As artificial intelligence drives an unprecedented surge in electricity demand, a once-niche asset class is suddenly at the center of global competition: power plants. Data centers, hyperscalers, private equity firms, and infrastructure funds are all racing to secure reliable generation capacity, pushing valuations higher and reshaping how these assets are bought and sold.
Few people are better positioned to explain this shift than Neel Somani, a former quantitative researcher at Citadel who specialized in energy markets. Drawing on his background in optimization, pricing models, and power market dynamics, Neel Somani offers a rare, inside-out view of how power plants are valued, and why the rules are changing.
Why Everyone Suddenly Wants a Power Plant
At a basic level, the demand story is straightforward. AI workloads require massive amounts of compute, and compute requires electricity. But unlike traditional industrial demand, AI-driven demand is:
This has created a bottleneck—not in chips, but in power.
As Neel Somani puts it: “If you want to understand where AI is going, you have to understand how electricity is priced and dispatched.”
That reality has turned power plants into highly sought-after assets. But buying one is not as simple as evaluating a traditional business. Power plants operate in markets governed by physics, regulation, and real-time pricing, making their economics both complex and highly dynamic.
The Core Unit: Generators and Capacity
At the heart of any power plant are its generators. These generators have a fixed capacity, typically measured in megawatts (MW), which defines how much electricity they can produce at maximum output.
But not all megawatts are equal.
Each generator has a cost of production, which depends on:
Power producers don’t simply run all generators at once. Instead, they follow a principle known as economic dispatch:
Use the most efficient (lowest-cost) generators first, and only bring on more expensive units as demand and prices rise.
This creates a natural ordering of supply, often referred to as the merit order curve.
The Key Rule: Don’t Produce Unless It’s Profitable
The single most important rule in power generation is deceptively simple: Do not turn on a generator unless the market price exceeds your cost of production.
If it costs $50 per megawatt-hour (MWh) to produce electricity, and the market price is $40, you stay offline. If the price rises to $60, you turn on and capture the margin. This decision framework is what makes power plants behave like financial instruments.
Power Plants as Call Options
One of Neel Somani’s key insights, drawn from his quantitative finance background, is that a power plant can be thought of as a call option on electricity prices.
This is where the concept of the Heat Rate Call Option (HRCO) comes in.
What is an HRCO?
A Heat Rate Call Option represents the right, but not the obligation, to convert fuel into electricity and sell it at market prices. Neel Somani of Eclipse explains that the “heat rate” refers to how efficiently a plant converts fuel into power.
In practice, this means:
This optionality is incredibly valuable, especially in volatile markets.
Financing a Power Plant: From Banks to Hedge Funds
Because power plant revenues are uncertain and tied to volatile market prices, traditional lenders like banks are often hesitant to finance new projects.
Instead, developers turn to more sophisticated financial structures:
In this sense, building a power plant is not just an engineering project; it’s a financial engineering problem.
The Traditional Approach: Discounted Cash Flow (DCF)
A more fundamental way to value a power plant is through projected cash flows.
The process looks like this:
This gives you a net present value (NPV), the “intrinsic” value of the plant.
But there’s a problem.
Why Traditional Valuations Are Breaking Down
In today’s market, the NPV often doesn’t match reality.
According to Somani’s framework:
In some cases, assets are trading at multiples 5x higher than historical norms.
Why?
Because buyers are not just valuing current cash flows, they’re valuing:
For a hyperscaler building AI infrastructure, owning a power plant isn’t just an investment; it’s a hedge against being unable to operate at all.
Regional Pricing: The Hidden Variable
Another critical factor is location.
Electricity markets are highly regional, with prices varying based on:
To estimate revenue, producers look at regional price maps, often referred to as nodal or zonal pricing.
Two identical power plants in different regions can have vastly different economics. Neel Somani of Eclipse understands that this is why site selection and grid interconnection are just as important as the plant itself.
How Producers Think About Their P&L
At a high level, a power producer’s profit and loss (P&L) comes down to:
Revenue:
Costs:
Profit:
But layered on top of this are:
The result is a business that looks simple on the surface but is deeply complex in practice.
The New Reality: Power as a Strategic Asset
What’s changed in the AI era is not just demand, it’s who the buyers are. Historically, power plants were owned by utilities or specialized energy companies.
Today, they’re being pursued by:
This shift is redefining how these assets are priced. As Neel Somani’s perspective suggests, the “right price” is no longer just about discounted cash flow; it’s about optionality, scarcity, and strategic value.
Final Thought
Buying a power plant used to be a niche, highly technical endeavor. Today, it’s becoming a central question in the future of AI infrastructure. And as Neel Somani’s framework makes clear, understanding power plants isn’t just about engineering, it’s about markets, math, and the ability to price uncertainty itself.
