Generative AI (Gen AI) is rapidly evolving, with investments driving its development and deployment across various sectors. As of 2024, spending on AI infrastructure has surged, with projections indicating even more substantial investments in the coming years.
As per NewStreet Research, global spending on AI-capable processing units (XPUs) is expected to reach $105 billion in 2024, representing a dramatic increase compared to previous years. The total capital expenditure (capex) on AI infrastructure is projected to hit $210 billion in 2024, a tenfold increase from 2020.
The rapid growth in AI capex is primarily driven by the widespread adoption of Large Language Models (LLMs) such as ChatGPT, which demand significantly more computing power for both training and inference processes.
“Nvidia captured the lion’s share of the opportunity, as its GPUs were ideally positioned, with very competitive specs, a flexible architecture, and an ecosystem built over decades,” analysts said.
The investment in XPUs is only part of the broader AI landscape. The total AI capex also includes spending on servers, networking, power distribution, cooling equipment, and building infrastructure, all of which are crucial to supporting the expanding AI workload.
Key players and their investments
The largest investments in AI infrastructure are being made by the top four hyperscalers—Microsoft, Meta (NASDAQ:META), Google (NASDAQ:GOOGL), and Amazon (NASDAQ:AMZN). In 2024, these companies are expected to account for a significant portion of the $210 billion AI capex:
Microsoft (NASDAQ:MSFT): $40 billion (19% of total)
Meta: $23 billion (11% of total)
Google: $29 billion (14% of total)
Amazon: $16 billion (8% of total)
In addition to these giants, Tier 2 cloud providers and enterprises are also making substantial investments, each contributing $52 billion to the total AI capex.
As AI infrastructure spending continues to rise, the corresponding revenues are expected to follow. By 2027, NewStreet Research projects that global AI capex will reach $450 billion, with XPU spending alone accounting for $225 billion.
This expansion in infrastructure will translate into significant operational costs (COGS), which are estimated to reach $210 billion by 2027.
The primary sources of revenue in the AI sector will stem from cloud services, enterprise AI solutions, and AI-driven products. The hyperscalers will continue to dominate the cloud AI revenue streams, leveraging their massive investments in infrastructure.
For instance, Microsoft's Azure is expected to benefit from its AI capabilities, with AI contributing substantially to its growth.
Tier 2 cloud providers, such as CoreWeave, Lambda, and Oracle (NYSE:ORCL), are also expected to generate significant AI revenues, with projections indicating around $9 billion in 2024.
Similarly, enterprises investing heavily in AI infrastructure, such as Tesla (NASDAQ:TSLA), Apple (NASDAQ:AAPL), and Bytedance, are poised to capitalize on the growing demand for AI-powered solutions.