Six months ago, I called attention to a quiet revolution in the energy markets.
At the time, I profiled small modular reactors (“SMRs”). The name says it all.
They’re nuclear fission reactors similar to the large nuclear plants I’m sure you’ve seen, but with two important improvements. As the name suggests, they’re smaller. The reactors themselves can fit on the back of a semitruck. That means there are more areas where they can be deployed.
They’re also modular, which just means they’re manufactured in a factory setting and assembled on site. That makes them cheaper and faster to stand up.
There are also other benefits, which we’ll revisit below.
As I said in the original essay:
The potential of this industry is almost limitless, and it resonates with hardcore capitalists and environmentalists alike.
Since then, the rally in nuclear tech stocks has been nothing short of explosive (more on that below).
But now the question is: Did the fundamentals follow the hype, or are we in bubble territory?
In this edition of the Wide Moat Daily, we’ll revisit the thesis, what changed, stock performance, new risks and tailwinds, and offer an updated view on one of the most promising parts of the Great Reindustrialization of America.
The Original Thesis
Most people still don’t understand the massive power demand coming from crypto mining, cloud computing, and artificial intelligence (“AI”).
Globally, the power need for data centers tripled from 58 terawatt-hours (TWh) in 2014 to 176 TWh in 2023, and has more than doubled since then. According to the latest projections, the current consumption of roughly 450 TWh is anticipated to double again by 2030. Global data-center power consumption already exceeds that of the entire German economy, the largest in Europe.
Notice I haven’t mentioned the fact that more than 40% of Fortune 500 companies and many governments are targeting to be “net zero” on carbon emissions by 2050. Data-center demand alone requires a seismic shift in the power industry. There are three options once you understand the economics and engineering: natural gas, renewables, or nuclear.
Natural gas is great and is the No. 1 generator of electricity in the U.S. for a reason. But there is a practical limit to how much extra capacity we can bring online, and distributing it to new areas of the country is costly. And while clean for a fossil fuel, it still generates carbon emissions.
Renewables help countries and companies accomplish environmental and political goals, but they can be expensive and are geographically sensitive. Wind and solar work well in some areas and aren’t economical in others. Renewables are also intermittent by nature. That’s a “no go” for data centers that need power 24/7. And in order to deliver that, renewables need to be paired with a lot of expensive (and environmentally unfriendly) battery support. Not ideal.
Modern nuclear, on the other hand, has all the benefits with few drawbacks.
New reactors require minimal inputs and can even recycle the spent nuclear fuel. NuScale’s 77 megawatt-hour (MWh) SMR can operate 24/7 for a year on less than two tons of uranium. Advanced SMRs require about half the fuel as older designs.
Their operating costs are among the lowest of any energy source, and their reliability is second to none. Even if nuclear takes off like many think, we have decades of uranium using existing deposits.
That, in a nutshell, is the bullish thesis I presented six months ago.
Let’s look at what’s changed since…
Regulatory Tailwinds
Like it or not, governments can’t keep their hands off the energy industry.
Regulations and politics have hampered nuclear for decades, and that needs to change for these new technologies to really take off. And that’s exactly what we see.
In May 2025, the U.S. Nuclear Regulatory Commission (“NRC”) approved an updated 77-megawatt (“MW”) SMR design from NuScale. That’s a big jump from the 50 MW system they approved before, and economics, scale, and marketability all improved.
There is momentum outside the U.S., too. This matters because the industry benefits as supply chains become larger and more durable. Rolls-Royce SMR was the winner of the U.K.’s SMR competition and sparked more interest in that nation.
The International Atomic Energy Association began tracking SMR designs around the world, although there are only a few in the advanced stages, and the U.S. is home to most. The pros and cons of renewables and batteries are better known today than in the past, and it makes the case for nuclear stronger. That’s whether you are a government bureaucrat focused on clean energy or an industrialist with the sole goal of maximizing profits long term.
But that’s not what has me most interested. It’s this.
What Has Really Changed
In my previous issue of the Wide Moat Daily on this topic, I went on the record with four SMR names that I had the highest conviction in. Let’s see how they did.
Oklo (OKLO) is up over 500% year to date, and effectively all of that was since I recommended it. Centrus Energy (LEU) wasn’t far behind, with gains of about 400% since that issue was published. Those two stand out, but the others did well, too. NuScale Power (SMR) is up about 120% since I wrote about it, and BWX Technologies (BWXT) is the laggard with “only” 60% gains.
These gains are driven by the global data-center electricity demand I described earlier. Big Tech knows this, and they need nuclear.
In June 2025, Amazon (AMZN) secured a power purchase agreement (“PPA”) with the Susquehanna nuclear power plant for 1.92 gigawatts.
Last year, Microsoft (MSFT) signed Constellation Energy’s (CEG) largest-ever PPA to add roughly 850 MW of nuclear power to the grid to power data centers. Around the same time, Google (GOOG) signed a master agreement for up to 500 MW of advanced nuclear capacity to serve its U.S. data centers. This cemented SMR-class technology into the hyperscale industry.
Announcements like this are accelerating, and it’s driving nuclear stocks higher.
So, are we in a nuclear bubble, or is the industry just beginning its takeoff?
Focusing on NuScale for a moment, it is the only SMR firm with double NRC approval and arguably has the most robust supply chain. It has several agreements to provide future revenue and has about $500 million of liquidity on the balance sheet.
Source: Getty Images
NuScale’s story has changed from “if” to “when” to “how fast” they can make their SMRs. Current production capacity is 20 modules a year. Quarterly revenue has increased from $1 million last year to over $8 million in the most recent quarter.
Now, the company is absorbing $40 million to $50 million in operating costs as it gets up and running, so real profitability is yet to be achieved.
At the same time, just one or two additional major contracts makes NuScale profitable. And the growth potential beyond that is 5 times to 10 times in the next few years if it can secure contracts to power more AI data centers.
Final Thoughts
What I like best about the industry isn’t their valuations. In fact, most traditional valuation metrics won’t be very helpful. There’s no point comparing NuScale with its recent earnings. It has none.
More important is their market capitalizations. Despite 218% gains in the past year, NuScale is worth $11.3 billion. Oklo is up over 1,200% over that same time, but its market cap is still less than $20 billion.
If this industry does half as well as I expect, I believe several of these companies will be worth $100 billion in the medium term. That assumes they manage to make a few billion in annual profits and trade for 20 times to 30 times earnings.
No one knows if that’ll be in three years or 10, or which will turn out to be the most successful. That’s why I recommend diversifying your exposure to the industry. And realize that there will be lots of volatility along the way as these companies are not yet proven, and a lot rides on NRC approvals nobody can control.
Nvidia (NVDA), the spear of AI, is up 59% in the last year. Palantir Technologies (PLTR) is up the most of the established AI companies, as far as I can tell, at 381%. Many of these nuclear companies have done better despite riding the same economic and technological waves.
This touches on one of Wide Moat’s core themes: We strive to find alternative ways to capitalize on major trends that others overlook. In this case, using SMRs to capitalize on the AI wave has paid serious dividends.
You and I might not be able to get into OpenAI’s recent private equity raise, valuing the firm at $500 billion. And thanks to outside-the-box research like what we’ve done with SMRs, we don’t need to.
While everyone else focuses on the flagship companies, we’ll continue finding the real gems of AI and broader reindustrialization of America.
Regards,
Stephen Hester
Chief Analyst, Wide Moat Research
|