Enterprise spending on cloud infrastructure providers climbed to $99 billion within the second quarter of 2025, in accordance with knowledge from Synergy Analysis Group, marking a 25% year-over-year improve.
Progress has accelerated from 2023, when it averaged 19%, with the methods that underpin generative AI rising as a key driver.
The report famous that since early 2023, cloud suppliers have seen quarterly revenues develop by $36 billion, with infrastructure-as-a-service (IaaS) and platform-as-a-service (PaaS) providers accounting for the majority of that development, up 27% in the newest quarter.
Drilling into the Knowledge
Amazon maintains a commanding 30% world market share by its AWS cloud portfolio, adopted by Microsoft (Azure) at 20% and Google (Google Cloud) at 13%, whereas specialised gamers reminiscent of CoreWeave, Oracle, Databricks, and Huawei posted a few of the highest development charges.
From a standing begin two years in the past, CoreWeave has already surpassed $1 billion in quarterly cloud income, putting it simply exterior the highest dozen suppliers.
Synergy estimates that trailing 12-month cloud infrastructure revenues have now reached $366 billion, with development distributed throughout all world areas.
The report discovered the US stays the most important and fastest-growing market, increasing by 25% in Q2. In Europe, the UK and Germany proceed to dominate in measurement, whereas Eire, Spain, and Italy are among the many fastest-growing markets.
Outdoors of Europe, Brazil, India, Australia, Indonesia, and Mexico all exceeded world averages when measured in native currencies, Synergy stated.
Neoclouds Collect on the Horizon
Whereas hyperscale suppliers stay dominant, a brand new phase of specialised neocloud suppliers is quickly reshaping the aggressive panorama. These GPU-focused, AI-optimized infrastructure suppliers noticed revenues exceed $5 billion in Q2, up 205% 12 months over 12 months.
These suppliers differentiate themselves from conventional hyperscalers by specializing in GPU-as-a-Service (GPUaaS), generative AI platforms, and high-capacity, AI-optimized knowledge facilities. They typically serve industries that require excessive compute density.
In keeping with Synergy, the neocloud phase is projected to generate $23 billion in 2025, reaching $180 billion in annual income by 2030 and rising at a median fee of 69% per 12 months.
“It’s a quickly rising market,” Synergy’s chief analyst and analysis director, John Dinsdale, informed DCN. “At the same time as neoclouds develop aggressively, others will proceed to develop shortly too. Market shares could transfer, however this market will assist robust development throughout a broad vary of corporations.”
A September report from JLL discovered the world neocloud phase is projected to develop at an 82% annual fee from 2021 to 2025 as enterprises compete for GPU capability – highlighting hyperscale infrastructure can’t at the moment preserve tempo with AI demand, creating bottlenecks in availability.
The report famous that AI workloads, drawing over 100 kW per rack, require specialised cooling and ground loading capacities, which pushes knowledge facilities to improve.
Main neoclouds embrace CoreWeave, Crusoe, Lambda, Nebius, and OpenAI, with OpenAI’s consumer-facing providers and infrastructure investments – significantly its Stargate initiative – making it the most important within the class.
Fierce Competitors for Infrastructure
The Synergy report notes that whereas neoclouds face “fierce competitors,” their concentrate on AI workloads offers them an edge in high-growth segments. Many are repurposing infrastructure from crypto-mining or launching as startups specialised in high-performance computing, permitting them to scale shortly.
“They’re growing rather more power-dense services and likewise finding them effectively away from costly metro markets with energy constraints,” Dinsdale defined. “In fact, hyperscale operators are doing a few of that too.”
An extended tail of rising gamers – together with Utilized Digital, Northern Knowledge Group, Collectively AI, and WhiteFiber – can also be coming into the market.
With conventional hyperscalers increasing their very own GPU and AI choices, the road between cloud and neocloud is starting to blur.
Nevertheless, Dinsdale stated the neocloud sector’s slender specialization and aggressive development trajectory counsel it’s going to proceed to achieve share within the fastest-expanding areas of digital infrastructure – significantly these driving the following wave of AI growth.
Because the market develops, he says all the standard elements are at play: availability and price of actual property and energy, networking infrastructure, ease of doing enterprise, native monetary incentives, political stability, availability of native expert workforce, and proximity to clients.
“Given the character of many AI workloads, proximity to clients and latency points could typically be much less essential than they as soon as have been,” Dinsdale stated.
