Industries · Technology
Data centres, cloud and AI compute
How data centres, cloud providers and AI compute businesses work: colocation versus hyperscale, why power and cooling are now the limit, what one megawatt and one rack earn and cost, how a GPU hour is priced, the capex cycle, and the sovereign AI build-outs in the Gulf, India and Europe.
Key takeaways
- A data centre turns electricity into computing.
- Two businesses sit in one building. Colocation earns rent per kW for 10 to 15 years on a building that lasts decades, so it looks like property.
- Data centre cases are usually bets on three things at once: will the demand be there, can you get the power in time, and will the price still cover a very large investment when the capacity arrives?
- Explain the data centre value chain from land and power to cloud and AI services
- Tell colocation, hyperscale and GPU cloud (neocloud) business models apart
- Calculate revenue and return per megawatt, racks from a power budget, and profit per GPU hour
- Explain why power and cooling are the binding constraint, and what that does to site choice
- Crack typical cases on site selection, pricing, capacity and sovereign AI
Lessons
How data centres, cloud and AI compute work
What a data centre is, who builds and rents it, how cloud and AI compute sit on top, and why power is the scarce input.
Data centre economics: per MW, per rack and per GPU hour
Work out rent and return per megawatt for colocation, racks from a power budget, and profit per GPU hour for an AI cloud, and see why utilisation, price and chip life decide everything.
Data centres and AI compute: players, trends and cases
Who the players are by region, the 2024 to 2026 build-out and sovereign AI programmes, regulation basics, and typical case prompts.
Worked cases in this module
Look it up
Key terms