EXECUTIVE SUMMARY
This Intersect360 Research report presents an overview of the technologies and trends shaping the HPC-AI market in 2025, with a focus on Storage and Data Management. Other technology segments are covered in further reports in the “State of the HPC-AI Market” series.
This report is based on surveys of members of the HPC-AI Leadership Organization (HALO). HALO is a global end-user organization, facilitated by Intersect360 Research, that helps steer the course of the HPC-AI industry by identifying key issues, providing input into planned Intersect360 Research studies, and participating in surveys. For more information on HALO, visit www.hpcaileadership.org. HALO membership is free of charge to HPC and AI users.
The Intersect360 Research “State of the HPC-AI Market” report series includes the following technology modules:
• Storage and Data Management
• Systems and Clusters
• Facilities, Power, and Cooling
• Interconnects and Networking
• Quantum Computing
• Processing Elements (CPUs, GPUs, Accelerators)
• Cloud Computing (Including On-Prem as-a-Service Models)
Each report in the “State of the HPC-AI Market” series contains three sections: 1) end-user survey data on installations, major trends, and “satisfaction gap” analysis to reveal features where buyers are currently not satisfied with available technologies relative to their importance; 2) Intersect360 Research analysis of these data and trends; 3) submitted content from invited top vendors, as determined by end-user surveys, responding to a fixed template of questions on target market, technology differentiation, and future vision.
Executive Summary
The current zeal for AI solutions is transforming the IT landscape. Multiple Intersect360 Research studies have shown that strategies and budgets of HPC user organizations are increasingly driven by AI. This survey reveals that while AI initiatives are significant in driving top-level influence, including funding, traditional HPC requirements are paramount in guiding purchase decisions.
Ninety-three percent of survey respondents agreed that “storage solutions are strategically important, carefully considered acquisitions” for HPC-AI. A majority said that the storage infrastructure was usually shared for HPC and AI workloads, even though a majority also said that storage and data management requirements for HPC and AI workloads were different.
The emergence of AI workloads has affected the acquisitions of solutions for storage and data management solutions. Vendors of solutions should mind the areas of greatest “satisfaction gap”—i.e., the greatest relative gap between the rated importance of a feature and how satisfied users are with that feature. HPC-AI users are encouraged to remain mindful of their most pressing present needs while considering new solutions. For many organizations, data is the most important asset. The best high-performance solutions
will extract insights without sacrificing data sovereignty.
TABLE OF CONTENTS
Introduction and Executive Summary – 2
Introduction – 2
Executive Summary – 2
Table of Contents – 4
State of the HPC-AI Market: Storage and Data Management – 6
Demographics – 6
Survey Results – 6
Figure 1: HPC and AI Workloads: Same Storage, or Separate? – 7
Figure 2: Storage System Capacities for HPC, AI, and Mixed HPC-AI – 8
Figure 3: Storage Requirements Same/Different For HPC and AI – 8
Figure 4: Percent of HPC-AI Storage In Public Cloud – 9
Figure 5: Annual Growth Rate for HPC-AI Storage Capacity – 10
Figure 6: Percent of HPC-AI Storage that is File, Block, Object – 10
Figure 7: High-Performance Storage Trends: Agree/Disagree – 12
Satisfaction Gap Analysis – 13
Figure 8: Importance of HPC, AI, and Enterprise Workdloads for Next Storage Purchase for HPC-AI – 14
Figure 9: Type of Storage Planned for Next HPC-AI Storage Purchase – 14
Figure 10: Satisfaction Gaps for HPC-AI Storage and Data Management 16
Conclusions – 17
Appendix A: Survey Demographics – 19
Figure A1: Economic Sector of Respondents – 19
Figure A2: Geographic Region of Respondents – 20
Appendix B: Data Tables – 21
Table 1: HPC and AI Workloads: Same Storage, or Separate? – 21
Table 2: Storage System Capacities for HPC, AI, and Mixed HPC-AI – 21
Table 3: Storage Requirements Same/Different For HPC and AI – 21
Table 4: Percent of HPC-AI Storage In Public Cloud – 22
Table 5: Annual Growth Rate for HPC-AI Storage Capacity – 22
Table 6: Percent of HPC-AI Storage that is File, Block, Object – 23
Table 7: High-Performance Storage Trends: Agree/Disagree – 24
Table 8: Importance of HPC, AI, and Enterprise Workdloads for Next Storage Purchase for HPC-AI – 24
Table 9: Type of Storage Planned for Next HPC-AI Storage Purchase – 24
Table 10A: Importance Ratings of Features of HPC-AI Storage and Data Management – 25
Table 10B: Satisfaction Ratings of Features of HPC-AI Storage and Data Management – 25
Table 10C: Satisfaction Gaps for Features of HPC-AI Storage and Data Management – 26
Appendix C: Survey Text – 27
Appendix D: Submitted Content from Vendors – 32