This Is What It Takes to Get a Data Center Financed
Timestamps are as accurate as they can be but may be slightly off. We encourage you to listen to the full context.
Podcast Summary
This episode dives deep into the complex world of data center financing, exploring how these infrastructure projects require intricate coordination between real estate, technology, power generation, and financial markets. (03:09) The hosts speak with Travis Wofford, a partner at Baker Botts law firm, who specializes in helping clients navigate the multifaceted legal and financial challenges of building data centers in the AI era.
- Main Theme: The episode examines how data centers represent a convergence of high-tech innovation, real estate investment, and massive power infrastructure requirements, creating unprecedented financing complexities in the age of AI.
Speakers
Joe Weisenthal
Co-host of Bloomberg's Odd Lots podcast, focusing on financial markets, economics, and complex investment themes. Known for his analytical approach to understanding market dynamics and emerging investment trends.
Tracy Alloway
Co-host of Bloomberg's Odd Lots podcast with expertise in financial markets and credit analysis. She brings deep knowledge of structured finance and alternative investment strategies to the show.
Travis Wofford
Partner and chair of the corporate department at Baker Botts law firm, based in Houston. He specializes in complex deal structuring, helping clients build, buy, finance, and sell major infrastructure projects and companies, with particular expertise in energy and data center financing.
Key Takeaways
Data Centers Require Sequential Financing Approaches
Travis explains that data center projects typically require three distinct financing phases: development capital (often equity-funded), construction capital through large construction loans once permits and power are secured, and takeout financing like securitizations once construction is complete. (24:28) This staged approach helps manage the enormous capital requirements and uncertain timelines inherent in these projects. The key insight is that different types of capital are appropriate at different risk stages of the project lifecycle.
Power Infrastructure Creates the Biggest Bottlenecks
The episode reveals that power interconnection, not chip availability or financing, represents the primary constraint for data center development. (22:35) Travis describes how getting both load interconnection (for drawing power) and generation interconnection (for producing power) can take up to five years, with timelines frequently extending beyond initial estimates. This creates significant planning challenges and affects project economics.
Location Strategy Balances Connectivity Against Saturation
Northern Virginia remains attractive due to subsea cable connectivity and regulatory environment, but market saturation is driving developers to seek alternatives like Texas and ERCOT. (20:59) Travis notes that while Virginia offers excellent connectivity to Europe, Africa, and other US locations, the saturated market makes it difficult to secure power, water, and other infrastructure quickly for new large-scale projects.
Securitization Structures Mirror Historical Infrastructure Finance
Despite the technological complexity, data center financing leverages proven securitization methods used for cell towers and residential solar projects. (08:36) The key differences lie in tenant quality assessment, lease duration analysis, and facility technology resilience rather than fundamental structural innovation. Rating agencies apply familiar frameworks while adjusting for data center-specific variables.
Private Credit Offers Speed and Customization Over Cost
Private credit provides non-dilutive capital with more flexible underwriting standards compared to traditional bank lending, typically targeting 15% IRR with 10-12% upfront costs. (31:54) While more expensive than public markets, private credit can move faster and accommodate riskier profiles that might not meet bulge bracket bank standards, making it attractive for developers needing rapid capital deployment.
Statistics & Facts
- Morgan Stanley forecasts $2.9 trillion in global data center spending through 2028, compared to just $950 billion in total S&P 500 CapEx spending in 2024. (04:16)
- Current generative AI revenue is approximately $16 billion, creating a significant gap compared to the projected infrastructure investment. (04:45)
- The interconnection queue contained 2,600 gigawatts two and a half years ago before AI became prominent, with expectations that only 20% would actually be constructed. (45:55)
