With the rapid development of artificial intelligence (AI) technology, investments in data center infrastructure are increasing globally. According to Bain & Company's "2025 Technology Report," the industry needs to generate $2 trillion in revenue annually by 2030 to meet future AI demands. However, Bain's analysis shows that there may be a funding gap of about $800 billion.
Image source note: The image is AI-generated, and the image licensing service provider is Midjourney
Recently, many AI-focused investment projects have been announced, such as OpenAI's plan to build five large data centers in the United States, while Microsoft claims to be constructing "the world's largest" data center in Wisconsin. Bain estimates that to meet AI needs, the U.S. will need an additional 100 gigawatts of power capacity, which means an annual investment of $50 billion for data center construction.
Despite these ambitious investment forecasts from Bain, many experts remain skeptical. For example, John Dinsdale, Chief Analyst at Synergy Research, believes that the $50 billion annual investment target for 2025 is overly optimistic. Although recent data shows that capital expenditures by hyperscale operators reached $12.7 billion in the second quarter of 2025, an increase of 72% year-on-year, there are still many uncertainties in distinguishing between AI and non-AI infrastructure investments.
At the same time, Sid Nag, Chief Research Officer at Tekonyx, estimates that the annual capital expenditure on AI data center infrastructure globally is around $30 billion, but this number could decline if demand decreases. Recent research also shows that 95% of AI projects fail to generate returns, which may lead to continued reductions in capital expenditures in the coming years.
Bain also points out that challenges in building data centers include limitations in power supply, construction services, computing equipment (such as GPUs), and ancillary equipment (such as electrical switches and cooling systems), which may prevent infrastructure development from keeping up with growing demand. Especially, increasing power supply is considered the most challenging task.
Although the industry is full of expectations for the future of AI, the gap between the huge investment needs and the current financial situation has raised concerns among professionals. Sam Altman, CEO of OpenAI, said that if AI can maintain its current trajectory, it could achieve remarkable breakthroughs.
Key Points:
🌟 A $2 trillion annual revenue is needed to meet AI infrastructure investment, but a shortfall of $800 billion is expected.
⚡ An annual investment of $50 billion is required for data center construction, but experts believe this target is overly optimistic.
🔌 Building data centers faces challenges in power supply, construction services, and other areas.
