Generic data center administrator image
– Getty Images

Neocloud revenue is scaling at an “unprecedented pace,” reaching $9 billion during the fourth quarter of 2025, and exceeding $25 billion for the full year, Synergy Research Group (SRG) has found.

In its latest neocloud market forecast, SRG said the neocloud market, which was up 223% year-over-year in Q4, will approach $400 billion by 2031, with a sustained 58% compound annual growth rate (CAGR).

Synergy Research Neoclouds 9 bil Q4 2025
– Synergy Research Group

SRG noted that AI demand is outstripping hyperscale capacity, leading to rapid expansion for neocloud providers, who are increasing their market share.

CoreWeave, Crusoe, Core Scientific, Lambda, Nebius, and Nscale lead the pack, with CoreWeave “standing out as the most direct challenger to traditional hyperscale cloud providers.”

However, there is also an influx of newer entrants, with notable buy-in from transitioning crypto miners as profitability issues paint AI infrastructure as a more lucrative avenue.

SRG found neoclouds are positioning themselves as “focused alternatives to traditional hyperscale providers,” and are “reshaping the competitive boundary between infrastructure and platform layers in the cloud ecosystem.”

In research from late last year, SRG found that, while a significant gap still exists, neoclouds are gradually claiming market share from cloud infrastructure services.

“What we are observing is not merely the emergence of a new class of cloud provider, but a deeper structural realignment in the architecture of computation itself,” Jeremy Duke, founder and chief analyst at SRG, noted. “Traditional hyperscale systems were conceived around a form of generalized elasticity, whereas AI workloads impose far more rigid constraints – particularly around parallelism, locality, and the concentration of compute. Neoclouds are, in effect, an architectural response to those constraints. As AI moves from exploratory phases into sustained, large-scale deployment, these underlying differences cease to be incidental and instead become determinative of how compute systems evolve.”