Table of Contents
Intersect360 Research has recently released its annual report on the Worldwide HPC-AI Market, providing insights into the growth and trends of the market in 2023 and forecasting for the years ahead. The report highlights the significant growth of the HPC-AI market, with a total worldwide market size of $85.7 billion in 2023, representing a 62.4% year-over-year increase. This growth can primarily be attributed to the substantial increase in spending by hyperscale companies on their AI infrastructure.
Growth in Hyperscale AI
In 2023, for the first time, hyperscale companies’ spending on AI infrastructure surpassed the spending on on-premises infrastructure for HPC-AI. This surge in hyperscale AI investment can be compared to a gold rush, as it defies normal market dynamics. The enthusiasm for AI and its promises has propelled the HPC-AI market to new heights, both in on-premises and hyperscale data centers. However, organizations will now need to evaluate the true business models for AI as they move forward.
Swings in the Hyperscale AI Market
According to Intersect360 Research, the hyperscale AI market segment is expected to experience some wild swings in the coming years. After an unprecedented period of investment, hyperscale spending on compute infrastructure will stabilize in the next couple of years. On the other hand, on-premises HPC-AI spending is projected to continue growing steadily, offsetting the slowdown in the hyperscale segment.
Forecast for the HPC-AI Market
The report forecasts that hyperscale companies’ spending on AI infrastructure peaked in 2023 and will decline in the following years. Over the next five years, the combined hyperscale and on-premises HPC-AI market is expected to remain relatively flat, with a compound annual growth rate (CAGR) of 0.3%. However, during this period, on-premises infrastructure for HPC-AI will continue to grow at a CAGR of 5.4%, while hyperscale HPC-AI infrastructure will experience a decline with a CAGR of -4.6%.
Return to Normalcy
Despite the projected dip in HPC-AI spending in 2024 and 2025, it still represents a significant increase compared to 2022. This decline indicates a return to normalcy from the rapid growth experienced in 2023. As steady growth resumes, it is expected that by 2028, total spending will surpass that of 2023. Moreover, the proportion of on-premises and hyperscale infrastructure will reverse, with more long-term investment in stable AI for enterprises shifting from the cloud to on-premises systems.
The AI Market Breakdown
Intersect360 Research has found that the AI market largely consists of markets that are already counted. Pure-AI infrastructure spending is predominantly confined to hyperscale and cloud environments. However, organizations with existing on-premises HPC expertise have expanded their environments to include AI. As the enterprise AI market evolves, there is expected to be a balance between cloud and on-premises infrastructure, similar to that seen in the HPC market.
Hyperscale Companies and the Market
Hyperscale companies are characterized by their internet-driven business models and substantial investment in IT infrastructure. Examples of hyperscale companies include Amazon, Apple, Google, Meta, and Microsoft. Additionally, non-U.S. companies such as Alibaba, ByteDance, Naver, and Tencent, as well as new AI-focused cloud services like CoreWeave, Denvr Dataworks, and Lambda, fall into the hyperscale segment.
Market Shifts and Surprises
According to Dan Olds, Chief Research Officer at Intersect360 Research, the HPC, AI, and hyperscale market segments have experienced several years of chaos, and this trend is likely to continue. The next five years are anticipated to bring significant market shifts, many of which will come as a surprise to many industry observers.
Conclusion
The Intersect360 Research report on the Worldwide HPC-AI Market in 2023 highlights the tremendous growth of the market, driven by hyperscale companies’ investment in AI infrastructure. While hyperscale AI spending is projected to decline in the coming years, on-premises HPC-AI infrastructure will continue to grow steadily. The report also emphasizes the need for organizations to evaluate the true business models for AI as the market evolves. As the enterprise AI market develops, a balance between cloud and on-premises infrastructure is expected to emerge, similar to the existing dynamics in the HPC market.
