Connor Teskey views the unprecedented global energy buildout for artificial intelligence as a multi-decade infrastructure play with enough market capacity to accommodate multiple competitors, according to an interview published by Semafor.
Capital Requirements and Energy Bottlenecks Through the Decade
Power availability remains the primary bottleneck for digital infrastructure. Connor Teskey stated that this supply-demand imbalance and its associated capital requirements will persist through the end of the decade and beyond. As all forms of energy generation face surging demand, different technologies are scaling at distinct paces based on deployment speed and generation type.
Fuel cells currently represent the fastest generation method to deploy. Meanwhile, onshore wind and solar remain the cheapest form of bulk electricity, while batteries and energy storage provide necessary stability for an increasingly volatile electric grid. Nuclear energy captures the large-scale baseload and energy security requirements demanded by large-scale operators.
Down-the-fairway, mature renewables such as onshore wind, solar, and battery storage attract the highest concentration of market participants. Consequently, return outcomes in those segments remain more range-bound. Conversely, nuclear infrastructure presents a massive opportunity defined by intense demand and steep capital requirements. Because fewer investors possess the specialized expertise to execute nuclear projects, return requirements for nuclear assets differ significantly from standard renewables.
The Evolution of AI Infrastructure Financing
The financing model for AI infrastructure follows a traditional historical trajectory. Independent capital providers initially build and hold the assets. As the asset class matures, commercial banks syndicate the financing, drawing in other institutional participants such as private credit funds and insurance companies. Teskey noted that this exact maturation process is occurring within the AI sector, albeit at a significantly larger scale and much faster velocity.
This rapid expansion brings major technology competitors to the same negotiating table with infrastructure investors. According to Teskey, the sheer magnitude of the global energy buildout and the rewiring of global supply chains for greater resiliency require capital far exceeding historical limits from governments or public markets. This broad capital deficit explains why five or six leading players routinely collaborate on single initiatives, operating under the premise that market demand is broad enough to accommodate multiple participants.
Evaluating Risk and Separating Hype from Physical Fundamentals
Market dispersion will ultimately separate top performers from underperforming ventures. Teskey’s firm applies a rigorous filter, executing only a tiny fraction of the total opportunities reviewed while still deploying substantial amounts of capital. Addressing the extreme demand for natural gas turbines among technology firms, Teskey pointed out supply chain realities: ordering a new gas turbine today means delivery will not occur within a decade beginning with the number two.
Public pushback and data center zoning friction do not threaten the core trajectory of energy demand growth. Teskey dismissed concerns that minor algorithmic adjustments in model efficiency could disrupt the fundamental supply-demand imbalance. Even if software efficiency models improve, current electricity and infrastructure supply levels remain drastically insufficient to service total projected load.
Current data center projects operate under strict 20- to 25-year contracts with major corporate counterparties. These agreements guarantee a return on invested capital during the initial contract lifecycle regardless of broader demand shifts decades into the future.
Building the Toll Roads of the Digital Era
Traditional infrastructure investments historically relied on public sector offtakers like municipal water departments or port authorities. Modern AI projects, by contrast, involve private corporate counterparties. Teskey characterizes this shift as a natural evolution where digital data requires transportation, processing, and storage infrastructure identical to physical commodities like toll roads, railroads, and pipelines.
