The race to build artificial intelligence infrastructure is heating up globally, but a critical warning has been issued for investors: backing the wrong data centre operators could lead to massive financial losses. According to a business analyst, the ability to scale facilities rapidly, not just own them, will separate the winners from the losers in the coming AI revolution.
The High Stakes of Power and Capital in the AI Era
Nnadozie Odinaka, a strategic finance professional completing his MBA at the Georgia Institute of Technology, delivered a stark message in an interview. He pointed out that while tech giants are pouring over $320 billion into data centres in 2025 alone, a significant increase from $241 billion in 2024, many investors are focusing on the wrong metrics.
"Investors are fixated on current occupancy rates and power capacity," Odinaka stated. "The real, urgent question is whether an operator has the balance sheet strength to scale at the blistering speed that AI development demands." The numbers underscore his point. AI workloads are incredibly power-hungry, consuming four to six times more energy per server than traditional computing tasks.
Modern racks equipped with advanced GPUs can demand up to 500 kilowatts—enough electricity to power an entire neighbourhood. Furthermore, a single AI training cluster can require over 100 megawatts, which is equivalent to the power needs of a small city's grid.
Two Critical Risks Most Investors Underestimate
Drawing from his experience in financial risk advisory at leading consulting firms, Odinaka identified two major risks that are consistently downplayed: capital intensity and execution complexity.
He issued a clear warning: "If your chosen data centre operator cannot secure power substations, advanced cooling infrastructure, and necessary construction permits well in advance, they will miss the market window completely. By the time they are scrambling to find megawatts of power, you are already losing."
Odinaka highlighted specific red flags for investors to watch:
- Insufficient Power Procurement: The leading operators today secured their power contracts years ago. An operator just starting negotiations is already far behind.
- Outdated Cooling Technology: The intense density of AI servers requires liquid cooling systems, which most older operators lack. Retrofitting existing facilities can lead to cost overruns exceeding 50% of the original budget.
"Cost overruns of 50 to 100 percent are common in this space," Odinaka cautioned. "Your operator needs substantial financial breathing room to absorb these shocks."
The Blueprint for Winning in the AI Infrastructure Race
For investors used to traditional real estate fundamentals, Odinaka recommends a completely new framework. He advises looking forward, not backward. "Stop looking at current capacity utilisation," he said. "Instead, ask if they can deploy two to three times their current capacity within 18 months. Inquire about reserved power capacity for the next five years. Examine if their capital structure is robust enough to handle the unexpected."
According to his analysis, the operators poised to succeed share three crucial characteristics:
- Strong, strategic partnerships with utility providers.
- Proven capabilities in modular, rapid construction.
- Diversified and resilient funding sources.
"The winners are treating scalability as a complex financial engineering problem, not merely a construction challenge," Odinaka explained.
The potential reward is monumental. Economic research projects that AI could add $15 trillion to global GDP by 2030, but this growth is entirely dependent on proportional infrastructure scaling. For savvy investors, this represents a chance for generational wealth creation. For the unprepared, it risks creating stranded, worthless assets.
"In five years, it will be painfully clear which operators understood that data centres are the foundational bedrock of the AI economy," Odinaka concluded. "Foundations determine how high you can build—and how much value you can ultimately capture." His final message to the investment community is blunt: the AI boom and its infrastructure demand are real, but profitability hinges on identifying operators with the financial discipline, strategic foresight, and execution capability to scale intelligently.
