Morgan Stanley Forecasts 32 GW of New Hyperscaler and Neocloud Data Center Capacity in 2027
Редакция ЦОД:Главное
Morgan Stanley forecasts that hyperscalers and neocloud providers will add 32 GW of new data center computing capacity in 2027, up 48% year on year. The bank expects Google, Amazon, Microsoft and Meta to account for roughly 85% of that new construction.
- Morgan Stanley projects 32 GW of new hyperscaler and neocloud data center computing capacity in 2027, up 48% year on year.
- Google, Amazon, Microsoft and Meta are forecast to account for about 85% of 2027 new construction.
- The bank forecasts an additional 39 GW of capacity additions in 2028, representing 21% year-on-year expansion.
- The projected build-out could intensify competition for power, grid interconnections, suitable sites and critical data center equipment.
- The estimates are forecasts, not evidence of commissioned capacity or a fully confirmed project pipeline.
Morgan Stanley forecasts a sharp acceleration in data center capacity additions by hyperscalers and neocloud providers, projecting 32 GW of new computing capacity in 2027. That would represent a 48% year-on-year increase, according to the bank’s estimate.
Google, Amazon, Microsoft and Meta are expected to contribute about 85% of new construction in 2027. The projection therefore implies a highly concentrated build-out cycle, led by the four largest hyperscale platforms rather than broadly distributed demand across the provider market.
For 2028, Morgan Stanley forecasts a further 39 GW of new capacity additions, equivalent to 21% year-on-year expansion. If the projections materialise, the scale of deployment could sustain substantial demand across the data center infrastructure stack, including development sites, grid infrastructure, generation, cooling and critical equipment.
The figures are forecasts rather than reported commissioned capacity or a confirmed construction pipeline. Morgan Stanley’s estimate does not provide a geographic split, project-level detail, facility delivery schedules or its underlying methodology. The source also does not specify whether the GW figures refer to IT load, electricity consumption or another measure of computing capacity.
If realised, the forecast points to an exceptionally large global data center expansion cycle, with substantial demand for sites, grid infrastructure, power generation, cooling systems and critical equipment. It also suggests that much of this demand could be concentrated among four major hyperscalers, increasing pressure on power availability, land, grid interconnections and supply chains.
- — / 16 октября 2026 г. · КитайскийПервоисточник →
Материал подготовлен редакцией ЦОД: Главное на основе указанных первоисточников.