Global AI Tally: McKinsey Estimates $7 Trillion in Data Center Investments by 2030
The digital age has formally arrived, but it is AI that is turning it into a race for computing power. According to new research by McKinsey & Company, the world will need to invest up to $7 trillion in building and operating data centers by 2030 just to keep pace with the growth of artificial intelligence.
The report, titled “The Cost of Compute: A $7 Trillion Race to Scale Data Centers,” warns that computing infrastructure must triple in the next 5 years, with up to 70% of that growth driven by AI workloads. These aren’t just more servers — it’s an architectural overhaul designed for scaling LLMs, generative models, and continuous real-time training.
Key projections:
– $5.2 trillion needed specifically for AI-related workloads
– Total power demand could reach 156 GW
– 125 GW of that must be added between 2025 and 2030
To illustrate the scale: the required investment is equivalent to laying nearly 5 million kilometers of fiber-optic cable — or wrapping the Earth 120 times around the equator.
McKinsey emphasizes that infrastructure efficiency is becoming a core factor in regional competitiveness. It’s not just about power, but energy usage. Advances in chips, LLM optimization, and efficient training may help mitigate strain on the grid — but the relentless growth in models and data demand offsets these gains.
Additional pressure comes from global chip shortages, regulatory uncertainty, and geopolitical fragmentation. Still, companies investing now in scalable, modular, and energy-efficient data centers will have a major edge in the next AI cycle.
McKinsey leaves no doubt: this race is no longer about algorithms. It’s about the infrastructure that powers them.
