Sitemap - 2024 - SemiAnalysis

Fab Whack-A-Mole: Chinese Companies are Evading U.S. Sanctions

Datacenter Anatomy Part 1: Electrical Systems

AI Neocloud Playbook and Anatomy

Clash of the Foundries: Gate All Around + Backside Power at 2nm

Multi-Datacenter Training: OpenAI's Ambitious Plan To Beat Google's Infrastructure

The Memory Wall: Past, Present, and Future of DRAM

Nvidia's Blackwell Reworked - Shipment Delays & GB200A Reworked Platforms

GB200 Hardware Architecture - Component Supply Chain & BOM

GB200 & GB200A Component & Supply Chain Model

100,000 H100 Clusters: Power, Network Topology, Ethernet vs InfiniBand, Reliability, Failures, Checkpointing

OpenAI Chip Team Is Now Serious

How Dell Is Beating Supermicro

Apple’s AI Strategy: Apple Datacenters, On-device, Cloud, And More

OpenAI Is Doomed? - Et tu, Microsoft?

Wafer Fab Model

AI Cloud TCO Model

Intel’s 14A Magic Bullet: Directed Self-Assembly (DSA)

Nvidia Blackwell Perf TCO Analysis - B100 vs B200 vs GB200NVL72

Is Intel Back? Foundry & Product Resurgence Measured

Nvidia’s Optical Boogeyman – NVL72, Infiniband Scale Out, 800G & 1.6T Ramp

Nvidia B100, B200, GB200 - COGS, Pricing, Margins, Ramp - Oberon, Umbriel, Miranda

Astera Labs IPO - The Next Connectivity Superhero or Steamrolled By Competition?

CXL Is Dead In The AI Era

AI Datacenter Energy Dilemma - Race for AI Datacenter Space

Groq Inference Tokenomics: Speed, But At What Cost?

Hybrid Bonding Process Flow - Advanced Packaging Part 5

Datacenter Industry Model

Compliance Policies

Accelerator Industry Model

Neural Network Quantization & Number Formats From First Principles

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