Data Center Energy Oversupply and Credit Risks
A contrarian analysis of the AI data center build-out reveals a potential energy oversupply by 2030. Utilities are committing to 110GW of capacity against a 95GW demand forecast, creating significant credit and ratepayer risks. Forward power curves and natural gas prices contradict the bullish narrative, suggesting a mispricing of demand.
The Contrarian Case for Energy Oversupply
The prevailing narrative suggests that the AI boom will drive an unprecedented surge in energy demand, benefiting utilities and energy infrastructure. However, a detailed analysis of supply and demand data reveals a starkly different picture. Utilities have committed to connecting approximately 110 gigawatts of data center capacity, while third-party forecasts project total demand of only 95 gigawatts by 2030. This significant gap suggests that the industry is overbuilding, creating a high risk of stranded assets and diminished returns for utility companies.
Market Signals vs. Narrative
Market data further undermines the bullish demand narrative. Forward power curves in key markets like Texas are flat or slightly downward sloping, indicating that traders do not expect a substantial increase in energy prices or demand. Similarly, natural gas forward curves are inverted, with prices expected to decline over the next decade. This is particularly notable given the expansion of LNG exports, which should theoretically tighten supply. The disconnect between the hype and the market pricing suggests that the AI energy demand narrative is overhyped.
Credit and Structural Risks
The financing structure of the data center build-out introduces additional risks. Private credit funds are increasingly lending to data center projects, often using off-balance-sheet vehicles to keep debt off the balance sheets of major tech companies. This practice obscures true leverage and complicates credit analysis. As competition for deals intensifies, covenant protections may weaken, increasing the risk of default, particularly for second-tier operators. Furthermore, the circular financing arrangements between chipmakers, cloud providers, and data center operators raise concerns about vendor financing and potential bubbles.
Ratepayer and Political Implications
The overbuild risk has significant implications for ratepayers. In many states, there are no robust mechanisms to protect residential customers from the costs of data center infrastructure. If demand does not materialize as expected, the burden of underutilized capacity may fall on consumers, leading to higher electricity bills and political backlash. Utilities in states with strong ratepayer protections, such as Indiana, are better positioned to manage this risk. Investors must carefully assess jurisdiction-specific risks when evaluating utility stocks.
Conclusion
While the AI revolution is real, the energy demand narrative is likely overstated. The combination of utility oversupply, flat forward curves, and structural credit risks suggests a more cautious approach is warranted. Investors should focus on companies with strong ratepayer protections and conservative financing structures, while avoiding those exposed to high-risk private credit and circular financing arrangements.
Key insights
-
Utilities have committed to 110GW of data center capacity, exceeding the 95GW demand forecast for 2030. This oversupply creates a high risk of stranded assets and reduced earnings growth.
Impact: Utility stocks may face downward pressure on earnings estimates as overbuilt capacity fails to generate expected returns.
-
Forward power and natural gas curves are flat or inverted, contradicting the bullish narrative of exponential AI energy demand. Traders are not pricing in the expected demand surge.
Impact: The disconnect between narrative and market pricing suggests a potential correction in energy-related assets if demand does not materialize.
-
Private credit funds are aggressively lending to data centers, using off-balance-sheet vehicles to obscure leverage. This practice increases credit risk and complicates analysis.
Impact: Investors in private credit and data center debt face higher default risk, particularly for second-tier operators with weaker covenant protections.
-
Ratepayer protection mechanisms vary significantly by state. In many jurisdictions, residential customers may bear the cost of underutilized data center infrastructure.
Impact: Utilities in states without strong ratepayer protections face higher political and regulatory risk, potentially leading to higher electricity bills and consumer backlash.
-
Existing grid capacity is likely sufficient to handle near-term AI demand, reducing the urgency for massive new build-outs. This supports the oversupply thesis.
Impact: The reduced need for new generation projects may limit growth opportunities for utility companies and energy infrastructure providers.
Action items
-
Analyze utility committed capacity against demand forecasts to identify oversupply risks. Focus on companies with high committed capacity relative to projected demand.
Impact: Identifying oversupply risks allows investors to avoid utilities with high stranded asset exposure and select those with more conservative growth strategies.
-
Monitor forward power and natural gas curves for signals of demand expectations. Use these curves to validate or challenge the AI energy demand narrative.
Impact: Forward curves provide real-time market signals that can help investors make more informed decisions about energy-related assets.
-
Assess the credit quality of private credit data center loans, focusing on covenant protections and the creditworthiness of borrowers. Avoid loans with weak covenants or high leverage.
Impact: Careful credit analysis helps investors avoid default risk in the private credit data center market, particularly for second-tier operators.
-
Evaluate ratepayer protection mechanisms in different states when investing in utilities. Prioritize companies in states with strong protections for residential customers.
Impact: Selecting utilities in states with strong ratepayer protections reduces political and regulatory risk, leading to more stable earnings and lower consumer backlash.
-
Scrutinize off-balance-sheet financing structures used by tech companies for data centers. Adjust credit analysis to account for hidden leverage.
Impact: Accurately assessing hidden leverage helps investors make more informed decisions about the creditworthiness of tech companies and their data center investments.
Quotes
“So the utilities are working on already connecting almost as much as you need by 2035.”
“The forward curve for gas is inverted. It goes from 370 to 360 by the end of the decade.”
“So that's what we're spending all our time on. And it's state by state, and even within the same state, you've got numerous jurisdictions.”