AI Picks and Shovels (Arc 3: Keeping it Alive): Constellation Energy (NASDAQ: CEG) — The Power Plant

Picks and Shovels Series | Arc 3: Keeping it Alive | Part 2 of 3

Last week Vertiv sold the loop that stops a 140-kilowatt rack melting the chip. This week the rack still has to be switched on.

Nvidia designs the brain. Arc 1 covered the foundry, the custom silicon, and the merchant challenger that try to turn that design into chips. Arc 2 covered the switching fabric that moves the traffic, the memory that holds the weights, and the interconnect landlord that houses the racks. Every one of those pieces converts electricity into computation and waste heat. Cooling only matters if there is power to cool. Arc 3 moves from the fluid to the electron.

AI data centres do not run on the average of the day. They run on the hour the model is training. Intermittent renewables can supply part of the load. They cannot, on their own, guarantee the continuous megawatts the cluster requires without storage or backup that itself has a cost and a timeline. Nuclear already exists at scale. It already runs at high capacity factors. It can be contracted for twenty years.

Constellation Energy (NASDAQ: CEG) is the largest owner of nuclear generation in the United States, with roughly 22 gigawatts of nuclear capacity across six states after its acquisition of Calpine. It is not a pure-play reactor designer. It is the operator that already owns the plants, already holds the licences, and is now signing long-term power purchase agreements with the hyperscalers that need the electrons.

The Contracted Fleet

In September 2024 Constellation agreed with Microsoft to restart the undamaged reactor at Three Mile Island Unit 1 — now the Crane Clean Energy Center. Microsoft takes the full output under a 20-year contract. Constellation is spending roughly US$1.6 billion to bring the 835-megawatt unit back. The Department of Energy has since provided a loan facility of up to US$1 billion. The restart is targeted for 2028. It is the most visible example of a hyperscaler contracting an entire nuclear unit.

The model has repeated. In early October 2026 Constellation and Google announced a 20-year power purchase agreement supporting 890 megawatts of additional nuclear capacity through uprates at 11 existing units in the PJM region. Constellation is backing that with more than US$4.3 billion of investment. Google also took a 15-year supply agreement for a further 2,700 megawatts across the PJM market. A separate deal with Amazon supports expansion at Calvert Cliffs. These are not merchant sales into a volatile wholesale market. They are contracted offtakes. They give Constellation revenue visibility. They give the hyperscaler a claim on carbon-free baseload that matches its data-centre load.

The PPA and the Visibility

The business model is the power purchase agreement itself. A PPA locks price and volume for a term long enough that both sides can underwrite the capital. For the generator it converts an asset that was once economically marginal into a contracted cash-flow stream. For the buyer it converts an uncertain future power price into a known cost. Federal clean-energy tax credits apply once the restarted units are online. The commercial driver is simpler: the hyperscaler’s need for 24/7 power that renewables alone cannot yet underwrite at the required reliability. Electricity demand forecasts tied to AI have been revised higher repeatedly. Nuclear is one of the few sources that can meet the reliability requirement without a multi-year interconnection queue for new transmission or a decade-long permitting fight for a greenfield reactor.

Friday’s close put Constellation at roughly US$271. That is a market capitalisation near US$96 billion. Trailing earnings per share sit around US$10, for a price-to-earnings ratio in the mid-20s. The stock is down more than 20% year-to-date after a strong prior run. The multiple has compressed from the levels that priced pure scarcity. The contracted offtakes are the reason the multiple still sits where it does. They convert what used to be merchant exposure into multi-year revenue visibility with investment-grade counterparties.

Competition and the Queue

Other nuclear operators hold assets that can be uprated or restarted. Merchant gas plants and new combined-cycle capacity can still clear interconnection queues faster than a greenfield reactor. Battery storage and demand-response programs can firm a portion of intermittent generation. None of them currently offers the same combination of existing licensed capacity, high capacity factors, and the ability to sign multi-decade offtakes with investment-grade counterparties for carbon-free baseload at the scale the hyperscalers are contracting. Constellation’s scale is the differentiator. No other US nuclear owner has the same concentration of operating capacity already under long-term hyperscaler contracts. The practical bottleneck is not reactor technology. It is the timeline between a signed PPA and electrons that can actually reach the rack.

The Aussie Angle

Australia has no operating nuclear plants and no near-term policy path to them. The baseload problem that US hyperscalers are solving by contracting existing reactors is the same problem Australian grids face whenever large, continuous industrial or data-centre load is proposed. The US solution is contractual access to existing nuclear. The Australian solution, for now, is a mix of gas, coal, renewables and storage. Capacity factors and firming costs are different.

For an Australian investor the usual overlays apply. The equity can be held directly via an international broker or indirectly through US indices and superannuation vehicles. Any capital gain or loss is measured against an AUD cost base, so the AUD/USD rate on entry and exit matters. There are no franking credits. Dividends, if paid, are subject to US withholding tax (typically 15% under the treaty with a valid W-8BEN). Holding Constellation does not change the domestic reality of Australian grid connection queues or the firming costs of the local generation mix. It does expose the holder to the US dollar and to the AUD cost base on any capital gain. The power plant is in Pennsylvania or Maryland. The tax return is in Australian dollars.

The Investment Case — In One Line Per Side

Bull case: Constellation already owns the largest US nuclear fleet, is restarting and uprating existing licensed capacity under multi-decade offtakes with Microsoft, Google and Amazon, and is converting assets that once traded on merchant power prices into contracted cash-flow streams with investment-grade counterparties.

Bear case: The capital is still being spent, the restarts still carry regulatory and construction risk, the stock already reflects a meaningful portion of the AI power narrative at a mid-20s trailing multiple, and contracted nuclear does not remove the interconnection and transmission constraints that still sit between the plant and the rack.

The honest read: Contracted nuclear offtakes improve the quality and duration of cash flows relative to pure merchant generation. The valuation assumes those contracts convert on the stated timelines without major regulatory or construction slippage, and that the hyperscalers continue to prefer existing nuclear over the alternatives available on similar timelines. That is a high bar, not an impossibility.

Nuclear solves grid-scale power. It does not solve the problem of a cluster that needs electrons on the same site, behind the same fence, on a timeline shorter than a transmission upgrade. Next week we look at the company that generates power on-site, next to the load, using fuel cells that can be stacked in months rather than years.

Data Sources:

• Constellation Energy announcements on the Crane Clean Energy Center restart and Microsoft offtake (September 2024)
• Constellation—Google 20-year PPA and uprate announcements (early October 2026)
• Constellation—Amazon Calvert Cliffs expansion
• Share price and market capitalisation data as of Friday close prior to publication
• Company filings and investor materials on nuclear capacity and Calpine acquisition

Wall St. Down Under | Australia

Subscribe | wallstdownunder.com.au

Disclaimer: Wall St. Down Under is an independent financial newsletter for informational and educational purposes only. Nothing published here constitutes financial advice, a recommendation to buy or sell any security, or a solicitation of any investment decision. Always do your own research and consult a licensed financial adviser before making investment decisions. Australian investors should consider their own financial situation, objectives, and risk tolerance. Past performance is not indicative of future results.