Gartner: AI Infrastructure Spending to Surge 53% on AI Demand, Account for 59% of Total AI Spending
TECHWORLD ·
✦ AI Summary
Gartner forecast that global AI spending will total USD 4 trillion in 2027, or about KRW 5,320 trillion, up 47.5% from the previous year.
Of that, global AI infrastructure spending in 2027 is expected to reach USD 2.4 trillion, or about KRW 3,192 trillion, accounting for 59% of total AI spending.
Gartner forecast that AI-optimized server spending will reach USD 957.2 billion in 2027, or about KRW 1,273 trillion, while total spending in the generative AI model market will rise 82% from the previous year to USD 51 billion, or about KRW 68 trillion.
BizCrushWant to follow Korean talks and live streams in real time?Try CapCrush on the web
Gartner forecast that global AI spending will total USD 4 trillion in 2027, or about KRW 5,320 trillion, up 47.5% from the previous year. Of that, global AI infrastructure spending in 2027 is expected to reach USD 2.4 trillion, or about KRW 3,192 trillion, accounting for 59% of total AI spending.
Gartner forecast that global AI infrastructure spending in 2027 will increase 53% from the previous year. It also presented the scale and share of global AI spending and AI infrastructure spending in 2027.
John-David Lovelock, a VP analyst at Gartner, said demand for scaling inference workloads is rising faster than expected, while demand for computing resources for training workloads is also expanding. He said this is driving continued growth in global AI infrastructure spending.
Gartner said AI-optimized server spending takes the largest share among the detailed categories of AI infrastructure spending. It projected AI-optimized server spending at USD 527.6 billion in 2026, or about KRW 702 trillion, and at USD 957.2 billion in 2027, or about KRW 1,273 trillion. Gartner analyzed that AI-optimized server spending will grow 81.4% in 2027.
Key factors behind the growth of the AI market include the continued expansion of AI infrastructure by hyperscalers and neo cloud providers, as well as higher prices for AI-optimized servers driven by "memflation."
Lovelock said service providers are supporting strategy development, data preparation, integration, and continuous deployment, and that this support is expanding companies' use of AI. He also said corporate spending is rising as existing software adds AI features, while investment in new platforms for enterprise-wide AI management and governance is also increasing.
Gartner forecast that global AI services spending will rise 29.5% from the previous year in 2027, while global AI software spending will increase 42% from the previous year. It also said that as the generative AI model market continues to grow, total spending in the generative AI model market in 2027 will increase 82% from the previous year to USD 51 billion, or about KRW 68 trillion.
Gartner said companies' use of AI is expanding into a wide range of business functions. As a result, demand for agentic AI that supports complex workflows, demand for multimodal applications, and demand for industry-specific AI were cited as factors supporting market growth.
Assuming accelerated adoption of domain-specific and specialized AI models, Lovelock forecast related spending will rise 99% in 2027. He said spending in that area will increase sharply as adoption of domain-specific and specialized AI models accelerates.
Companies are expanding their adoption of domain-specific models. The goal is to improve AI accuracy for specific use cases, enhance contextual understanding, and improve business outcomes.
Foundation models, however, continue to lead the market. Foundation models account for 81% of total generative AI model spending, and Lovelock said they are the foundation for AI agents, industry-specific models, and enterprise AI applications.
Source: TECHWORLD · Lee Gwang-jae
Original: https://www.epnc.co.kr/news/articleView.html?idxno=410020
References
This article was produced with the help of an automated content generation algorithm.
Source: TECHWORLD
View originalThis article was summarized and organized by BizCrush based on the original article from TECHWORLD. For exact quotations and full details, please refer to the original article.
