FEATURED

Nuclear Is Crowded. The Real Play Isn’t.

Everyone knows nuclear is hot. That’s precisely the problem.

The moment a chip CEO mentions an energy stock, three ETFs file new products and retail flows show up before the fundamentals have even been stress-tested. The primary names — Constellation, Talen, Vistra — have already moved. Some of them significantly. What hasn’t moved yet is the infrastructure layer underneath, and that’s where the conversation gets more interesting.

The Scale of the Problem

Goldman Sachs Research now forecasts US data center power demand will climb from 31 gigawatts in 2025 to 66 gigawatts by 2027 — more than doubling in two years from a single use case. The Electric Power Research Institute goes further: data centers could consume between 9% and 17% of all US electricity generation by 2030, up from about 4% in 2023. Those estimates are roughly 60% higher than EPRI projected just two years ago.

That doesn’t account for EV adoption, manufacturing reshoring, or industrial electrification running simultaneously. The grid was not built for this.

On the supply side, President Trump signed four executive orders on May 23, 2025, targeting an increase in US nuclear capacity from approximately 100 gigawatts today to 400 gigawatts by 2050. A 4x increase over 25 years. The companies building the pipes, turbines, fuel enrichment facilities, and grid interconnections to support that expansion have barely moved.

Where the Crowding Is

The crowded trade is spot power pricing exposure through large existing nuclear operators. Good business. Well-understood. Already priced. The consensus owns it.

What isn’t priced: the fuel supply chain.

Advanced reactor designs like small modular reactors require high-assay low-enriched uranium fuel — HALEU — which currently has almost no commercial production capacity in the United States. The DOE’s own program page states it plainly: HALEU is not currently available from domestic suppliers, and gaps in supply could delay the deployment of advanced reactors. In January 2026, DOE awarded $2.7 billion to four contractors to expand domestic enrichment capacity over the next decade. The plants that will supply that fuel at scale don’t exist yet. The companies building them are still small-cap or private.

The SMR Timeline Is Now Real

Construction of the Western world’s first grid-scale SMR began in May 2025 at Ontario Power Generation’s Darlington site in Canada, using GE Vernova Hitachi’s BWRX-300 design. The first unit is currently scheduled to be operational in 2030. OPG applied for its operating licence in March 2026 — the foundation pour already cleared its first regulatory hold point.

2030 is four years away. Fuel contracts for that reactor need to be locked in well before commissioning. The uranium enrichment capacity to service the broader SMR wave does not currently exist at the required scale. That’s a structural supply gap, not a cyclical one.

Slight tangent, but it matters: the Breakthrough Institute found that US enrichment capacity is currently sufficient to cover only 10% to 25% of projected annual HALEU needs by 2050. Even with the $2.7B DOE investment, multiple analysts note the central constraint going forward is time, not capital.

The Grid Problem Nobody Talks About

Here’s what’s interesting: even if every nuclear expansion plan executes on schedule, transmission and grid capacity will almost certainly lag behind demand. Interconnection queues have swelled to historic levels, with gigawatts of ready-to-build projects waiting years for grid connection studies and upgrades. New solar farms and battery storage facilities can be constructed in months. The transmission lines needed to deliver their power often require a decade or more of planning, permitting, and construction.

That creates a second investable theme entirely. Grid hardening and transmission buildout. Utilities building new high-voltage lines, transformer manufacturers, and companies providing grid management software are all structural beneficiaries of the same AI power demand wave, with considerably less crowding than the nuclear headline trade.

Three Ways This Plays Out

Optimistic case: SMR deployment accelerates ahead of schedule. HALEU fuel demand materializes in 2027-2028. Uranium prices spike as civilian demand intersects with defense enrichment requirements. Grid operators become bottleneck assets with regulated returns. The whole energy infrastructure complex moves higher.

Middle case: Nuclear expansion proceeds, but permitting delays push commissioning timelines 12-18 months. Grid investment runs at current pace. Power prices stay elevated but not explosive. Companies with contracted revenue outperform merchant generators.

Downside case: Government funding or policy support for nuclear energy changes mid-decade. A future administration resets priorities. Or solar and battery storage costs fall fast enough to compete with nuclear economics, reducing urgency. The fuel cycle buildout stalls for lack of committed offtake.

The Part People Skip

Real assets, and power capacity in particular, have become the dominant allocation theme among capital providers in 2026. Three forces are behind it: AI infrastructure demand is creating physical bottlenecks in electricity supply; the softening higher-for-longer rates environment is reviving appetite for long-duration assets; and energy transition capital is concentrating in markets where the gap between supply and need is widest.

The nuclear headline trade is owned. The fuel cycle and grid buildout beneath it mostly aren’t. That gap tends to close eventually — usually faster than people expect once the primary story gets crowded enough.

Worth keeping an eye on.