Learn · track 1 of 4
The stack
An AI datacentre is a building, a connection to the grid, a cooling plant, racks of accelerators and the network that joins them into one machine. Each layer has its own suppliers, its own lead times and its own ways of failing, and money in this sector tends to follow whichever layer is scarce at the time.
This track takes the stack from the ground up. It starts with what a datacentre is and how an AI campus differs from the cloud halls built before it. It then covers the two kinds of business that build and run them: the neocloud, which rents out GPU clusters by the hour or under contract, and the AI factory, a datacentre run as a production line for tokens. The last pieces cover power and cooling, which decide how many chips a site can run, and networking and optics, which decide how well those chips work together.
The scale of the build-out shows in the filings: Microsoft, Alphabet, Amazon and Meta paid $201.9B
SEC EDGAR 10-Q/10-K (XBRL): purchases of property and equipment, cash-flow statement (calc.)As of 30 Oct 2026Source
Sources
- AI Infrastructure Primer 2026, Compute Terminal.
- Microsoft, Alphabet, Amazon and Meta: quarterly and annual reports (10-Q and 10-K, SEC EDGAR), purchases of property and equipment; how we compute it.
1 · 5 min read
What is a datacentre?
What a datacentre is, the parts every one shares, how its capacity is measured and what changes when the tenant is an AI cluster.
2 · 4 min read
What is a neocloud?
A neocloud is a cloud provider built around one product: large, tightly networked pools of the latest AI accelerators, rentable with minimal friction.
3 · 3 min read
What is an AI factory?
The phrase AI factory — popularised by NVIDIA and now standard industry vocabulary — is more than marketing. It describes a datacentre engineered as a production line: the raw materials are electricity and data, the machinery is racked accelerated computing, and the output is tokens — the units of text, code, images and actions that AI models consume and produce.
4 · 6 min read
Power and cooling
How electricity gets from the grid to a GPU, what is lost on the way, and how the heat gets back out: the two systems that decide how many chips a site can run.
5 · 6 min read
Networking and optics
How thousands of GPUs are joined into one machine: copper inside the rack, fibre between racks, the fabric wars and the optics where the money and the watts went.
Next track: The economics → · Glossary · Start here
Free · every trading day · 08:00 ET