Midnight MarketsV4.0
⚠ Not financial adviceNot financial advice. Independent research, not investment recommendations. The author and website make no representation as to accuracy or completeness. Do your own due diligence before acting on anything you read here. Read the full disclaimer →
Sector III · 4 chokepoints · 14 names

Energy

The Other Watts — Non-AI Energy

Imagine the US electrical grid is a 100-year-old apartment building. Most of the wiring was installed when the only thing you plugged in was a lamp and a refrigerator. Now suddenly everyone on every floor wants an electric car, a heat pump, induction stoves, a home battery, and — yes — a small data center in the basement. The building does not have the capacity. The fuse box is melting.

To fix it you need: more electricity (which means more uranium, more natural gas, more reactors), much bigger fuse boxes and wires (transformers, switchgear, transmission lines), and the crews to install it all (electrical contractors are now sold out 18+ months in advance).

Strip out the AI data-center demand entirely and the grid still needs trillions in spending over the next 15 years. That spending lands on a tiny handful of public companies — and unlike the AI story, these names are not yet trading like the future already happened.

Choke 01

The Fuel — Uranium & Nuclear Fuel Cycle

Uranium Mining · Conversion · Enrichment · HALEU

Nuclear is back. The fuel supply chain is not.

Three years ago, nuclear was a dying industry. Now every major economy is racing to restart reactors and order new ones — driven by AI, yes, but also by EV electrification, industrial decarbonization, and a brutal lesson from European gas dependence. The uranium fuel cycle has four stages : (1) mining the raw ore, (2) conversion to uranium hexafluoride gas, (3) enrichment to reactor-grade material, and (4) fabrication into fuel pellets. Russia controlled ~35% of global enrichment until 2024 sanctions. The West is scrambling to onshore it. The new "HALEU" (high-assay low-enriched uranium) needed for next-gen small modular reactors is essentially a two-supplier-on-Earth situation: Centrus and Russia's TENEX.

Why this is a chokepoint

You cannot build a new reactor without fuel. You cannot enrich fuel without centrifuges that take 5-7 years to build. The bottleneck is structural, multi-year, and getting tighter.

Price

World's largest Western-aligned uranium producer + 49% of Westinghouse

Canadian miner with the highest-grade uranium deposits in the world (Cigar Lake, McArthur River). Owns 49% of Westinghouse Electric (joint with Brookfield) — the dominant Western reactor OEM. Three segments: uranium mining, fuel services (conversion + fabrication), and Westinghouse (reactor sales + services). "Tier-1 utility supplier" — meaning every Western nuclear utility has them under long-term contract .

Price

Only US HALEU producer — the bottleneck-of-the-bottleneck

Operates the only commercial HALEU (high-assay low-enriched uranium) production line in the United States — at Piketon, Ohio, under DOE contract. HALEU is required for essentially every advanced reactor design (Oklo, X-energy, TerraPower, etc.). Until Centrus scales, every next-gen reactor is gated on this one facility.

Price

Largest US-based uranium producer — ISR-method low-cost mines

Wyoming and Texas ISR operations producing at low cost. Hub-and-spoke model centralized at Irigaray. "America First" uranium — political tailwind plus contracts with Energy Fuels' partners. Smaller and more leveraged to spot price than CCJ. Higher beta both up and down.

Price

Owns Rook I in Saskatchewan — the largest undeveloped uranium project

No revenue yet. The entire bet is on getting Rook I built on schedule and budget — typical for a junior miner, but at scale (~30M lbs/year potential, vs world annual production of ~150M lbs). If/when it comes online, it's a top-3 uranium producer overnight. Dilution risk between now and then. Highly speculative.

Choke 02

The New Reactors — SMR & Advanced Nuclear

Small Modular Reactors · Microreactors · Advanced Fission

The first commercial SMRs ship 2027-2029. The stocks are pricing that already.

Conventional reactors take 10-15 years to build, cost $10-20B, and routinely run over budget. Small Modular Reactors (SMRs) are the proposed answer: factory-built reactors of 50-300 MW, shipped to site, identical units in series. None have shipped commercial power yet . NuScale was first NRC-certified (2023), then lost its UAMPS deployment, then signed a Standard Power deal in 2024 for two data-center sites. Oklo is the most ambitious (fast reactor, sodium-cooled, designed for hyperscaler private use). BWXT is the boring profitable name that already builds nuclear components for the US Navy and operates as fuel fabricator for SMRs. Highly speculative group — but the eventual winner is a multi-bagger .

Why this is a chokepoint

If even one SMR design hits commercial deployment at promised cost ($65-80/MWh), it reshapes the entire grid build-out. Hyperscalers want behind-the-meter reactors for AI campuses. Industrial customers (steel, cement, fertilizer) want them for process heat.

Price

Fast reactor "powerhouse" model — pre-revenue moonshot

Sodium-cooled fast reactor developer targeting 15-75 MW "Aurora powerhouse" units — designed to run on recycled nuclear fuel (its second business). Pre-revenue: Q1 2026 net loss $33M. Sam Altman is chairman, which both helps (network) and hurts (governance concerns). First commercial Aurora targeting 2027-2028 at Idaho National Lab site.

Price

Only NRC-certified SMR design in the US — light-water, modular

Light-water-reactor SMR — the conservative, regulatory-blessed approach. The only design the NRC has fully certified for commercial deployment in the US . Lost the UAMPS Idaho project in 2023, which crushed the stock — but signed Standard Power for 24 modules across two data-center sites and Romania. Closest to "real" of any SMR name. Still pre-revenue from commercial deployment.

Price

US Navy nuclear-component monopoly + SMR fuel fabrication

Old, profitable, government-funded. BWXT is the sole supplier of nuclear propulsion to the US Navy — a contracted, multi-decade business. Building HALEU fuel facilities. Won contract to make TRISO fuel for X-energy. Less explosive upside than Oklo, but actual profits, actual revenue, and bipartisan political support. The "boring uncle" of the SMR thesis — and the one professional managers actually own.

Choke 03

The Pipes — Natural Gas & LNG Midstream

LNG Export · Pipelines · Permian Gas

The bridge fuel that's now a 30-year fuel. Permian gas is structurally constrained.

The "energy transition" has quietly become an "energy addition" — solar and wind are growing, but so is natural gas, and so is nuclear, all simultaneously. The reason: load growth is bigger than any single source can fill . US LNG export capacity is on track to double by 2028. European gas dependence on Russia is gone, replaced by US/Qatar LNG. Permian Basin associated gas (gas that comes out as a byproduct of oil drilling) is so abundant it's often flared — and the constraint is pipeline takeaway. Three names dominate: Cheniere (the largest US LNG exporter), Williams (gas pipelines), and Kinder Morgan (the largest US midstream by mile of pipe).

Why this is a chokepoint

You can build a new LNG export terminal in 4-5 years if you start now. You cannot build a new long-haul pipeline in less than 8 years anywhere in the US (permits). The infrastructure that exists today is structurally undersized.

Price

Largest US LNG exporter — Sabine Pass + Corpus Christi terminals

Operates the two largest US LNG export terminals — Sabine Pass (Louisiana, 30 mtpa) and Corpus Christi (Texas, 15 mtpa + expansion to 35 mtpa). Long-term sale-and-purchase agreements ("take-or-pay") with European utilities, Korean utilities, and oil majors. ~90% of revenue is contracted — closest thing to a utility in the LNG space. Stage 3 Corpus Christi expansion ramping through 2027.

Price

Largest US gas pipeline operator — Transco backbone

Operates ~33,000 miles of natural gas pipelines. Transco connects Gulf Coast LNG terminals to East Coast cities — the most important single piece of gas infrastructure in the US. Multiple expansions underway. Steady, slow compounder with a real yield. Data-center gas demand is a small but accelerating tailwind.

Price

Largest US midstream by total pipeline mileage — 82,000 miles

Boring, steady, ~6% dividend yield. Carries about 40% of US natural gas at some point in its journey to end users . Permian-to-Gulf takeaway projects (Gulf Coast Express, Permian Highway) are the growth pieces. Re-rating possible if data-center gas demand becomes a Wall Street narrative. Bond-like steadiness, equity-like upside if rates fall.

Choke 04

The Crews & Wire — Grid Build-out (Non-AI-Specific)

Transmission · Electrical Contractors · Cables · Geothermal

Even with zero data-center demand, the grid is one of the biggest secular spends of this decade.

Two-thirds of US transmission lines were built before 1970. Now we're trying to add EVs, heat pumps, factory electrification, and yes data centers on top. The result: multi-year backlogs for the firms that actually build the grid . MYR Group, MasTec, Quanta Services — these are the electrical contractors that put up high-voltage lines. Prysmian, Nexans — the European cable giants that make the actual transmission wire (very few producers worldwide; subsea cable is a true duopoly). Ormat — the dominant geothermal developer for behind-the-meter clean baseload. None of these names need AI to thrive — they need EVs, heat pumps, and reshoring, all of which are happening simultaneously.

Why this is a chokepoint

The grid is the rate-limiter on the entire electrification thesis. You cannot decarbonize, you cannot reshore, you cannot deploy AI without rebuilding the wires. And the wires are sold out.

Price

Pure-play electrical contractor — transmission + distribution

One of the largest pure-play transmission & distribution (T&D) electrical contractors in North America. Build power lines, substations, distribution networks. Customers: investor-owned utilities, public power. Backlog at all-time high. Smaller than Quanta (PWR, already in AI section) so the same thesis applies with higher beta.

Price

Multi-segment infrastructure contractor — pipelines + power + comms

Diversified infrastructure: builds pipelines, electric transmission, communications networks, and wind/solar projects . Less pure than MYRG but with broader exposure. Heavy industrial backlog. Operating leverage is significant as utilization rises.

Price

Global cable giant — subsea HVDC duopoly with Nexans

Cables of every kind — from telecom fiber to massive subsea HVDC interconnects that take 5-7 years to manufacture per project. North Sea Link, Viking Link, IFA2 — Prysmian on every one. Acquired Encore Wire (US distribution) in 2024 to crack the North American market. Boring but extremely difficult to disrupt.

Price

Largest pure-play geothermal — behind-the-meter clean baseload

Operates ~1.3 GW of geothermal plants globally, with another ~1 GW in development. The only commercially proven non-nuclear, weather-independent, 24/7 clean baseload . Fervo Energy is the venture-backed competitor; Ormat is the established incumbent. Long PPA structures with hyperscalers (Google was first — others followed). Slow grower but compounding capacity quietly.