The artificial intelligence infrastructure boom has flipped the venture capital playbook: access to cheap, reliable electricity is now more valuable than funding itself. Companies that invested in securing large amounts of low-cost power before the AI rush began are positioned to capture billions in recurring revenue, while competitors with plenty of capital struggle to find grid capacity. Grid connection wait times in major markets already stretch beyond four years, making "speed to power" the single most important factor driving data center expansion.

Why Has Power Access Become More Valuable Than Money?

The AI data center boom is expected to require nearly $7 trillion in global infrastructure spending by 2030, with roughly $5.2 trillion tied directly to AI workloads alone. But securing that capital is no longer the hardest part. Instead, companies are running into transformer shortages, interconnection delays, grid congestion, and multi-year energization timelines. This has completely reversed traditional venture capital priorities.

Investors are now aggressively rerating companies based on their ability to deliver energized capacity and contracted compute infrastructure, rather than traditional metrics like mining hash rates or revenue growth. Bitzero Holdings exemplifies this shift. The company started as a Bitcoin miner, using cash flow from cryptocurrency mining to secure large amounts of low-cost electrical power across Norway, Finland, and the United States. Now, that early investment in power infrastructure is paying off dramatically.

On May 5, 2026, Bitzero signed a binding agreement with OneQode Networks to lease its full 110-megawatt data center capacity in Namsskogan, Norway for AI workloads under a 15-year contract valued at roughly $2.6 billion. This single deal demonstrates how valuable pre-existing power infrastructure has become in the AI era.

How Are Companies Structuring These Power-for-AI Deals?

The economics of these arrangements reveal why power access has become so valuable. Under Bitzero's agreement with OneQode, the structure differs from traditional data center leases. Bitzero generates revenue by leasing the site's power capacity and infrastructure to OneQode, while OneQode pays the electricity bill for running the AI systems. This means Bitzero captures recurring infrastructure revenue without directly absorbing the massive ongoing power costs associated with operating large-scale AI workloads.

The Namsskogan agreement is structured at roughly $135 per kilowatt per month with a 3 percent annual escalator. At full utilization, the 110-megawatt site could generate roughly $176 million to $178 million in annual revenue, with potential annual net operating income of roughly $151 million based on an 85 percent margin profile.

"If I want exposure to crypto, I only need three positions now. I own Bitzero because they mine Bitcoin and they're actually a power company," said Kevin O'Leary.

Kevin O'Leary, Shark Tank investor and strategic investor in Bitzero Holdings

O'Leary's investment reflects a broader market recognition that power infrastructure companies are fundamentally different from typical cryptocurrency enterprises. He became one of Bitzero's earliest and biggest backers because he sees the company as rooted in energy infrastructure, not speculation.

Which Regions Are Winning the Data Center Power Race?

Geography is playing a decisive role in determining which companies and regions will capture the most value from the AI infrastructure boom. The Nordic region, particularly Norway and Finland, has emerged as a gravity center for digital infrastructure because of its combination of cheap power and clean energy sources.

Norway and Finland benefit from immense hydroelectric and nuclear baseloads that provide the kind of stable, long-duration electricity that AI workloads demand. Power prices across parts of the Nordic region are significantly below many major European markets. Additionally, cold climates drastically reduce cooling costs for data centers, which is critical since keeping servers cool consumes enormous amounts of water and energy.

Bitzero's expansion strategy reflects this geographic advantage. The company's flagship Namsskogan operation in Norway already supports Bitcoin mining and is now becoming the foundation for its AI infrastructure business. Bitzero also controls additional Norwegian expansion capacity tied to a broader development pipeline that management says could eventually scale well beyond 300 megawatts as grid upgrades continue. In Finland, Bitzero has secured a massive one-gigawatt development campus in Kokemäki tied directly into low-cost Nordic power infrastructure.

Texas is also emerging as a major hub for data center development, though for different reasons. The state has at least 248 data center projects planned across its territory, according to analysis by the Texas Tribune. Texas attracts developers because of cheap land, available power, fiber lines, and importantly, its lack of local business restrictions. However, nearly half of the planned data centers in Texas are set to be built in unincorporated areas not governed by cities or towns, marking a shift from existing data centers where only 12 percent are currently in unincorporated areas.

What Are the Key Factors Driving Data Center Location Decisions?

• Power Availability and Cost: Access to cheap, reliable electricity is now the primary constraint on data center expansion, making regions with abundant hydroelectric, nuclear, or natural gas capacity the most attractive locations for development.
• Regulatory Environment: States and regions with business-friendly regulatory frameworks and minimal zoning restrictions allow developers to build faster and more cost-effectively than areas with strict local oversight.
• Cooling Infrastructure: Cold climates reduce the water and energy required to cool servers, making Nordic regions and other northern locations particularly valuable for large-scale AI data center operations.
• Grid Connection Speed: The ability to connect to existing electrical infrastructure quickly has become more important than securing capital, since grid connection wait times in major markets already exceed four years in many cases.

How Is the Data Center Boom Affecting Water and Energy Resources?

The rapid expansion of data centers is creating pressure on regional water and energy supplies. Data center servers generate heat, and most water used by data centers goes toward keeping their systems cool. The amount of water required varies depending on the cooling system used and the center's location, since hotter regions require more cooling.

Google plans to use closed-loop cooling systems in water-stressed regions of Texas, where the liquid carrying heat away from equipment gets reused over years by circulating inside the system rather than being consumed immediately. An initial fill for such a system per building can range from 1.5 to 2 million gallons, equivalent to the average use of more than 6,000 U.S. households.

"An initial fill for the system per building can range from 1.5 to 2 million gallons, equivalent to the average use of more than 6,000 U.S. households," explained Ben Townsend.

Ben Townsend, Head of Infrastructure and Sustainability at Google

Water experts warn that data centers will put more pressure on Texas's water supply. One estimate shows data centers could account for between 3 percent and 9 percent of Texas's total water use by 2040, up from less than 1 percent today, according to research from the University of Texas at Austin. By comparison, manufacturing accounts for about 7 percent of the state's water use.

Data centers also use water indirectly through the extensive energy they consume. Texas's energy production, especially from coal, nuclear, and natural gas plants, requires massive amounts of water. As AI computing chips get more powerful and hotter, the need for both electricity and cooling water will continue to increase.

What's Next for the AI Infrastructure Power Race?

The convergence of massive AI demand, limited grid capacity, and the value of pre-existing power infrastructure is creating a new class of infrastructure giants. Companies like Bitzero that made early investments in securing cheap power are now being rewarded with billion-dollar contracts. Meanwhile, newer entrants with plenty of capital are discovering that money alone cannot solve the power constraint.

This dynamic is likely to accelerate consolidation in the data center and power infrastructure space, as companies with existing power assets become acquisition targets for larger technology firms desperate to secure capacity. The next wave of AI infrastructure winners will be determined not by who has the most capital, but by who controls access to the electricity that powers artificial intelligence.