Colorado’s electric regulatory regime is nearly 100 years old, governed by laws written when Franklin Roosevelt was in the White House, when telephones and radios were cutting-edge technology, and electricity demand grew slowly and predictably. Those laws created vertically integrated monopoly utilities with government-granted exclusive service territories in exchange for state regulation and cost recovery through captive ratepayers.
That framework doesn’t match today’s rapidly changing power landscape.
The regulated monopoly model
Under Colorado law, monopoly utilities must serve every new megawatt of load. That means utilities must expand generation and infrastructure in anticipation of demand and then socialize those costs across all ratepayers. When projected load growth fails to materialize, residential and small business customers are left holding the bill — a risk that understandably gives ratepayers heartburn.
Ratepayers are right to be concerned about speculative demand forecasts. John Dinsdale, an analyst with Synergy Research Group told the Colorado Sun late last summer, “Colorado really hasn’t been a major target for hyperscale operators and there is very little activity in terms of developing large data centers or campuses.” I’ve heard similar assessments from multiple data center developers.
Yet under the current regulatory framework, utilities are still expected to plan, build, and invest ahead of certainty.
Utilities do have tools to manage that risk. High tariffs, nonrefundable deposits, minimum bills, collateral requirements, and exit fees can reduce exposure if large loads fail to materialize. But these tools also increase cost, complexity, and delay — and they do not solve the underlying mismatch between hyperscale demand and a monopoly utility model built for slow, predictable growth. They protect utilities after the fact; they do not provide speed to power.
A modern upgrade
This moment presents an opportunity to rethink how we provide electricity to large loads. Instead of forcing hyperscale demand into a monopoly framework that was never designed for it, Colorado should consider giving data centers what they value most: speed to power through ownership of their own generation, provided those systems remain physically islanded from the grid.
Senator Tom Cotton’s (R-Arkansas) bill to amend the Federal Power Act is the first serious effort in Washington to address this problem. It would remove federal regulatory barriers that prevent large electricity consumers from building and operating their own power plants and private electrical networks to serve their own load. These islanded systems would be self-contained, would not rely on the central grid, and would not distort utility integrated resource plans.
This is the relief valve our grid needs.
Under Cotton’s framework, hyperscale loads would carry their own infrastructure costs. They would build and operate their own generation mix, including gas, nuclear, geothermal, or hybrid systems, and manage their own reliability. They would not dilute capital risk into regulated rate bases or distort Colorado’s integrated resource plans.
The cost of serving these enormous loads would rest where it belongs, on the entities creating them.
Just as important, self-supplying large loads gain speed to power by avoiding the years-long regulatory proceedings and permitting cycles that delay projects under the monopoly model.
Islanded systems also remove stress from the grid. Instead of chasing uncertain forecasts with billion-dollar transmission projects, Colorado planners can focus on serving existing customers, including households, small businesses, schools, hospitals, and essential public services, while meeting the growing demand driven by electrification.
Reliability improves as well, an area where Colorado is already under strain. When a data center owns its power system, failures are not absorbed through a rate case. They are punished immediately through downtime, contract breaches, and lost capital. Performance replaces blame shifting.
Best of all, there is no cost shifting from tech companies that may be playing states against each other onto current Colorado customers. Large loads would pay for their own generation.
Colorado must follow suit
Forward-thinking utilities will welcome this evolution. Smart ones will see new opportunities, creating divisions to offer private power engineering, operations, maintenance, and reliability guarantees. They will become infrastructure partners rather than monopoly toll collectors.
Cotton’s bill is only the first step because it fixes federal law. The next step must come from the states. Colorado’s legislature should allow physically islanded grids, contract-only power systems exempt from traditional utility regulation, as New Hampshire already does and as a bipartisan bill in Illinois proposes. This preserves the existing grid while relieving it of large loads it’s not designed to absorb, and it addresses both the demand surge and the need for speed to power without burdening ratepayers already facing rising bills.
This islanded-grid option for large loads is drawing praise from researchers across the political spectrum.
Colorado is trying to meet 21st-century energy demands under a Depression-era regulatory scheme, and it is failing. Requiring current ratepayers to subsidize speculative hyperscale infrastructure will not grow our economy. The solution begins with Congress passing Sen. Cotton’s bill to allow new loads to self-supply and self-pay outside the regulated grid. It ends with the Colorado legislature doing the same.
Amy Cooke is the co-host of Independence Institute’s Power Gab podcast, and president of Always On Energy Research
