Cloud providers are bringing data center capacity online for higher-priced AI services as enterprise cloud infrastructure spending accelerates, reinforcing demand for construction even as power constraints complicate delivery.. Worldwide spending on cloud infrastructure services reached $168 billion in the third quarter of 2026, up 55% from a year earlier, according to a Synergy Research Group report.
The market has doubled in eight quarters, and AI-specific services now account for roughly one-third of total revenue.. John Dinsdale, chief analyst at Synergy, told Data Center Knowledge that three factors are driving the surge: additional capacity coming online, higher pricing, and a continued shift toward more expensive AI services..
“A lot of the additional capacity that is coming online is specifically targeting more valuable, higher-priced AI services,” he said.. That shift in pricing and service mix helps explain how cloud revenue can grow faster than the data center capacity supporting it. JLL’s August midyear report illustrated the construction demand behind that expansion..
Related:Zombie Workloads Haunt Data Center Efficiency Efforts. Hyperscalers accounted for 59% of North American data center tenant demand, while 28 GW of owner-occupied capacity was under construction, according to JLL. Neoclouds – specialized cloud providers offering GPU-as-a-Service to enterprise customers – represented another 11% of demand, while AI companies accounted for 7%, broadening the customer base competing for facilities..
The buildout is expected to continue. Global data center capital spending is projected to exceed $3 trillion by 2030 as hyperscalers, sovereign AI programs, and specialized cloud providers expand their infrastructure to support growing AI workloads, according to an August report from Dell’Oro Group..
AI Changes the Cloud Customer Mix. The cloud market’s expansion is also reshaping which companies need infrastructure and how they compete. Amazon retained the largest share of worldwide cloud infrastructure revenue at 26%, followed by Microsoft at 19%, and Google at 16%, Synergy reported..
“AI has totally transformed the cloud market in terms of service capabilities, customer demand, and growth opportunities,” Dinsdale said.. Microsoft and Google grew faster than Amazon, while Oracle and neocloud providers, including CoreWeave, Crusoe, and Nebius, were among the fastest-growing companies outside the top three.
Synergy said neocloud expansion is eroding the market shares of some established providers.. For data center operators, that changing mix raises questions about tenant durability and contract stability, particularly if some fast-growing providers consolidate. Dinsdale said he expects the underlying infrastructure to retain value even if an individual neocloud’s prospects change..
Related:The Vanishing Value of VMware: A Wake-Up Call. Construction Demand Persists Despite Power and Tenant Risks. CBRE’s August report on North American data center trends found 7,481 MW under construction across eight primary markets in the first half of 2026, up 24.8% from a year earlier.
More than 80% was preleased, up from 74.3% a year earlier. Power constraints continued to extend completion schedules, leaving customers competing for capacity that has yet to be delivered.. Those commitments mean operators evaluating expansion must weigh construction schedules and tenant stability alongside cloud revenue growth..
Dinsdale said he expects data centers, installed equipment, and existing customer contracts to find new owners if neocloud providers consolidate or fail. That suggests consolidation would redistribute infrastructure rather than leave large amounts of it unused, though the financial outcome would vary significantly by asset..
“Whatever happens to a neocloud company and its prospects for carrying on as an independent entity, there will be valuable assets that will not go to waste,” he said. “Valuations will vary greatly, but I don’t imagine too many stranded assets being trashed.”. Related:Cloud Region Migration: Optimize Costs, Latency, and Compliance
