This Intersect360 Research report presents the 2021 total market size and five-year forecast for the overall market for infrastructure for High Performance Computing (HPC) and artificial intelligence (AI) training, segmented into vertical markets across industry, government, and academia. The forecast horizon is from 2022 through 2026, with compound annual growth rates (CAGRs) using 2021 as a base.
This report creates a unified model for HPC and AI training infrastructure, including a description of the full influence of hyperscale companies as both buyers of infrastructure for AI training and of providers of cloud computing resources for both HPC and AI. For year-over-year consistency, we also examine traditional HPC budgets; these often include machine learning or AI training workloads. The unified model brings in additional spending on AI training that is not associated with any HPC budget.
Intersect360 Research defines HPC as the use of servers, clusters, and supercomputers—plus associated software, tools, components, storage, and services—for scientific, engineering, or analytical tasks that are particularly intensive in computation, memory usage, or data management. Additionally, this report tracks similar infrastructure for machine learning and deep learning for the training of AI models.
Intersect360 Research reports available in this series include the following segmentations:
The pandemic economy has continued to create tremendous short-term disruptions. As COVID-19 continued to affect multiple industries, the previously forecasted rebound was blunted, and the market experienced only slight growth. Total HPC market revenue grew to $41.0 billion ($40,961 million), up 5.2% from 2020. Within HPC, cloud computing continued extraordinary growth, buoyed by the uncertainty and supply chain disruptions brought about by the pandemic.
Including both HPC and AI training, the total market was $46.4 billion ($46,359 million), not including AI infrastructure purchased by hyperscale companies. Of the additional $5.4 billion in pure AI training spending not associated with HPC budgets, only about one-quarter is on-premises infrastructure; the other three-quarters is cloud computing supporting AI training. When hyperscale companies’ own use of AI is considered, it is apparent how dominant the role of hyperscale companies is.
Hyperscale companies are those with web-scale infrastructure supporting a wide variety of applications, such as search, social media, gaming, media distribution, and cloud services, to name a few. The hyperscale market—first described by Intersect360 Research as “ultrascale internet” in 2007 and broken out as a distinct market beginning in 2014—has grown to encompass nearly half of all data center infrastructure spending, with over 80% consumed by the twelve top spenders.
Over the next five years, HPC and AI training will rebound to stable growth levels. HPC budgets, conforming to previous years’ methodology, will grow to $59.2 billion ($59,241 million) in 2026, a 7.7% compound annual growth rate (CAGR). Not including hyperscale consumption, HPC and AI training combined will grow to $70.0 billion ($70,014 million), an 8.6% CAGR. The hyperscale market has been undergoing continuous growth and consolidation but is also subject to substantial risks; these are discussed in this report, and Intersect360 Research is not forecasting the hyperscale market at this time.
TABLE OF CONTENTS
EXECUTIVE SUMMARY 2
TABLE OF CONTENTS 4
Multiple Market Views of HPC and AI Training 8
What Are HPC, AI, and Hyperscale? 9
HPC AND AI TRAINING MARKET 2021: VERTICAL MARKETS 12
2021 Market Size: Traditional HPC View 12
Table 1: Total HPC Market Revenue ($M), 2021 vs. 2020, by Vertical Market 13
Figure 1: Total HPC Market Revenue ($M), 2021 vs. 2020, by Vertical Market 13
2021 Market Size: Expanded HPC and AI Training View 14
Table 2: Total HPC and AI Training Market Revenue ($M), 2021, by Vertical market 15
Figure 2: Total HPC and AI Training Market Revenue ($M), 2021, by Vertical market 15
Figure 3: Total HPC and AI Training Market Revenue ($M), 2021 by Vertical market, if Hyperscale Were Included 16
2022–2026 HPC AND AI TRAINING FORECAST: VERTICAL MARKETS 18
2022-26 Market Forecast: Traditional HPC View 18
Table 3: Total HPC Market Revenue ($M), 2020-21 Actuals, 2022 to 2026 Forecast, by Vertical Market 19
Table 4: HPC Market Revenue Share, 2020-21 Actuals, 2022 to 2026 Forecast, by Vertical Market 19
Table 5: HPC Market Year-over-Year Revenue Change, 2021 Actual, 2022 to 2026 Forecast, by Vertical market 20
Figure 4: Total HPC Market Revenue ($M), 2020-21 Actuals, 2022 to 2026 Forecast, by Vertical Market 21
2022-26 Market Forecast: Expanded HPC and AI Training View 22
Table 6: HPC and AI Training Market Revenue ($M), 2021 Actuals, 2022 to 2026 Forecast, by Vertical market 22
Table 7: HPC and AI Training Market Revenue Share, 2021 Actuals, 2022 to 2026 Forecast, by Vertical Market 23
Table 8: HPC and AI Training Market Year-over-Year Revenue Change, 2022 to 2026 Forecast, by Vertical Market 23
Figure 5: Total HPC and AI Training Market Revenue ($M), 2021 Actuals, 2022 to 2026 Forecast, by Vertical Market 24
APPENDIX A: METHODOLOGY 27
APPENDIX B: HPC, AI, AND HYPERSCALE: DEFINITIONS AND INTERDYNAMICS 29
What Are Artificial Intelligence and Machine Learning? 29
What Is HPC? 31
What Is Hyperscale? 31
What Makes Hyperscale “Hyper”? 32
Hyperscale and Business Computing 33
Hyperscale and HPC 33
How Hyperscale is Unique 35
APPENDIX C: ECONOMIC DYNAMICS OF THE COVID-19 PANDEMIC 37
APPENDIX D: HPC MARKET DYNAMICS MODEL AND FUNDAMENTAL FORECAST ASSUMPTIONS 39
Market Maturity 39
Fundamental Market Dynamics Model 40
Figure D1: Traditional HPC Market Dynamics 41
Fundamental Market Assumptions 42
Fundamental Drivers 42
Fundamental Market Dampeners 43
Model-Based Assumptions 44
Basic Market Drivers 44
Basic Market Dampeners 45
Single-User Access ($490): Report is for the sole use of one person, such as for forecasting or product planning. Redistribution or sharing of the reports is prohibited, even internally within the purchasing organization or department. Citation of quotes or data, such as on a web site, in a press release, or for enterprise use, requires approval from Intersect360 Research.
Enterprise-Wide Usage ($2,400): Report can be shared within the purchasing organization—for example, across product, strategy, sales, or forecasting teams. This does not extend to partner organizations, such as suppliers or resellers, but does include contractors working for the purchasing organization. Public citation of quotes or data, such as on a web site or in a press release, requires approval from Intersect360 Research. Advisory service clients are granted enterprise-wide usage rights for all reports in the subscribed service (HPC Advisory Service or Hyperscale Advisory Service), but not reprint rights unless purchased separately.
Reprints ($4,800 per report): Full redistribution rights for some or all of a report, such as for PR purposes, including email, print, web, or other distribution. The reports may not be edited in any way except as approved by Intersect360 Research (such as to excerpt a section or to highlight particular text). An organization must have enterprise usage of a report—either through individual report purchase or advisory service access—to purchase reprints. Contact Intersect360 Research to purchase.