GPUs-as-a-Service (GPUaaS) are gaining traction because enterprises racing to build and deploy AI and HPC workloads are increasingly constrained by GPU availability, cost, and operational complexity. Such demand for GPU capacity has given rise to neoclouds, a new type of cloud service provider that offers AI-focused cloud infrastructure.
Unlike hyperscaler GPU instances, which are bundled into massive cloud ecosystems with broad but often rigid service catalogs, GPUaaS platforms are typically purpose-built for performance, low latency, and customer access.
Neocloud and other GPUaaS providers often offer more specialized configurations, faster provisioning, and customer support tuned to AI workloads rather than generalized cloud services. This differentiation enables organizations to bypass the hyperscaler queue and gain access to GPUs quickly, often at lower cost, while still maintaining the flexibility of cloud consumption models.
See also: What Are Neoclouds and Why Does AI Need Them?
A deeper look: What exactly are neoclouds?
Neocloud providers offer an alternative to the two traditional ways enterprises typically meet their compute demands. Instead of investing millions in on-premises GPU clusters, which require specialized expertise to procure, cool, and manage, organizations can tap into GPUaaS providers for flexible, on-demand access to the latest accelerators.
Additionally, the neocloud provider model addresses the supply crunch many enterprises encounter with hyperscalers, where wait times for premium GPUs can stretch from weeks to months.
Companies considered to be neocloud providers include CoreWeave, Crusoe, Lambda Labs, Nebius, Vast.ai, and others. They, and other GPUaaS providers, provide predictable pricing, easier scaling, and the ability to align consumption directly with project cycles, making it particularly appealing for AI research groups, startups, and enterprises seeking agility.
See also: Report Uncovers the Realities of AI Adoption
Use of neoclouds surges
A recent study from HostingAdvice, conducted last month among 350 U.S. directors and senior managers in data engineering, cloud architecture, platform engineering, data science, and finance, indicates a tidal shift in AI infrastructure. The study found that 90% of organizations have already adopted, or plan to adopt, neoclouds. According to the report, this signals a strategic pivot away from traditional hyperscaler dependency.
The dominant driver is GPU wait times, according to the study. Nearly a third (31%) of survey respondents identified long delays accessing GPUs from the likes of AWS, Azure, and Google Cloud as the top reason for migrating to neocloud services.
Further compounding frustration, many businesses reported waiting two to four weeks for the most advanced and newest GPUs, with another 20% experiencing delays exceeding three months. The study infers that hyperscalers appear to prioritize internal usage over external customer demand.
Cost savings from the switch to neoclouds are reported as substantial:
- 36% of organizations enjoy 10–24% cost savings.
- 33.5% realize 25–49% in savings.
- 15.6% report savings of 50% or more.
Performance improvements are also notable:
- 42% report 25–49% faster performance.
- 14.6% achieved 50%+ performance gains.
- Only 2% saw no difference, and zero respondents saw worse performance.
The bottom line: Enterprises using neocloud providers got better compute at lower cost.
Benefits of the neoclouds model
For enterprises seeking AI compute power without the exorbitant costs, neocloud providers can deliver several benefits, including:
Lower Cost per GPU Hour: Neoclouds often charge a fraction of what hyperscalers do for comparable GPU instances. They claim that their lean operating models and hardware reuse strategies translate to real savings.
Dedicated Access: In many cases, neocloud providers offer dedicated bare metal access to GPUs, reducing contention and ensuring predictable performance.
Rapid Availability: With flexible procurement and provisioning processes, neocloud providers claim they can often deliver capacity much faster than traditional vendors, helping teams iterate and deploy AI models without delay.
A final word
Neocloud providers are filling the need for fast and affordable access to the GPUs that enterprises need to support their AI efforts. Additionally, such providers deliver measurable performance improvements and cost efficiencies, positioning their services as an alternative to mainstream AI infrastructure offerings.
Furthermore, neocloud providers are likely to compete in the near future with hyperscalers in the GPUs-as-a-Service marketplace. That market was valued at $3.23 billion in 2023 and is projected to grow to $49.84 billion by 2032, representing a 36% growth rate, according to Fortune Business Insights. (That estimate includes both hyperscalers and neoclouds.)
Salvatore Salamone is a physicist by training who has been writing about science and information technology for more than 30 years. During that time, he has been a senior or executive editor at many industry-leading publications including High Technology, Network World, Byte Magazine, Data Communications, LAN Times, InternetWeek, Bio-IT World, and Lightwave, The Journal of Fiber Optics. He also is the author of three business technology books.
