Amanda Van Dyke
“Everyone is talking about building reactors. Almost nobody is asking where the fuel will come from.”
For years, the energy debate has revolved around wind versus solar, fossil fuels versus renewables, climate versus economics. But quietly, almost unnoticed, the conversation has changed.
The world’s largest technology companies are no longer asking how cheap electricity is.
They’re asking whether there will be enough electricity at all, because every single one of their growth plans are dependent on it.
Artificial intelligence, data centres, electrification, industrial reshoring and energy security have fundamentally altered the equation. Vastly increased, reliable, 24-hour power has become a strategic asset. Suddenly, nuclear power, dismissed by many for decades, is back at the centre of the conversation.
Microsoft is restarting Three Mile Island.
Amazon is buying nuclear-powered data centres.
Google is investing in small modular reactors.
More than thirty countries have pledged to triple nuclear capacity by 2050.
The nuclear renaissance appears inevitable.
Or is it?
Because while politicians are announcing reactors and investors are piling into uranium miners, there is one uncomfortable question almost nobody is asking.
Can we actually fuel all these reactors?
I recently finished reading one of the best deep dives on nuclear markets I have come across, There Is No Alternative – The Definitive Guide to Nuclear: Uranium, Enrichment and the Supply Chain Behind 440 Reactors, a comprehensive examination of nuclear power, uranium and the global fuel cycle by TSCW. While the paper makes a compelling long-term case for nuclear, it also reinforced something I’ve become increasingly concerned about.
The constraint isn’t demand.
The constraint isn’t technology.
The constraint is the supply chain, and in its present form, in the West it cannot keep up.
We’ve Finally Realised We Need Nuclear...
The investment case for nuclear has become increasingly compelling.
Three irreversible trends are converging simultaneously:
-exploding electricity demand from AI;
-widespread electrification of transport and industry;
-rising energy consumption across developing economies.
For decades we believed renewables alone could deliver the energy transition.
Physics is proving otherwise.
The full solar supply chain for delivered electricity is not cheap, and in built intermittency (it only works when the sun shines) means not only does it need expensive back up batteries and significant grid expansions compared to regular baseload, but it also needs accompanying baseload to be built alongside it.
Wind has many of the same issues, and while it is slightly more efficient, but even harder to permit.
Modern batteries solve hourly balancing, they do not solve, daily, weekly and monthly let alone seasonal storage (think winter).
And data centres don’t care whether it’s cloudy or windy.
A hyperscale AI campus needs power every second of every day.
Electricity has quietly become infrastructure rather than simply a commodity.
That changes everything.
Physics Doesn’t Care About Politics
One of the reasons I enjoyed the paper is that it keeps returning to first principles.
Politics can change.
Subsidies come and go.
Governments change every few years.
Physics doesn’t.
Nuclear possesses three characteristics that no other low-carbon technology currently combines:
approximately 92% capacity factors;
extraordinary energy density;
continuous 24-hour generation.
One uranium fuel pellet the size of the end of your finger contains roughly the same energy as:
one tonne of coal;
149 gallons of oil;
17,000 cubic feet of natural gas.
That isn’t an engineering breakthrough.
It is simply how nature works.
Which is why nuclear has quietly become indispensable, and the only commercially viable solution for meeting the surge in global electricity demand.
Investors Are Looking in the Wrong Place
Most investors think uranium.
Some think reactor builders.
A few think small modular reactors.
I think they’re all missing the most important part of the story.
Nuclear isn’t one industry.
It’s an ecosystem.
Mining is only the first step.
Before uranium reaches a reactor it must be:
mined;
milled;
converted;
enriched;
fabricated into fuel assemblies.
Every one of those stages requires specialised facilities that take years—often decades—to permit and construct.
And unlike copper or gold, you cannot simply ship concentrate to whichever refinery offers the best price.
The nuclear fuel cycle is among the most regulated industrial systems on Earth.
Which means capacity cannot appear overnight.
The Real Bottleneck Isn’t Just Uranium
Everyone talks about uranium supply.
Almost nobody talks about enrichment.
Yet enrichment may become the defining geopolitical bottleneck of the next energy era.
Today:
Russia controls roughly 44% of global enrichment capacity.
Conversion capacity is concentrated in only a handful of facilities worldwide.
Kazakhstan produces almost 40% of global uranium mine supply.
Much of Kazakhstan’s historic export infrastructure runs through Russia.
People often assume that once uranium is mined, the problem is solved.
It isn’t.
Natural uranium contains only around 0.7% fissile uranium-235.
Commercial reactors require approximately 3–5% enrichment.
That enrichment process requires extraordinarily sophisticated centrifuge technology, billions of dollars of investment and years to construct.
It cannot simply be scaled because governments announce new reactors.
My Bigger Concern
This is where I begin to diverge slightly from the otherwise excellent paper.
I think the industry may actually be underestimating the scale of the coming fuel problem.
Today, everything appears manageable.
Utilities still have inventories.
Secondary supplies continue to support the market.
Underfeeding has helped increase effective uranium availability.
Some military material has been downblended into civilian fuel.
Mine supply plus these secondary sources have, so far, been enough.
But we are at a tipping point.
There is no more military material to downblend
Utilities have run down all their inventories
We don’t have time to keep overfeeding which was always something we could only do while we has a small oversupply available.
No mines or development stage projects are going to be able to significantly increase production in the next 5 years, let alone get it converted to fuel rods.
Typically 12–24 months from mined U₃O₈ (”yellowcake”) converted, enriched, and load loaded fuel rods that can fuel a reactor.
And that is assuming you have spare conversion, enrichment and fuel assembly plant capacity, which we don’t.
We are mistaking temporary adequacy for long-term security.
Because several trends are converging simultaneously.
Secondary supplies are gradually diminishing.
Western countries are rebuilding defence capabilities.
Russia is becoming a less acceptable supplier.
New uranium mines typically require well over a decade from discovery to production.
New enrichment facilities take years to build.
Meanwhile governments continue announcing reactor programmes, but the typical western build takes 20 years through permitting to production.
4x more expensive and 4x longer than Russia or China, using the exact same technology.
Hyperscalers continue signing nuclear agreements.
And investors continue assuming fuel will simply appear.
I am not convinced it will.
The SMR Assumption
Small Modular Reactors have become one of the hottest investment themes in energy.
I understand why.
They promise:
lower capital costs;
faster deployment;
factory manufacturing;
distributed generation.
They may well succeed.
But there is one assumption built into virtually every SMR presentation I have seen.
That assumption is this:
Fuel will be available.
What if that assumption is wrong?
Every new reactor ultimately competes for the same upstream supply chain.
Many advanced reactor designs also require specialised fuels that currently have even more limited production capacity.
Building reactors is one challenge.
Fueling them is another entirely.
The Market and Higher Prices cannot solve this
Many investors instinctively assume higher uranium prices solve everything.
Mining doesn’t work that way.
Commodity markets eventually respond.
Mining responds very slowly.
Nuclear mining responds even more slowly.
Permitting is lengthy.
Environmental approvals are extensive.
Processing facilities are specialised.
Conversion plants are scarce.
Enrichment plants are scarcer still.
Building an entire fuel cycle takes decades, not years.
Which means by the time markets recognise shortages, there may be little anyone can do in the short term.
Goldman Sachs and Sprott are Right...
Goldman Sachs and the Sprott Physical Uranium Trust are now explicitly framing uranium as entering a multi‑decade structural demand uptrend, driven by reactor growth, SMRs and AI‑era baseload needs.
Goldman’s “Nuclear Nuggets” work argues this is structural demand shift, that long‑term uranium demand assumptions have risen materially, while supply forecasts have barely moved, creating a widening cumulative deficit through 2045.
Their updated model bakes in roughly 20 new U.S. reactors between 2025–2045, several restarts, resumed construction at shelved projects, and raising assumed reactor lifetimes from about 75 to 80 years, all of which adds incremental uranium burn each year.
SMRs are now in the model. A 2026 update explicitly incorporates small modular reactors, projecting nearly 46 GW of SMR capacity by 2045, which alone adds an estimated 62 million pounds of extra uranium demand (almost double present demand) and pushes overall 2045 demand up ~17% versus Goldman’s prior forecast.
Across conventional reactors and SMRs, Goldman now sees a cumulative uranium supply‑demand gap on the order of 1.9–2.3 billion pounds of U₃O₈ between 2025 and 2045, implying a structural shortfall of roughly 30%+ if new mines lag build‑out.
AI, electrification and policy are the drivers. They tie the nuclear build‑out to AI/data‑centre baseload, EV penetration, hydrogen and desalination, arguing that these electrification trends push policymakers toward round‑the‑clock low‑carbon power where nuclear is one of the few scalable options.
Goldman’s conclusion is that this long‑term imbalance is “laying the groundwork for a new contracting cycle,” with rising term contracting volumes and sustained upward pressure on uranium prices rather than a transient squeeze.
Sprott’s recent commentary emphasises that spot prices and miners have rallied on tight supply and a surge in utility contracting, reinforcing investor conviction that the bull case is underpinned by fundamentals rather than speculation.
Utilities re‑entering the term market. They highlight accelerating long‑term contracting by utilities, driven by concerns over future availability and geopolitical risk, which pulls more pounds into long‑term commitments and leaves less discretionary volume in the spot market.
Supply response is slow and fragile. Sprott stresses that many mines were idled after Fukushima, new projects have long lead times, and key suppliers like Kazakhstan and Canada face both technical and political constraints, so supply cannot quickly match the new demand trajectory.
Both reports position physical uranium and uranium miners as vehicles to express a view on this emerging structural deficit, arguing that the combination of growing reactor fleets, SMR deployment and decarbonisation makes uranium a “scarce energy transition metal” with asymmetric upside.
Taken together, the sell‑side (Goldman) and buy‑side/vehicle (Sprott) narratives are converging on the same core idea: uranium demand is being revised up meaningfully for the next 20 years, while supply additions are slow, geographically concentrated and policy‑sensitive, setting up both an investment story and a geopolitical one.
But even those analyses largely focus on uranium itself.
I suspect the bigger story may lie further downstream.
We’ve Seen This Movie Before
The mining industry has repeatedly underestimated future demand.
Copper.
Nickel.
Lithium.
Rare earths.
Each time the market assumed supply would respond.
Each time permitting and project development proved much slower than expected.
Nuclear may simply represent the same phenomenon on a much larger geopolitical scale.
Except this time the bottlenecks are not confined to mining.
They’re embedded throughout the entire fuel cycle.
The Calm Before the Storm?
Perhaps I’m wrong.
Perhaps enough mines will be financed.
Perhaps conversion capacity expands quickly.
Perhaps enrichment plants are built faster than history suggests.
Perhaps we re-learn how to build nuclear on time on budget.
Perhaps governments finally coordinate industrial policy effectively.
I hope so.
Because if they don’t, we could find ourselves in an extraordinary position.
For the first time in decades, the world may genuinely want to build nuclear power at scale.
Yet the limiting factor may not be engineering.
It may not be finance.
It may not even be politics.
It may simply be the inability to produce enough nuclear fuel.
That is not a problem you solve with subsidies.
It is not solved with speeches.
It is solved twenty years earlier through patient investment in mines, conversion plants, enrichment facilities, skilled workforces and long-term industrial strategy.
Unfortunately, much of the West stopped doing exactly that after Fukushima.
China didn’t.
Russia didn’t.
Today we are trying to rebuild capabilities that quietly disappeared over several decades.
Industrial ecosystems are easy to lose.
Very difficult to recreate.
The West Isn’t Just Racing Against Demand—It’s Racing Against Strategy
There is another dimension to this story that receives remarkably little attention.
While much of the West spent the past 25 years allowing its nuclear fuel cycle to atrophy, China and Russia spent the same period systematically consolidating theirs. They didn’t just build reactors; they invested across the entire value chain—uranium mining, conversion, enrichment, fuel fabrication, reactor technology and long-term fuel contracts.
Over the past two decades, China has added roughly 2–3 nuclear reactors per year on average (accelerating to 6–10 approvals or construction starts annually in recent years), growing its operating nuclear fleet at an approximate 9–10% compound annual rate, while Russia has added around one reactor every 1–2 years domestically with nuclear capacity growing at roughly 1–2% CAGR, instead focusing much of its expansion on exporting reactors abroad.
At the same time, they continued investing in the technologies likely to define the next generation of nuclear power. China has commercialised the world’s first Generation IV high-temperature gas reactor, is leading development of thorium molten salt reactors, is bringing its Linglong One SMR to market, and continues expanding its nuclear submarine fleet. Russia has leveraged decades of experience operating nuclear-powered submarines and the world’s only fleet of nuclear icebreakers to develop floating nuclear power stations and land-based small modular reactors capable of powering remote Arctic and Siberian communities, while remaining the global leader in commercial fast breeder reactor technology.
What strikes me is that neither China nor Russia sees nuclear simply as a way to generate electricity.They see it as a strategic industrial ecosystem that underpins energy security, AI infrastructure, military power, Arctic development, shipping, mining and geopolitical influence.
They are building an integrated strategic capability that spans:
civilian electricity generation
industrial heat
hydrogen production
desalination
AI data centres
Arctic logistics
remote mining
naval propulsion
submarine fleets
floating cities
export finance
fuel-cycle dominance
advanced reactor R&D
The West often asks, “How many reactors should we build?”
China and Russia are asking a much bigger question:
“How do we build a complete nuclear industrial ecosystem that gives us strategic, economic and geopolitical advantage for the next century?”
That is why I believe the coming challenge extends far beyond uranium itself. If global demand accelerates faster than Western fuel-cycle capacity can be rebuilt, countries that already control much of the world’s enrichment, conversion and advanced reactor capability are unlikely to prioritise supplying strategic competitors. After spending a quarter of a century securing these advantages, it would be surprising if China and Russia chose to surrender them just as they become most valuable.
The world may discover that the greatest constraint on the nuclear renaissance is not building reactors—but rebuilding an industrial ecosystem that others never allowed to disappear.
The Bottom Line
The nuclear renaissance is real.
The demand is real.
The physics are undeniable.
The strategic importance is obvious.
But investors, policymakers and hyperscalers appear to be focusing almost entirely on reactors while paying surprisingly little attention to the extraordinarily fragile supply chain that sits behind every kilogram of nuclear fuel.
The world currently has just enough.
That is very different from having enough for the future.
My concern is that we are standing in the calm before the storm.
Everyone is celebrating the return of nuclear.
Very few are asking whether we have invested enough to fuel it.
If we have the industrial ecosystem to support it.
History suggests that when demand meets decades of underinvestment, commodity shortages rarely announce themselves politely.
They simply arrive.
And by then, it is already too late.
Conclusion
Nuclear power is no longer primarily an environmental issue, that issue has been settled, nuclear over the last 75 years is the safest form of civilian mass power generation, more so than oil gas solar or wind. Today it is question of economic competitiveness, energy security and geopolitical resilience.
The world’s electricity system has reached a point where reliable, high-capacity, low-carbon generation is indispensable. Nuclear’s unique combination of reliability, energy density and scalability makes it exceptionally difficult to replace. While reactors receive most public attention, the real strategic value lies within the complex global fuel cycle, mining, enrichment and conversion, where decades of underinvestment and geopolitical concentration have created critical vulnerabilities. As AI, industrial electrification and energy security concerns continue to accelerate demand, the countries and companies that control these bottlenecks are likely to hold disproportionate strategic and economic advantages over the coming decades. The coming nuclear renaissance is not being driven by ideology but by the immutable constraints of physics, infrastructure and national security, and what is very clear to me is we are nowhere near ready…..
No comments:
Post a Comment