policy analysis of semiconductor manufacturing, datacenters, export controls, and AI infrastructure
- OpenAI and Microsoft are publicly identified as supporters/users of the open InferenceMAX/InferenceX benchmarking initiative.
- NVIDIA, AMD, Oracle Cloud Infrastructure, CoreWeave, Crusoe, Nebius, TensorWave, Together AI, Supermicro, Vultr, PyTorch Foundation, vLLM, and SGLang are publicly identified as supporters, collaborators, compute-resource providers, or ecosystem participants in the InferenceMAX/InferenceX initiative.