Blackwell, Decoded

A plain-spoken, interactive course on NVIDIA's Blackwell architecture — the chips, engines, and rack-scale systems built for the age of AI reasoning. It turns the dense 31-page NVIDIA Blackwell Architecture Technical Brief (V2.1) into eight short, visual modules plus a glossary, written for people who know software but are new to AI hardware.

It's a fully static site — no build step, no dependencies, no server required. Just open index.html.

Quick start

Open it directly in a browser:

open index.html          # macOS

Or serve it locally (recommended — keeps progress tracking reliable across browsers):

python3 -m http.server 8000
# then visit http://localhost:8000

Your progress is saved per-device in the browser via localStorage (key nvbw-progress-v1); the site degrades gracefully if storage is blocked (e.g. on file://). Use Reset progress on the home page to clear it.

Contents

#	Module	Section	~Time
01	Why Blackwell?	Foundations	5 min
02	The Three Scaling Laws	Foundations	6 min
03	The Blackwell GPU	The Architecture	8 min
04	The Engines	The Architecture	9 min
05	NVLink & Scaling Up	The Architecture	6 min
06	GB300 NVL72 Rack	The Systems	9 min
07	GB200 & HGX Servers	The Systems	8 min
08	Inference Deep-Dive	Going Deeper	7 min
§	Glossary	Reference	—

Work through them in order for a complete picture, or jump to whatever you need.

Project structure

.
├── index.html              # Course home (overview, progress, module cards)
├── 01-why-blackwell.html   # Module pages 01–08
├── 02-scaling-laws.html
├── 03-blackwell-gpu.html
├── 04-the-engines.html
├── 05-nvlink.html
├── 06-gb300-nvl72.html
├── 07-gb200-and-hgx.html
├── 08-parallelism.html
├── glossary.html           # Term reference
└── assets/
    ├── style.css           # All styling
    ├── app.js              # Progress tracking, nav, glossary tooltips, mobile sidebar
    └── glossary.js         # Glossary term data (powers hover/tap tooltips)

The source PDF (NVIDIA Blackwell Architecture Technical Overview) is kept locally for reference but is not tracked in git — see .gitignore. Download it from NVIDIA: https://resources.nvidia.com/en-us-blackwell-architecture

Source & disclaimer

Educational material generated by Claude from NVIDIA's published Blackwell Architecture Technical Brief (V2.1). Figures are NVIDIA projections "subject to change" and are reproduced here for learning. All trademarks belong to NVIDIA Corporation. This project is not affiliated with or endorsed by NVIDIA.