America Has the Power. Can It Get Through the Paperwork?

AI data centers can be built faster than the power system can connect them. Meanwhile, America has used nuclear fuel that still contains energy, new reactors moving through test programs and isotope projects producing products the country needs. The question is whether the rules can protect the public without trapping workable solutions in paperwork.

AI is increasing electricity demand faster than America can approve the grid connections, advanced reactors and fuel systems that may be able to supply it.

“We have computers waiting for electricity, nuclear material waiting to become fuel, companies waiting for permission—and government offices waiting for another government office to finish its paperwork.”

A data center can go from a set of drawings to a finished building faster than the power system can decide where to plug it in.

That is becoming one of the strangest problems in America’s artificial-intelligence boom. Technology companies are preparing facilities that may use as much electricity as a town or small city. The buildings can be designed, financed and constructed in a few years. New generating plants, transmission lines, substations and regulatory approvals can take much longer.

The result is a race between two clocks.

On one clock, artificial intelligence is expanding rapidly. On the other, utilities and government agencies are trying to determine whether the grid can serve these enormous new customers without weakening reliability or shifting the cost to everyone else.

At the same time, the United States is sitting on tens of thousands of tons of used nuclear fuel that still contains substantial potential energy. Advanced reactors may eventually recover some of that energy. Reactors can also produce valuable isotopes used in medicine, manufacturing, research and space exploration.

The ideas are not imaginary. The machines are beginning to move beyond paper. But each piece of the solution travels through a different regulatory door, and the doors do not always open at the same speed.

When “Red Tape” Is More Than a Figure of Speech

The federal government has openly acknowledged that the approval process for data centers and the infrastructure supporting them can be too slow. In 2025, the White House ordered federal agencies to accelerate permitting for large AI data centers, power projects and transmission facilities.

Then, in June 2026, the Federal Energy Regulatory Commission took a more direct step. FERC ordered the six regional grid organizations under its jurisdiction to justify or reform the rules governing how data centers and other large electricity users connect to the transmission system.

That action matters because the largest obstacle is often not a local building permit. It is the electrical interconnection process.

Before a giant new customer connects, engineers must study whether the grid can handle the load. They may determine that a new substation, transmission line, generating plant or other equipment is necessary. Those upgrades can be expensive, and the studies can become tangled when several proposed projects are competing for the same capacity.

FERC also identified another problem: some developers submit requests in several locations while searching for the quickest and least expensive connection. Those duplicate or speculative applications can clog the system, inflate demand forecasts and force utilities to study projects that may never be built.

So, yes, some red tape is outdated or needlessly repetitive. But some of what gets called red tape is the work required to answer four questions that affect every household connected to the system: Is the project real? Is enough power available? What must be built to serve it? Who pays if the project disappears?

The Extension-Cord Problem

A company may promise to “bring its own power,” but that phrase can mean several things.

It might build an on-site natural-gas plant, contract with a nuclear facility, purchase power from a solar or wind project, install batteries or support the development of new generation nearby. Even then, the data center may still rely on the public grid for backup, balancing, transmission or power during maintenance.

In other words, buying enough electricity on paper is not the same as operating independently from the grid.

FERC is exploring arrangements that could connect large customers more quickly when they bring new generation, accept interruptible service or agree to reduce consumption during grid emergencies. The commission is also pushing cost-recovery agreements intended to keep residential customers from paying for infrastructure built for a project that never arrives.

That is the sensible middle ground. A genuine project should not wait years because six offices are studying the same paperwork. But a family should not be handed the bill for a billion-dollar substation built for a computer campus that was later canceled.

The Fuel We Call Waste

The nuclear side of this story begins with a misleading word: waste.

Used nuclear fuel is intensely radioactive and must be handled with extraordinary care. It is not harmless, and recycling it does not make every radioactive byproduct disappear. But it is also not the nuclear equivalent of ashes after a fire.

The U.S. Department of Energy says more than 90 percent of the fuel’s potential energy can remain after several years in a conventional reactor. Other nations, including France, recycle portions of used fuel. The United States currently has no operating commercial reprocessing facility.

Recycling separates materials that may be made into new fuel from materials that still require long-term management. Certain advanced fast-reactor designs may be able to use fuels made from materials older reactors leave behind, extracting more energy and changing the amount and character of the remaining waste.

That does not mean an old fuel assembly can be carried directly from a storage cask to a new reactor. A recycling plant must separate usable material. A fuel-fabrication facility must turn that material into fuel built for a particular reactor. The reactor then converts it into heat and electricity. Each step requires specialized equipment, security, safeguards and regulatory approval.

The Department of Energy began seeking private partners in 2026 to design, build and operate used-fuel recycling, processing and fuel-fabrication facilities. That is evidence that the government is not uniformly suppressing the idea. It is also evidence that America does not yet have the commercial system needed to do it at scale.

The Money Trail Is Real. The Conspiracy Is Not Proven.

There is a reason rumors about nuclear storage sound believable: an enormous amount of money is tied to the current system.

The federal government agreed decades ago to begin taking commercial used fuel from nuclear utilities in 1998. It failed to do so. Utilities kept the material at reactor sites, built dry-cask storage systems and sued the government for costs they say they would not have incurred if the federal promise had been fulfilled.

The Government Accountability Office reports that the federal government has paid billions of dollars in damages to utilities and may owe tens of billions more if the disposal problem remains unresolved.

That is a real financial incentive embedded in a system that has existed for decades. Contractors build casks, provide security, maintain facilities and perform technical work. Communities and companies may depend on jobs connected to storage and cleanup.

But that does not prove storage companies are secretly blocking recycling to protect their income. The documented payments largely compensate utilities for the federal government’s failure to take the fuel as promised. Meanwhile, the federal government is actively soliciting recycling proposals.

The stronger and more responsible question is not whether a hidden group has buried the technology. It is whether a political, legal and commercial system built around storage can change direction quickly enough to begin treating some of this material as a resource.

The Isotope Door

A nuclear reactor can make more than electricity.

When selected materials are exposed to neutrons inside a reactor, they can become radioisotopes used for cancer diagnosis and treatment, industrial inspection, scientific research, space missions and national security. Many of these products are difficult to manufacture, have short useful lives and are supplied by a limited number of aging or foreign facilities.

That creates an opening for companies developing advanced reactors. A smaller non-power reactor may produce valuable isotopes while the company develops its engineering team, supply chain, operating procedures and safety record.

Oklo and its isotope subsidiary, Atomic Alchemy, offer a current example. Their Groves project near Lockhart, Texas, is a low-power test reactor authorized through the Department of Energy’s Reactor Pilot Program. In July 2026, DOE granted startup authorization, allowing the company to load fuel, conduct tests and move toward first criticality.

Groves is not a commercial electric power plant. It is designed to demonstrate construction and reactor operations that may support future isotope-production facilities. Separately, Atomic Alchemy has an isotope-production construction-permit application under review by the Nuclear Regulatory Commission, and Oklo’s commercial power reactors must follow their own licensing path.

The distinction matters. The company is not legally slipping a commercial power plant through an isotope loophole. It is using a faster test-reactor pathway to build experience, produce evidence and reduce the uncertainty surrounding later projects.

Supporters see that as common sense: build something small, prove it works and use real operating data instead of debating drawings for another decade. Critics will reasonably insist that a successful test must not become a substitute for the independent review required before commercial reactors begin supplying cities and data centers.

Both points can be true. A faster path can still require a serious safety path.

One Machine, Several Businesses

The advanced-nuclear companies attracting the most attention are not always proposing a single stand-alone power plant. Many envision connected businesses involving power generation, fuel supply, recycling, isotope production and industrial heat.

That is why the story can sound confusing. The reactor creates heat and neutrons. Other equipment converts the heat into electricity, prepares recycled material as new fuel or separates and packages isotopes. These operations may share one campus and one company, but they are not all performed by the reactor vessel itself.

The business logic is easy to understand. Electricity may be the largest future market, particularly if data centers continue expanding. Isotopes can provide a valuable product sooner. Fuel recycling may eventually turn a liability into a supply source. Experience gained in one operation can help train workers and qualify suppliers for another.

What looks like a collection of side projects may actually be an attempt to rebuild an American nuclear industrial system that has been fragmented for decades.

Why the Timelines Still Do Not Match

Even with faster approval programs, advanced nuclear power is unlikely to solve every near-term data-center connection problem.

A test reactor reaching criticality is a milestone, not the same as a commercial plant delivering hundreds of megawatts around the clock. Recycling demonstrations are not yet a nationwide fuel supply. An isotope license does not authorize electricity sales. A successful pilot still has to become a repeatable, financeable and licensable commercial project.

Data-center developers, meanwhile, are asking for power now.

That means the near-term answer will probably include a mixture of new natural-gas generation, existing and restarted nuclear plants, renewables, batteries, transmission upgrades, more efficient cooling and contracts allowing large computing facilities to reduce demand during grid emergencies.

Advanced reactors and recycled fuel may become a larger part of the answer in the 2030s. The work being done today determines whether that possibility becomes operating infrastructure or remains another promising idea trapped between agencies.

Speed Without Foolishness

America does not appear to be running out of energy ideas.

It has natural resources, existing nuclear plants, advanced-reactor developers, enormous quantities of used fuel, talented engineers and companies willing to invest billions in new infrastructure. It also has a regulatory system divided among federal agencies, state commissions, regional grid organizations, local governments and utilities.

Some of those layers protect the public. Some protect the grid. Some prevent companies from leaving taxpayers with abandoned costs. Others were built for a slower era and do not fit projects moving at the speed and scale of artificial intelligence.

The job is not to tear down every gate. It is to make sure each gate has a purpose, one responsible decision-maker and a deadline.

Data centers should prove that their power plans are real and pay for the infrastructure built to serve them. Nuclear developers should be allowed to test new designs and gather operating experience without pretending a test is the same as a commercial power license. Fuel recycling should be evaluated on evidence, cost, safety and security—not dismissed because the present system has grown comfortable storing the material.

The country may already possess much of the technology it needs. The harder challenge is aligning the rules, money and construction schedules before demand outruns the system.

America has the power. The question is whether it can get through the paperwork in time.