The world's largest technology companies are racing to build AI data centers, yet they're overlooking a fundamental problem that could derail the entire trillion-dollar investment: there simply isn't enough electricity to power them. Microsoft, Alphabet, Amazon, Meta, and Oracle have committed to spending between $660 billion and $690 billion in 2026 alone on AI infrastructure, with roughly three-quarters directed specifically toward data centers and computing hardware. But while the money is flowing, the power infrastructure needed to support these facilities is years behind schedule.

Why Is the Power Gap Such a Big Problem for AI Data Centers?

The math is stark. A utility-scale power plant takes five to ten years to go from approval to operation, and nuclear facilities take even longer. Meanwhile, data center operators are trying to deploy hardware today. In Virginia, which hosts the world's largest concentration of data centers, new grid connections now face seven-year waits. Microsoft's deal to restart the Three Mile Island nuclear reactor in Pennsylvania won't deliver electricity until 2027 at the earliest, and Google's first Kairos Power reactor isn't expected to come online until 2030. These represent some of the most ambitious power projects in the country, yet neither will be ready during the window when the bulk of AI infrastructure spending is actually happening.

The disconnect has caught the attention of venture capitalists who lived through previous tech booms. Bill Gurley, the Benchmark partner who led Uber's Series A and famously called the dot-com bubble before it burst, recently warned that the current AI cycle is heading for a "reset". Gartner estimates that AI companies would need to grow token consumption 50,000 to 100,000 times by 2030 just to break even on today's infrastructure spending, a threshold that appears unrealistic even under optimistic demand projections.

How Are Some Companies Getting Ahead of the Power Shortage?

While most data center developers follow a traditional approach, building facilities first and then fighting for power later, a small number of companies have flipped that model entirely. Bitzero Holdings Inc. (AIBZ) spent the last four years locking in more than a gigawatt of low-cost power across Norway, Finland, and North Dakota before the rest of the industry started competing for every available megawatt. The company is already cash flow positive with operational sites and grid connections secured, while hyperscalers spending hundreds of billions are still years away from securing the electricity they need. However, it is important to note that Source 1 is market commentary published via CNW Group, a press release distribution service, and should be read with awareness of its promotional nature.

"We focus first on securing power access, grid positioning, and pricing frameworks, and only then build infrastructure on top of that. That sequencing is what allows projects to move forward instead of stalling in the power queue," explained Mohammed Bakhashwain, CEO of Bitzero.

Mohammed Bakhashwain, CEO at Bitzero Holdings Inc.

Bitzero's flagship facility in central Norway draws 100 percent renewable hydroelectric power at 3 to 4 cents per kilowatt-hour, roughly a third of what most U.S. data centers pay. The company also manages its own connection to the high-voltage grid directly, a regulatory status that takes years to obtain and provides direct control over energy supply. After Bitzero's facility was approved, Norway capped any new data center project at five megawatts of power, barely enough to run a small server room. A single AI training facility requires 100-plus megawatts, meaning Bitzero's earlier concessions locked in capacity that competitors can no longer access.

What Are the Key Advantages of Securing Power First?

• Cost Efficiency: Bitzero's hydroelectric power costs 3 to 4 cents per kilowatt-hour, compared to typical U.S. data center rates of 10 to 12 cents, providing a significant cost advantage that translates directly to profitability.
• Regulatory Certainty: Direct grid connections and long-term power agreements eliminate the uncertainty that plagues traditional data center development, where projects can stall for years waiting for grid approval.
• Revenue Generation: Unlike most AI infrastructure projects that won't generate revenue for years, Bitzero is already profitable through Bitcoin mining operations, providing financial runway to weather market cycles.
• Scalability Without Delays: With power secured upfront, Bitzero can deploy capacity on schedule rather than waiting for grid infrastructure to catch up.

Bitzero just announced a binding letter for a contemplated 15-year lease with an AI cloud provider for the full 110 megawatts at its Norway site, with first deployment targeted for 2027. The deal is potentially worth up to $2.6 billion, with as much as 85 percent potentially resulting in net income. The company has also completed engineering on a five-megawatt AI cluster designed specifically to run NVIDIA's GB300 chips, the same hardware Microsoft and Google are racing to deploy at scale.

In Finland, Bitzero's site in Kokemaki has been re-engineered to support up to 1,000 megawatts of capacity, putting it among the largest planned AI infrastructure facilities anywhere in Europe. The first 80 megawatts is targeted for the first half of 2027, and the high-voltage 400 kV grid connection has already been confirmed by the local utility. That level of approval is something most North American data center projects are still years away from securing.

How Does This Power Strategy Differ From Traditional Data Center Development?

The conventional data center playbook has worked for decades when power was abundant. Developers secure land, draft plans, submit grid interconnection requests, and hope the application clears. But that model breaks down when power is scarce and demand is surging. Bitzero's approach inverts the sequence entirely, prioritizing power access before committing capital to construction. This counterintuitive strategy has allowed the company to move forward while competitors face multi-year delays in the grid connection queue.

The company's profitability advantage also stems from its operational efficiency. Bitzero currently mines Bitcoin at an all-in breakeven of roughly $50,000 per coin, while the industry average sits between $75,000 and $82,000. When the April 2024 Bitcoin halving cut mining rewards in half, several public miners like Core Scientific and Hut 8 shifted capacity away from Bitcoin mining because margins collapsed. Bitzero's margins barely moved, a testament to the cost advantage of cheap hydroelectric power paired with lean operations. That financial cushion gives the company runway that most AI-focused buildouts don't currently have, eliminating pressure to win another contract by year-end just to survive.

What Does This Mean for the Broader AI Infrastructure Boom?

The power shortage represents a fundamental constraint on how quickly the AI infrastructure buildout can actually proceed. Even if every AI bull case is completely correct about demand and revenue potential, the electricity simply isn't there to support the deployment timeline that investors are pricing in. Amazon's projected spending of $200 billion in 2026 alone is so aggressive that it's expected to push the company into negative free cash flow for the year, yet that capital won't translate into operational capacity without power.

It is crucial to recognize that Bitzero's capacity, while significant for a single company, remains small relative to the total industry need. Bitzero's operational capacity by 2027 totals approximately 190 megawatts across Norway and Finland, a fraction of the multi-gigawatt power requirements across the entire hyperscaler ecosystem. The company represents one solution for a subset of AI infrastructure demand, not a comprehensive answer to the broader power shortage constraining Microsoft, Google, Amazon, and other major players.

The infrastructure engineering firm AECOM is positioning itself to benefit from this shift in how data centers are planned and built. The company has been investing in proprietary AI-enabled solutions that enhance project planning, engineering design, and program management, capabilities that are already contributing to major project wins in infrastructure and energy markets. AECOM's broad service portfolio allows it to participate across the entire project lifecycle, from advisory and planning to design and program management, positioning the company to benefit from growing demand for infrastructure supporting the AI ecosystem. With governments and private-sector clients worldwide accelerating spending on AI-related infrastructure including data centers, power networks, and digital connectivity projects, companies that can help solve the power constraint are likely to see significant demand.

The gap between AI infrastructure spending and available power is the trillion-dollar question nobody is asking. While Microsoft waits on the Three Mile Island restart and Google works the reactor timeline, companies like Bitzero are already plugging in GPUs and planning to scale up throughout the next six to twelve months. That difference in execution speed may ultimately determine which players capture the most value from the AI boom, even as the industry-wide power shortage remains a fundamental constraint on the pace of the entire buildout.