Hyperscale
Hyperscale data centres: facilities of typically 100 MW or more operated by major cloud providers, characterised by standardised infrastructure at massive scale.
Last refreshed: 26 April 2026 · Appears in 1 active topic
Have AI training requirements made the original hyperscale design obsolete?
Timeline for hyperscale
Mentioned in: OpenAI pauses Cobalt Park Stargate site
Data Centres: Boom and BacklashMentioned in: Maine passes first US statewide DC freeze
Data Centres: Boom and BacklashIEA: 17% growth, $700B capex run-rate
Data Centres: Boom and BacklashStargate US: $500B headline, 1.2 GW operational
Data Centres: Boom and BacklashMentioned in: Blackstone £10B Blyth, Amazon €33.7B EU
Data Centres: Boom and Backlash- What is a hyperscale data centre?
- A hyperscale data centre is a very large facility — typically 100 MW or more of IT load — designed and operated by a major cloud or technology company. AI training campuses of 2025-2026 are being designed for 500 MW to 1 GW, FAR larger than earlier hyperscale designs.Source: Lowdown data-centres briefing
- How much are hyperscalers spending on data centres in 2026?
- The IEA projects the five largest hyperscale operators will spend approximately $700 billion on data centres in 2026, up 75% from the $400 billion-plus they collectively spent in 2025.Source: IEA April 2026
Background
hyperscale refers to a class of data centre designed to scale computing resources rapidly and efficiently to support massive workloads — principally public cloud services, AI training, and large-scale content delivery. The defining characteristics are size (typically 100 MW or more of IT load), standardised modular design, very high server density, and ownership by a major cloud or technology company rather than a colocation provider.
As of 2025, the five largest hyperscale operators — Amazon, Microsoft, Google, Meta, and Apple — account for the majority of global data centre investment. The IEA's April 2026 report found these five collectively exceeded $400 billion in capex in 2025, expected to rise 75% in 2026 to approximately $700 billion. Synergy Research Group tracks hyperscale facility count and capacity as a market indicator.
The AI training wave has elevated hyperscale's power and land requirements beyond what earlier designs assumed. A conventional hyperscale facility of 2018 vintage might draw 100-150 MW; AI-optimised hyperscale campuses of 2025-2026 are being designed for 500 MW to 1 GW, and the Stargate US programme involves campuses at 1.2 GW. This scale increase is the primary driver of grid-connection constraint and BTM generation demand.