Enterprise investment in cloud infrastructure services surged to $99 billion in Q2 2025, a 25% annual increase, per Synergy Research Group. This growth, up from a 19% average in 2023, is primarily fueled by generative AI systems.
Growth has accelerated from 2023, when it averaged 19%, with the systems that underpin generative AI emerging as a key driver.
The report noted that since early 2023, cloud providers have seen quarterly revenues grow by $36 billion, with infrastructure-as-a-service (IaaS) and platform-as-a-service (PaaS) services accounting for the bulk of that growth, up 27% in the most recent quarter.
Drilling into the Data
Amazon maintains a commanding 30% global market share through its AWS cloud portfolio, followed by Microsoft (Azure) at 20% and Google (Google Cloud) at 13%, while specialized players such as CoreWeave, Oracle, Databricks, and Huawei posted some of the highest growth rates.
From a standing start two years ago, CoreWeave has already surpassed $1 billion in quarterly cloud revenue, placing it just outside the top dozen providers.
Synergy estimates that trailing 12-month cloud infrastructure revenues have now reached $366 billion, with growth distributed across all global regions.
The report found the US remains the largest and fastest-growing market, expanding by 25% in Q2. In Europe, the UK and Germany continue to dominate in size, while Ireland, Spain, and Italy are among the fastest-growing markets.
Outside of Europe, Brazil, India, Australia, Indonesia, and Mexico all exceeded global averages when measured in local currencies, Synergy said.
Neoclouds Gather on the Horizon
While hyperscale providers remain dominant, a new segment of specialized neocloud providers is rapidly reshaping the competitive landscape. These GPU-focused, AI-optimized infrastructure providers saw revenues exceed $5 billion in Q2, up 205% year over year.
These providers differentiate themselves from traditional hyperscalers by focusing on GPU-as-a-Service (GPUaaS), generative AI platforms, and high-capacity, AI-optimized data centers. They often serve industries that require extreme compute density.
According to Synergy, the neocloud segment is projected to generate $23 billion in 2025, reaching $180 billion in annual revenue by 2030 and growing at an average rate of 69% per year.
“It is a rapidly growing market,” Synergy’s chief analyst and research director, John Dinsdale, told DCN. “Even as neoclouds grow aggressively, others will continue to grow quickly too. Market shares may move, but this market will support strong growth across a broad range of companies.”
A September report from JLL found the global neocloud segment is projected to grow at an 82% annual rate from 2021 to 2025 as enterprises compete for GPU capacity – highlighting hyperscale infrastructure cannot currently keep pace with AI demand, creating bottlenecks in availability.
The report noted that AI workloads, drawing over 100 kW per rack, require specialized cooling and floor loading capacities, which pushes data centers to upgrade.
Leading neoclouds include CoreWeave, Crusoe, Lambda, Nebius, and OpenAI, with OpenAI’s consumer-facing services and infrastructure investments – particularly its Stargate initiative – making it the largest in the category.
Fierce Competition for Infrastructure
The Synergy report notes that while neoclouds face “fierce competition,” their focus on AI workloads gives them an edge in high-growth segments. Many are repurposing infrastructure from crypto-mining or launching as startups specialized in high-performance computing, allowing them to scale quickly.
“They are developing much more power-dense facilities and also locating them well away from expensive metro markets with power constraints,” Dinsdale explained. “Of course, hyperscale operators are doing some of that too.”
A longer tail of emerging players – including Applied Digital, Northern Data Group, Together AI, and WhiteFiber – is also entering the market.
With traditional hyperscalers expanding their own GPU and AI offerings, the line between cloud and neocloud is beginning to blur.
However, Dinsdale said the neocloud sector’s narrow specialization and aggressive growth trajectory suggest it will continue to gain share in the fastest-expanding areas of digital infrastructure – particularly those driving the next wave of AI development.
As the market develops, he says all the usual factors are at play: availability and cost of real estate and power, networking infrastructure, ease of doing business, local financial incentives, political stability, availability of local skilled workforce, and proximity to customers.
“Given the nature of many AI workloads, proximity to customers and latency issues may sometimes be less important than they once were,” Dinsdale said.
