Cray: Enterprises Go Hybrid for AI Workloads
More than 70% of survey respondents run AI applications on-premises and in cloud
Most enterprises that are running artificial intelligence (AI)-based applications are using a combination of on-premises IT infrastructures and the public cloud, with almost 40% using supercomputers, high-performance computing (HPC) systems, or dedicated AI hardware, according to a survey by supercomputer maker Cray.
The Lowdown: The Cray survey, which was released this week, shows that organizations are bullish on applications that leverage AI and see hybrid cloud environments as the way to run them.
The Details: Cray surveyed 300 IT professionals across a range of industries to gauge the adoption of IT in the enterprise and the infrastructure the organizations are using for these workloads. What the system OEM found was a mix of on-premises and cloud infrastructure being used. The survey found that 39% of respondents are using high-performance systems in their own data centers and that 65% said they needed to expand their on-premises systems to keep pace with growing performance and storage needs. In addition, 35% said they plan to move some cloud workloads back on premises.
In addition, the survey found that more than 70% of enterprises using on-premises systems also have implemented AI in the public cloud and that more than half of all public cloud users also use on-premises systems for AI workloads.
Whatever infrastructures they’re using, enterprises view AI as an important part of their IT efforts. According to the survey, 25% of respondents already consider AI to be critical and 77% said it will become critical in the next three years. Almost 70% said AI will improve operational efficiency, and more than half said the technology will enhance the customer experience. In addition, 44% said it will help to control costs and grow revenue.
The Impact: The numbers from the survey look good for both public cloud providers and system makers like Cray. Enterprises are adopting hybrid cloud strategies at a rapid pace, with a report by Flexera earlier this year finding that 69% of organizations are embracing such plans. On-premises systems that run AI workloads need the performance, latency, throughput, and storage capabilities that such applications demand.
Background: Cray has been making HPC and supercomputers for more than four decades. Hewlett-Packard Enterprise announced in May that it was buying the OEM for $1.3 billion, expanding its own capabilities in the fast-growing HPC market, which Hyperion Research said will hit $44 billion by 2023.
The Buzz: “The research highlights that supercomputing plays an important role in enabling mainstream AI adoption in the enterprise,” said Frederick Kohout, senior vice president and chief marketing officer at Cray. “IT professionals see the need to expand their use of on-premises infrastructure, like supercomputers and HPC clusters, to meet the expected growth of business-critical AI applications.”
“This survey by Cray is consistent with our own AI research studies,” said Addison Snell, CEO of Intersect360 Research. “Training for machine learning requires high-performance computing, either on premises or in the cloud, and there’s a significant overlap between AI investment and systems for HPC and hyperscale. One way or another, those interested in AI need to access a high-performance infrastructure.”
Related Links:
CHANNELNOMICS:
> HPE Makes Super Play for Cray
> AI, Blockchain Making Way Onto Mainframes: Ensono Report