Posts tagged Data Centers

Why Power Is the New Constraint in Data Center Strategy

The Electricity Reckoning

For decades, data center operators competed on connectivity, latency, and redundancy. Today, a new variable has moved to the top of the site-selection checklist: access to electricity. The explosive growth of AI workloads has transformed power availability from a background planning assumption into the primary strategic constraint facing the industry — and the fintech firms that depend on it.

The scale of projected demand is staggering. Goldman Sachs Research forecasts that global power consumption from data centers will grow by as much as 165% by 2030 compared to 2023 levels. More recent estimates suggest that figure could reach 220% as AI server shipments exceed earlier projections. Avid Solutions, citing Goldman Sachs Research, notes that data centers will require approximately $6.7 trillion in investment by 2030 to match growing compute power needs — the largest infrastructure investment cycle in modern history. In the United States alone, data centers are projected to account for 8% of total national power demand by the end of the decade, up from roughly 3% in 2022.

Grid Constraints and Interconnection Delays

The problem is not simply one of scale — it is one of timing and geography. AI data centers create unusually concentrated, around-the-clock electrical loads. Unlike residential or industrial demand, which fluctuates throughout the day, AI facilities run at maximum draw continuously. These concentrated 24/7 loads stress grid planning assumptions that were designed for a very different demand profile.

Deloitte's 2025 AI Infrastructure Survey of 120 US data center and power company executives found that grid stress was the leading challenge for data center infrastructure development, with 72% of respondents rating power and grid capacity constraints as very or extremely challenging. Perhaps most strikingly, there is currently a seven-year wait on some requests for grid interconnection — meaning that data centers in planning today may not be able to secure power access until well into the 2030s. Deloitte estimates that US AI data center power demand could grow more than thirtyfold between 2024 and 2035, reaching 123 gigawatts from just 4 gigawatts today.

Infrastructure planning is struggling to keep pace. Goldman Sachs Research estimates that approximately $720 billion in grid spending will be needed through 2030, and that US utilities will need to invest around $50 billion in new generation capacity just to support data center load growth. New pipeline capacity will also be required, as incremental data center power consumption is expected to drive around 3.3 billion cubic feet per day of additional natural gas demand by 2030.

Risk Management Implications for Fintech Leaders

For fintech executives, the power constraint is not just an infrastructure problem — it is a risk management challenge. Business continuity plans that assume reliable, scalable compute availability must now account for the possibility of power-driven capacity shortfalls. A firm whose cloud provider or colocation partner cannot expand its footprint due to grid interconnection delays faces real exposure to latency, throughput, and uptime risk at exactly the moment when AI systems are becoming operationally critical.

The constraint also reshapes competitive dynamics. Avid Solutions highlights that power constraints, not capital limitations, represent the main bottleneck for building new data centers. This means that well-capitalized technology firms with long-term power agreements and dedicated facilities will have structural advantages over firms that rely on spot capacity or shorter-term contracts. For fintech companies that have not yet locked in infrastructure partnerships, the window for securing favorable, long-term compute access may be narrowing.

Proactive fintech leaders are beginning to treat power access as a strategic input, alongside capital, talent, and regulatory positioning. Understanding where compute capacity is being built, how it is being powered, and what risks surround its availability is no longer the exclusive domain of the data center team — it is a board-level concern.

Citations

Avid Solutions. "13 Data Center Growth Projections That Will Shape 2026–2030." Avid Solutions, January 2026. https://avidsolutionsinc.com/13-data-center-growth-projections-that-will-shape-2026-2030/

Goldman Sachs. "How AI Is Transforming Data Centers and Ramping Up Power Demand." Goldman Sachs Insights, 2025. https://www.goldmansachs.com/insights/articles/how-ai-is-transforming-data-centers-and-ramping-up-power-demand

Goldman Sachs. "AI to Drive 165% Increase in Data Center Power Demand by 2030." Goldman Sachs Insights, February 2025. https://www.goldmansachs.com/insights/articles/ai-to-drive-165-increase-in-data-center-power-demand-by-2030

Deloitte. "Can US Infrastructure Keep Up with the AI Economy?" Deloitte Insights, June 24, 2025. https://www.deloitte.com/us/en/insights/industry/power-and-utilities/data-center-infrastructure-artificial-intelligence.html

Serverwala. "Data Centers and Fintech: Powering Financial Innovation." LinkedIn Pulse. https://www.linkedin.com/pulse/data-centers-fintech-powering-financial-innovation-serverwala-saisc

Peter CavicchiaJuly 22, 2026Data Centers, Electrical Grids

Sustainable Data Centers as a Competitive Advantage

The race to build next-generation data centers is no longer decided by compute density alone. Sustainability has emerged as a decisive selection criterion—one that touches operating costs, regulatory exposure, investor expectations, and long-term resilience. For enterprises and regulated financial firms evaluating colocation or build-to-suit options, a facility’s environmental profile is now as material as its uptime record.

The Scale of the Problem

Data centers are voracious consumers of resources. Beyond the enormous electricity demands required to power thousands of servers, facilities can consume between one and five million liters of water per day—a volume that strains communities already experiencing drought, particularly across the American Southwest. As machine learning workloads and large language model inference continue to scale, those demands will only intensify. Harvard SEAS researchers studying the challenge noted that advanced computing tasks “need more and more computing power,” driving the need for more data center capacity—and with it, a proportionate growth in carbon emissions.

Four Levers for Greener Operations

Research from Harvard’s John A. Paulson School of Engineering and Applied Sciences identifies four primary engineering levers that data center operators can pull to reduce environmental impact: alternative energy generation, cooling efficiency, waste-heat recycling, and improved water management.

Cooling is the most immediate target. Conventional air-circulation systems account for roughly 40 percent of a typical facility’s electricity consumption, much of it wasted on components that generate little heat. Newer direct-to-chip cooling—where coolant chills metal plates placed against the highest-heat components—can substantially cut that load. The shift toward liquid cooling is already underway, and facilities that adopt it early gain a durable power-usage efficiency (PUE) advantage over peers.

Waste-heat reuse turns a liability into an asset. Dublin provides a leading example: an Amazon facility is redirecting thermal exhaust to supply space heating and hot water for public buildings and housing. Operators that quantify and monetize their waste heat unlock an additional revenue stream while reducing net carbon output—an argument that resonates with both ESG-focused boards and local regulators.

Renewable energy procurement completes the picture. Interactive siting tools can overlay transmission infrastructure with solar, wind, and small modular nuclear reactor potential, helping operators identify locations where clean power is both available and affordable. Water stewardship—mapping local supply constraints before breaking ground—rounds out the discipline, reducing the risk of regulatory friction or community opposition.

From ESG Checkbox to Competitive Moat

Leading colocation providers are turning sustainability commitments into measurable business advantages. Digital Realty, for example, has issued $7.2 billion in green bonds and secured 1.5 gigawatts of contracted solar and wind capacity. Its US portfolio is 69 percent ENERGY STAR-certified by managed IT capacity, and it has certified more than 15 million square feet to LEED, BREEAM, and IGBC standards—with 61 percent achieving Gold or higher. The U.S. EPA has listed the company among the top ten largest buyers of renewable energy nationally. Those credentials are not merely marketing; they represent verifiable cost structures and risk profiles that enterprise procurement and compliance teams can audit.

For regulated financial firms, the stakes are even higher. Energy-efficient facilities with verifiable green certifications translate directly to lower Scope 2 emissions disclosures, stronger alignment with SEC climate-risk reporting requirements, and reduced exposure to future carbon pricing. Investors increasingly scrutinize infrastructure vendors through the same ESG lens they apply to portfolio companies, making a data center partner’s sustainability credentials a factor in due diligence.

Ultimately, sustainable data center design is not a concession to environmental sentiment—it is an engineering and financial discipline. Facilities that consume less energy, manage water intelligently, reuse waste heat, and source power renewably will carry lower operating costs, attract better financing, and face fewer regulatory headwinds than those that do not. In a market where differentiation is hard-won, sustainability is becoming one of the clearest signals of operational excellence.

Peter CavicchiaJune 17, 2026Data Centers, Big Data, Sustainable Data Center Design

The Blog

Bio