Why Neel Somani Believes Energy Markets Are a Systems Problem

Most people think of electricity as a utility. You flip a switch, the lights come on, and you pay a bill at the end of the month. But for Neel Somani, former quantitative researcher at Citadel, energy markets are far more revealing: a living demonstration of how complex systems behave under pressure and how decisions made at the architectural level shape every outcome downstream.

Somani spent years covering power markets at Citadel, one of the world’s most sophisticated quantitative trading firms. That experience gave him a window into how energy infrastructure operates not as a simple supply-and-demand equation but as an interconnected system where every component influences every other.

For leaders trying to understand why energy costs behave as they do and what that means for their organizations, Neel’s perspective offers a starting point.

The Price is Set at the Margin

While most people assume that electricity prices reflect an average cost across all generators feeding the grid, the reality is that the price is determined by the last, least-efficient unit of power that must be turned on to meet demand.

“The price for power is based on the last megawatt of power that’s produced,” Somani explains. That single principle has enormous consequences for how markets behave, especially when demand spikes.

In a typical grid, operators stack generation sources from cheapest to most expensive. Renewables come first because their marginal cost is effectively zero. Then come nuclear plants, followed by progressively less-efficient natural-gas generators. At the very bottom of the stack sit oil generators, expensive, inefficient, and almost never used. Except when they have to be.

New England as a Case Study in the System

To illustrate how margin-based pricing interacts with the physical constraints of infrastructure, Neel points to New England in winter as one of the clearest examples of a cascading systems problem.

New England faces a structural challenge that most other regions do not. Natural gas serves a dual purpose: it heats homes, and it powers generators. In the winter, residential heating demand is not optional. If people do not heat their homes, pipes freeze, and the resulting damage is costly. So when temperatures drop, a fixed and substantial portion of natural gas supply gets committed to heating before the grid ever sees it.

That constraint forces grid operators to turn on oil generators to make up the difference. And because the price of electricity is set by the most expensive unit running, the moment oil generators come online, power prices climb sharply.

But the cascade does not stop there. “Think about what happens if you’re a natural gas salesman,” Somani explains. “You’re thinking that anyone you sell natural gas to is probably making a lot of money because they can put it in a natural gas generator, and they can get paid the oil price while only paying natural gas cost.” That arbitrage opportunity drives natural gas sellers to raise their prices as well, up to the point where the cost of generating a megawatt from natural gas equals that from oil.

The result is that power and natural gas prices spike simultaneously, not because of any market failure, but because of rational behavior by every player. Each actor is optimizing locally. The system outcome, however, can appear chaotic or predatory to anyone who does not know the underlying structure.

This is the kind of interdependence that Neel believes decision-makers should internalize. Systems do not fail because individuals behave irrationally. They often yield difficult outcomes because individuals behave exactly as they are incentivized to.

California and the Complexity of the Energy Transition

New England’s challenge is largely one of material constraints. California’s challenge is a different kind of systems problem, one driven by the interaction between renewable energy policy and grid architecture.

Somani describes California’s generation stack as cleaner but more volatile. The state has eliminated most of the coal and dirtier generation units that complicate grids elsewhere. In their place is a heavy reliance on renewables, primarily solar power, combined with natural-gas backup. During the day, solar supply is abundant and cheap. But the transition from afternoon to evening creates a problem.

“During the daytime, you have the sun out, so you have this very cheap, zero marginal cost solar power, and then in the evening, the sun goes down, so we lose that solar power, and everyone turns on their power all at once. So then the price spikes.”

It is not a failure of the renewable transition. It is an architectural feature of a grid that adds significant solar capacity without proportionally increasing the fast-response generation needed to fill the gap when that solar disappears.

Making matters more complex is the nature of the backup generation that fills that gap. “There are basically two types of gas turbines,” Somani notes. “There are combined cycle gas turbines and simple cycle gas turbines, and the ones that turn on really fast are the simple cycle gas turbines, but they’re also less efficient.” Speed of response and efficiency trade off against each other, and that trade-off gets priced into every evening electricity bill in California.

The solution Somani points to is not simply to add more renewable capacity. It is investing in ancillary services, the storage and demand-response infrastructure that can absorb cheap midday power and release it when prices climb.

“Batteries buy that cheap solar power during the daytime, they dispatch it in the evening, and they make that difference as profit.” This is systems thinking applied to energy: identifying where the structural mismatch is and designing around it rather than adding capacity at the wrong point in the stack.

What Energy Markets Teach About Systems Governance

Neel Somani’s work sits at the intersection of technology, economics, and governance, and his analysis of energy markets reflects the same mental model he applies across domains. The central lesson is that systems produce outcomes based on their architecture, not their intentions.

In energy, that means pricing spikes are often not the result of manipulation or policy failure but the predictable output of rational actors operating within a structure that creates those incentives. The governance question is not why the actors behaved badly. It is whether the architecture should be redesigned.

That framing is directly applicable to how organizations think about other complex systems, including the AI systems, financial infrastructure, and supply chains that increasingly drive business outcomes. In each case, Somani’s approach is the same: map the stack, identify where marginal costs and interdependencies create cascades, and design around structural vulnerabilities rather than patching outcomes after the fact.