Chip production
Examines equipment, materials, foundries, fabrication, and packaging to connect manufacturing capabilities with capacity and semiconductor production costs.
Newsletter directory
01 02 03 All newsletters
Perimeter Reading / Technology
SemiAnalysis
SemiAnalysis research team
Detailed research on semiconductors and AI infrastructure, connecting chip design, manufacturing, datacenters, and model economics with the businesses and supply chains behind them.
Subscriptions are managed by the publisher. Saving to your Desk keeps this profile in your research library.
The publication
What to expect.
SemiAnalysis follows the semiconductor and AI industries from physical production through the systems and models that use the resulting compute. Its research links technical choices to capacity, costs, and commercial consequences, drawing on supply chain work and industry engagement. The newsletter covers equipment, fabrication, packaging, networking, datacenters, and training or inference economics. Institutional models and consulting are separate offerings. This is a demanding read for anyone who needs to understand the infrastructure beneath technology companies' investment plans and product claims.
Who it’s for
Technology investors, engineers, and industry operators who need detailed explanations of semiconductor supply chains, AI infrastructure constraints, and the economics of large compute systems.
Inside the reading
Its areas of focus.
Compute infrastructure
Analyzes accelerators, networking, server systems, power, and cooling as interdependent parts of the expanding AI datacenter supply chain.
AI economics
Studies model training, inference performance, and software choices to explain the cost and efficiency of delivering AI capabilities.
The Desk’s reading note
Make the read useful.
Keep a glossary beside the article and trace each technical claim through to its effect on capacity, cost, or performance.
Keep a question, an idea to investigate, or a source to return to in your saved notes.
Add another perspective
