MacroVoices notes: buyers contracted half their needs and the term price still quadrupled

In the 12 November 2026 episode of MacroVoices, Erik Townsend brings back Sachem Cove CIO Mike Alkin. Alkin argues there isn't enough economically viable uranium for the world's 440 reactors plus the 80 in the pipeline, and that the gap has nothing to do with AI data centres; Kazakhstan's Kazatomprom forecast a 2026 cost of $13–14 a pound at its IPO and sits near $40 today. Educational notes, not investment advice, no tickers, no price targets, and my reading may be wrong.
Contents
- What the episode is about
- There’s no AI in the uranium thesis, but there is AI in the basket
- Where the cost curve broke: the cheapest producer sits on the industry average
- The reported price is not where the market trades
- Half the needs contracted, and the term price quadrupled
- Nobody pays the fuel buyer for calling the bottom
- Services moved first, the product moves later
- If the fundamentals are this good, why is my position still falling?
- Everyone says this is a structural bull market — how do I know it isn’t a story?
- Sources worth reading
- One thing to take with you

It is the material energy of the country — the universal aid — the factor in everything we do.
—— William Stanley Jevons, The Coal Question (1865)
In the 12 November 2026 episode of MacroVoices, episode 558, host Erik Townsend brings Sachem Cove chief investment officer Mike Alkin back on. Alkin’s thesis fits in one line: there is not enough economically viable uranium to feed the world’s 440 reactors — about 420 of them actually generating — plus the 80 new units in the pipeline, and that gap has nothing to do with AI or data centres. Two of his numbers: Kazakhstan’s state-owned Kazatomprom said in its 2018–19 IPO prospectus that it expected a 2026 cost of roughly $13–14 a pound, and sits near $40 today, while the term price has reached $98–99 and utilities have contracted only about half of what they consume. These are one uranium fund manager’s own figures and inferences; he declines to forecast a price on the show and says nobody knows what day the turn comes.
What the episode is about
MacroVoices is Townsend’s weekly macro show, aimed at professionals and experienced private investors. Alkin runs a uranium-only fund and has been on this trade for about ten years; he says he started expressing the view publicly when uranium was $17–20 a pound.
The episode lands at an awkward moment. Uranium prices are rising and fuel buyers are turning bullish one by one, yet the mining shares have been beaten back to what Townsend calls the liberation-day lows. The pressure comes from the Wall Street end — fear that the AI capex boom goes bust, plus post-election talk that Democrats could make AI and data centres public enemy number one — and everything nuclear gets sold with it. Alkin also brought the deck he used at the Nuclear Energy Institute fuel conference; they ran out of time and never got to most of the slides.
There’s no AI in the uranium thesis, but there is AI in the basket
Townsend’s first question has two halves: does it make sense to tie uranium fundamentals to AI capex, and is the market doing it anyway? Alkin says yes to the second and no to the first. If not a single data centre opened, the structural deficit is still there; nuclear went from a half-percent grower when this trade started to a 4–5% grower, and none of that depends on which AI model turns a profit.
Then comes the uncomfortable part. As a portfolio manager he has to know what the market cares about, because the market can stay irrational for a long time — and some of the junior uranium miners did this to themselves, putting “we’re part of the AI trade” into their decks. His line is: be careful what you ask for. Once you’re labelled, people start asking whether that capex gets spent, and that is a fair question. Then Goldman and Morgan build the baskets, a client calls up and asks for the AI basket, and nuclear rides along. At that point flows move the price and the facts don’t.
Where the cost curve broke: the cheapest producer sits on the industry average
This was the part I most wanted to write down. The industry’s two forecasters publish a weighted average cost of production every two years: $24 a pound in 2019, $27 in 2021, $30 in 2023, $38 in 2025. Kazatomprom is the largest and lowest-cost producer in the world — Kazakhstan is 41% of global supply and the state company is half of that — and its cost today is around $40.
The world’s cheapest producer sits on the industry-wide weighted average. Alkin’s conclusion is that the whole curve has to be reset up and to the right. He adds the currency layer: Kazatomprom’s costs are in tenge and its sales in dollars, so a weak tenge props up the margin; normalise back to the 340 the prospectus assumed instead of the 450–500 of recent years and the cost clears $55 a pound.
The same logic bites harder on development projects. Project after project has missed badly — costs 80–90% above what was promised, some a couple of hundred percent higher, production targets cut by 50–70%. Those misses are the risk an equity holder carries, since the one you own could be next, and they also squeeze an already undersupplied market. So when Townsend asks whether to own the miners or the metal, Alkin says a case exists for either and the difference is risk tolerance: there should be a real scarcity value on companies that can actually mine, and what that value is worth, he says, is art rather than science. He stops there and says he stays away from advice on structuring or hedging.
The reported price is not where the market trades
Alkin says the real takeaway from Cameco’s call got missed. The reported fixed-price contract is $98–99 a pound, and on the call Cameco said plainly that it is getting triple digits.
How do those coexist? Because the industry’s price reporting takes the lowest offer heard on a deal it knows about, not where the contract was struck — so even if that utility awarded the business to Cameco in triple digits, the price gets pegged to whoever bid lower. The reporters’ rationale is that uranium is a fungible commodity. Alkin calls that absurd: 70% of uranium is produced in the east and 70% consumed in the west, the west runs a large deficit, and the small eastern surplus feeds the Russian, Chinese and Indian programmes.
Half the needs contracted, and the term price quadrupled
The deck he cites says utilities are contracting a little over half of their requirements. In a normal year, 75–85% of the 160–200 million pounds consumed is bought under long-term contracts, because uranium has no substitute — you can coal-to-gas switch, you cannot uranium-switch — so security of supply comes first.
Then he lays the contradiction out. Commodities 101 says less demand plus more supply equals lower prices. Buyers have contracted half their needs, the paper supply stack looks able to produce more than that demand, and the uranium term price has quadrupled. It is a clean cut: the market is already telling you something, and you don’t have to take his word for it.
Then the inventory layer. The industry destocked for fifteen years, and destocking has to be reversed: with demand at 200–230 million pounds, actual buying has to run at 250–270 million to refill the coffers. He gets to more than two billion pounds that need contracting by 2040, and the mines that could supply it are neither in production nor in development. 2040 sounds far away, he says; on mining timelines it is two days from now. Another thermometer: utilities issuing RFPs a few years ago might get ten or twelve responses; now very few come back, and the big producers say they aren’t interested.
Nobody pays the fuel buyer for calling the bottom
This is the most human stretch of the episode. In 2017 a commercial officer at a large uranium producer took Alkin to dinner with some big fuel buyers. He had just been through global inventories: the industry line was that 1.4 billion pounds sat out there against annual use of under 200 million, so they would never have to buy again. His own work said only 400–500 million of that was mobile; the rest needed re-enriching, was depleted tails, or was locked in sovereign funds.
He got laughed at, and then asked a question: how are you compensated — do you get paid for calling bottoms? He says they reacted as if he’d asked for their bank balances. He pressed: if your analyst called the bottom and uranium doubled or tripled, do you get a piece of that? They looked at him like he had three eyes, and said no.
That no explains much of what followed. Fuel buyers are mostly nuclear engineers, plenty smart, but the job doesn’t ask them to be commercial the way a gas or oil trader is. The whole fuel cycle is roughly 25% of the cost of power and uranium itself is a mid-single to low-double-digit percentage, against 80–90% for coal and gas at a coal or gas plant. Here’s the asymmetry: nobody gets fired for paying up for fuel, and running out would do it.
He also tells the Boston story from 2018. He got laughed off the stage, and a woman came up and asked how he dared come and talk to them — a Wall Street bad guy scaring hardworking people who buy fuel for a living. He walked her to an alcove and opened his laptop onto thousands of rows of supply-demand work. Twenty minutes later she told him she owed him an apology.
Services moved first, the product moves later
This is the comparison he leaves in the episode. In the fuel cycle, uranium is the product; conversion, enrichment and fabrication are services. Conversion sat at $4 per kilogram for years, unprofitable, a dying part of the cycle nobody watched — then one country put out an RFP for conversion services and nobody bid at all. Conversion is over $60 now. Enrichment went from 30 to over 200, fabrication is up too. Uranium went 4x, and quietly.
His inference is that the product catches up to the services, because the services mean nothing without enough product, so the price has to go and call the product out. Today’s term price is $98–99, with contract ceilings at 150–160 and floors at 70–75, spot at 91–92.
He runs one historical comparison: uranium went from $7 a pound in 2000 to a brief $137 spot in 2007, which in today’s purchasing power is roughly $200. The interesting part is that if you redo the supply-demand work for the five years after that peak, the market was in surplus. Do the same exercise today and you get large deficits. He immediately adds that he doesn’t forecast prices publicly.
If the fundamentals are this good, why is my position still falling?
This is the easiest place to get hurt by an episode like this: you read everything, you feel like you understand it, and the number on the screen keeps going down.
What I took away is to keep two separate books. Book one is the thesis: is the deficit still there, where is the term price, has the contracting rate moved, are projects missing. Book two is the price. The whole episode is about those two books pointing in opposite directions — flows are moving and not one of his four fields has flipped. Keeping them separate gives you something to check against: if price falls and the four fields are unchanged, flows did it; if the term price rolls over and contracting picks up while price still falls, the thesis needs repair.
Writing it down rather than holding it in your head has a second use. Alkin’s own warning is worth borrowing: no thesis creep. The moment someone adds a new reason to explain the share price, that book is void, because it can no longer be proved wrong. The juniors labelling themselves AI plays did exactly that — they swapped in a hotter story, and when the hotter story sold off they went with it.
Everyone says this is a structural bull market — how do I know it isn’t a story?
I’ve asked myself this many times, and the episode gives two usable entry points.
The first is to go find the number that violates common sense. Supply-demand says halved demand should mean lower prices; the uranium term price quadrupled instead. You can check that contradiction without trusting anyone, and it is harder than any narrative. The reverse also holds: a theme that reads smoothly and feels exciting but offers no counter-intuitive number is probably just a well-told story.
The second is to ask what the people making decisions in that market get paid for. Anyone can build Alkin’s supply-demand table — he says the data is all public — but nobody at that table had compensation tied to a price call, so for years nobody fixed the table. Once you know that, you know how much evidentiary weight “everyone in the industry says it’s fine” carries.
One more layer I want to keep. Alkin says he assumes he’s wrong every day and asks people to tell him where — and at these conferences he cannot find a single bear willing to argue. Townsend’s joke is that Alkin had one job and failed. When I can’t find the other side of a case, I try to separate two situations: there is no other side, or the other side has no incentive to talk to me. The second looks like the first and means something very different.
Sources worth reading
- MacroVoices episode 558, 12 November 2026, the source for this piece; the show’s homepage at macrovoices.com carries that week’s downloads
- Alkin’s Nuclear Energy Institute nuclear fuels conference deck, linked in the show’s subscriber research email
- Kazatomprom’s 2018 IPO prospectus and its recent quarterly cost disclosures, for the $13–14 versus $40 comparison
- The World Nuclear Association’s biennial fuel report, which carries the official version of supply-demand and inventories
One thing to take with you
What a market looks like follows from what the people deciding inside it get paid for and get punished for. The question Alkin asked at that 2017 dinner — do you get paid for calling bottoms — came back as a no, and that no explains why, over the years since, some of the smartest nuclear engineers in the world collectively got the price direction backwards. There was no line for it on their scorecard.
Here’s something I’ve tried and would invite you to try: pick one person you’ve recently decided is being unreasonable — a colleague, a call-centre agent, someone at home. Write two lines on paper. First line: what they got praised for this month. Second line: what they got in trouble for this month. Then look again at the decision that annoyed you. By the time I finish the second line my irritation has usually drained away, because on their scorecard that decision was right.
This article is an educational discussion of investment method. It is not advice to buy or sell any individual security, offers no target prices, and does not analyze any current holding. Investing carries risk; make your own decisions or consult a qualified professional.
Comments
Loading comments…