Live with the Oak Bloke — special situations, blown-out countries, and how he expresses uranium

Georgia Capital is a holding company — a business whose business is owning other businesses — in the Republic of Georgia, the small ex-Soviet country in the Caucasus, not the US state. It owns stakes in local companies (a bank, healthcare, water, retail) and its job is to grow the combined value of those stakes and then hand cash back to shareholders.

Polomny's case has three legs. The country reformed itself two decades ago and money started arriving, so the pond itself is getting bigger. Management set a public target — ten-bag the company in ten years — and published the plan to get there. And when the stock market refuses to pay full price for the assets, management uses spare cash to buy back its own shares, which quietly increases every remaining holder's slice. That last habit is what he means by "Walter Schloss methodology," after the old-school value investor.

The part worth copying is not the stock, it's the patience. It took five years, and it handed him "two or three bites at the apple" — chances to buy more at lower prices when something scary happened. The scariest was the Ukraine invasion, which knocked the shares down 40% on the assumption that anywhere near Russia was dangerous. He argued the opposite: money would flee Russia and Ukraine into Georgia, which is what happened.

This is the bank inside the Georgia Capital story, now renamed Lion Finance Group after buying Armenia's largest bank. It is listed in London in its own right.

The single number that made it obvious, in the host's telling, is return on tangible equity in the thirties — meaning the bank earns roughly thirty pence a year on every pound of real shareholder money it holds. Ordinary Western banks earn about a third of that. When a bank sustains a number like that, it doubles its own capital base every few years without asking anyone for money.

Polomny's live interest is what Georgia Capital does with its stake rather than the bank itself: it has been selling shares down and pushing the proceeds back to its own shareholders. That is a deliberate wind-down of a concentrated position into cash returns — a sign, in his reading, of allocators who know what they own.

Uzbekistan is a 37-million-person Central Asian country that spent the post-Soviet decades closed and badly run — the state used to conscript citizens to pick cotton by hand. A change of leadership opened it up, and it was the only country in the world that grew through COVID. UZNF is a London-listed fund that buys Uzbek companies, so a Western investor can own the economy without opening a brokerage account in Tashkent.

Polomny's argument is deliberately boring: he does not need to pick winners or time anything. If an economy compounds at 7–8% a year for twenty years while its stock market is still tiny and ignored, the whole pond gets far bigger and any competently run fishing operation catches more fish. "All they have to do is just keep doing what they're doing."

The risks are the frontier-market ones and he does not dress them up: one leader's reform programme can be reversed, liquidity is thin, and the fund only listed recently, so there is very little track record to judge.

Aimia is a Canadian-listed shell of a former loyalty-points company that now exists to buy other businesses. It has roughly $280 million of cash and, crucially, more than a billion dollars of accumulated tax losses — meaning any profits it earns from here can be shielded from tax for a very long time. That makes cash inside Aimia worth more than the same cash inside a normal buyer.

Polomny is not buying the assets; he is buying the person. "I like jockeys" — the horse-racing metaphor for backing the manager rather than the business. The jockey is Rhys Simmerton, who arrived after a previous management "bought a bunch of companies they shouldn't have," sold the mistakes, and now wants to buy cash-rich UK companies at what he calls generational-low valuations, use their cash flow to buy the next one, and repeat.

Notice how this squares a circle he set up earlier: he thinks the UK is "tremendously cheap but uninvestable right now" because of policy. Owning a Canadian vehicle run by someone who will buy UK cash flows gets him the cheapness without owning the country. The obvious risk is that everything depends on one person continuing to allocate well — jockey bets fail when the jockey leaves or loses his touch.

TechnipFMC builds the equipment that sits on the seabed at offshore oil wells — including blowout preventers, the safety hardware whose failure caused Deepwater Horizon. Only a handful of companies on earth can engineer this, which is what he means by "oligopoly."

This entry exists to record an exit. Polomny mentions almost in passing that he "just sold out of portfolio" — an eight-bagger, roughly eight bucks to eighty, over five years. He is not saying anything negative about the company; the trade simply reached its objective and the capital is being redeployed.

What made it a special situation in the first place is worth keeping: management got distracted by an offshore-wind business, spun it off, and refocused on what they were uniquely good at — right as offshore oil spending was about to restart. A cheap price plus a fixable distraction plus a cycle turning is the shape he hunts for. And, as with Georgia Capital, the path was not a straight line: bad headlines about offshore or a drop in the oil price gave him "several bites at the apple" along the way.

This is a fund that does one thing: it buys physical uranium and stores it. It does not mine, drill, or operate anything, so it cannot suffer a cost blowout, a flooded shaft or an incompetent CEO. Owning it is close to owning the raw material itself.

Because it trades on a stock exchange, its share price can drift away from the value of the uranium in the vault. When the shares trade below that value — a "discount to net asset value," recently as much as 13% — you are buying a dollar of uranium for 87 cents. Polomny's rule is mechanical: when the discount goes deep, buy; then hold.

The reason he insists on this rather than mining shares is that the mining companies have stopped being free options. In the early years they were pure hope with nothing to disprove; now they have to actually produce, and many are producing badly. His condition for changing his mind is precise and public: when a Rio Tinto or a BHP finally commits billions to build a large new mine, the supply deficit starts closing and the trade changes.

Uranium out of the ground cannot be put in a reactor. It has to be converted into a gas (uranium hexafluoride), then enriched, then made into fuel. Conversion is the step almost nobody can do — there is essentially one operating commercial plant in the United States, and Solstice, spun out of Honeywell, owns it.

Polomny's interest is structural rather than speculative. When a supply chain has one narrow gate, the owner of the gate collects the toll no matter which miner eventually wins. He heard on recent earnings calls that the company intends to spend money widening that gate, which both relieves an industry bottleneck and grows the toll booth.

Two honest caveats. He places the plant in Kentucky; the US conversion works is actually in Metropolis, Illinois — a slip that does not change the argument. And conversion is only part of Solstice: the company also sells refrigerants, electronic materials and packaging, so a uranium buyer here is buying a diversified chemicals business with a strategic asset inside it, not a pure-play.

LandBridge owns land — ranch acreage in West Texas — and the rights that come with it, especially water. It does not drill; it charges other people for using its surface, its water and its infrastructure.

The reason that matters now is that AI data centres have become politically radioactive. They consume enormous amounts of electricity and water, and voters in Texas and Pennsylvania have noticed their power bills. Some jurisdictions have blocked them outright, and the President has said new ones should bring their own power rather than draw from the grid. So the scarce thing is no longer computing hardware; it is a place you are allowed to build.

LandBridge happens to own exactly that combination: empty land in counties with a few hundred residents, legal rights to underground water plus an existing business recycling the water oil wells produce, and access to natural gas that is so abundant locally that producers sometimes pay to have it taken away. Being a landlord rather than an operator means it collects fees whether or not any individual data centre is a good business. The risk is concentration — one region, one boom, and a valuation that already assumes a lot of the deals get signed.

Texas Pacific Land is the older, larger version of the same idea, and its existence is a historical accident: a railroad went bust in the 19th century and its creditors were left holding a vast quantity of West Texas dirt nobody wanted. Then oil was found under it. Today TPL collects royalties on production from that land and sells water services, with very few employees and almost no capital spending.

Polomny credits the insight to Murray Stahl of Horizon Kinetics, whom he calls his hero and the smartest person he has read. (He speaks of Stahl in the past tense; that is recorded here as his statement rather than as established fact.) He also points people at Horizon Kinetics' free published research, which works through the physical arithmetic — how much water a megawatt of data centre actually needs.

The general lesson he is drawing is about where to stand in a boom. Rather than pick which AI company wins, own the non-reproducible thing every winner needs. Nobody can manufacture more West Texas land with water rights under it.

Holtec is private, so you cannot buy it today; it has filed to go public and Polomny says he is looking forward to that.

Most coverage of Holtec is about restarting the mothballed Palisades nuclear plant in Michigan with government support — a headline-friendly but risky project. His point is that he read the company's IPO filing and found something more durable underneath: Holtec makes and manages the dry casks that hold spent nuclear fuel on power-station sites, and controls roughly 80% of that market.

That is a much better business than it sounds. Every reactor produces waste, the waste has to be stored somewhere for decades, nobody else is licensed and trusted to do it at scale, and the customer cannot walk away. It is a monopoly-ish service annuity attached to an industry that is growing again. The lesson he is illustrating — read the filing rather than the press release — is transferable to any IPO.

Cameco is the largest Western uranium miner — the safe, liquid, institution-friendly way to own the theme. Polomny's objection is simply price: "it's always very expensive." When a stock is permanently the consensus choice, the discount he needs never appears.

He does, though, give Cameco the strongest hand in the industry. It knows where every good deposit is, so it can wait for struggling developers to run out of money and then buy their assets cheaply — "let these guys bleed out and then we can swoop in." Together with Kazakhstan's state producer, he thinks the two of them are effectively managing how much uranium reaches the market.

That is a bullish argument for the price of uranium and a neutral one for the shares. It is also the logic behind buying the metal instead: if the two dominant producers are restraining supply, you want to own the thing being restrained, not the companies doing the restraining at 30 times earnings.

Kazatomprom is Kazakhstan's state-controlled uranium producer and the largest in the world. For years the market feared it would flood the market and crush the price. It keeps not doing that: production targets get cut, and the sulphuric acid plants it needs (acid is used to dissolve uranium out of the ground in situ) keep being delayed — Polomny relays that, according to Cameco people who visited, ground had not even been broken on one.

His reading is that this is not incompetence but policy: "managed decline." His paraphrase of the government's logic is the whole argument in one sentence — why sell a finite national resource cheaply when you believe a multi-decade bull market is coming?

That makes him bullish on uranium and neutral on the shares. Buying the equity means buying a state-directed company in a landlocked country that sells much of its output eastward, with all the political risk that carries. Buying the metal captures the same restricted supply without the ownership question.

Paladin runs the Langer Heinrich mine in Namibia and is one of the few juniors that has actually restarted production. Polomny's complaint is specific and worth understanding, because it is the trap in every commodity bull market: the price of the commodity is not the price the company receives.

Uranium was trading around $88 in the spot market and $97 on long-term contracts, but Paladin's realised price — what it actually got paid, after legacy contracts signed in leaner years — was $57. "Guys, what did you do?" A miner locked into old low-price contracts gives you the operational risk of mining with only a fraction of the upside you were trying to buy.

He is not saying the business is broken; he acknowledges they finally "got their act together" operationally. He is saying it is the wrong instrument. It is also a useful check for any resource stock: before buying it, find the realised price in the accounts and compare it to the headline commodity price.

This is the host's holding, not Polomny's. ASA International is a London-listed microfinance lender: it makes very small loans — around $200 — mostly to women running tiny businesses in South Asia and parts of Africa such as Ghana, Kenya and Pakistan.

The economics are unusual. Interest margins are enormous by Western standards (thirty or forty percent), because there is no alternative source of credit for these borrowers, while losses stay in the low single digits because the loans are tiny, short and socially enforced within borrower groups. The result is a return on tangible equity above 30% — comparable to Bank of Georgia's.

The host's explanation for why it trades cheaply anyway is investor reflex: "oh, Africa, hyperinflation risk — just not going to touch it." That is the same home-country bias Polomny then discusses in general terms. He does not endorse or dispute the specific company; be clear that this row records the host's view, and that lending at very high rates to poor borrowers carries political and reputational risk as well as currency risk.

Also the host's, not Polomny's. RTW is a London-listed investment company that owns a portfolio of biotech assets, run out of New York by a team the host describes as roughly thirteen PhDs and MDs who are also commercially minded.

The logic is delegation. Polomny had just explained why he avoids single-name biotech: judging whether a drug will work is a full-time speciality, and "you can't kiss all the girls." Buying a listed vehicle staffed by specialists is one answer to that — you are hiring the expertise rather than pretending to have it. The value shows up as net asset value stepping up whenever a holding is acquired or hits a milestone.

Polomny's own answer to the same problem is different and cheaper: buy a biotech ETF for the cycle exposure, then read the regulatory filings of the best biotech funds to see which names they all hold. Two routes to the same admission — that this is a sector where you should rent judgement rather than fake it.

The host's third biotech position, and he frames it as a special situation rather than a drug bet. PureTech creates biotech companies, spins them out, and keeps stakes and royalties — most famously Karuna, whose schizophrenia drug was sold to Bristol Myers Squibb, leaving PureTech with ongoing royalty income.

The cheapness argument is a NAV discount: you can buy it, he says, for roughly half the value of what it owns. So the cash, the royalties and the stakes are available at fifty cents on the dollar, and the drug candidates still in trials come free.

His way of thinking about the pipeline is a stable of racehorses: several are at phase two and phase three, each addressing markets he sizes in the ten-billion-a-year range, and you only need one to win. That is a fair description of how holding-company biotech works — and also of the risk, since trials fail more often than they succeed and a discount to NAV can persist for years. This is the host's view; Polomny offers none.

The host's example, used to demonstrate a tool rather than to make a recommendation. Thungela is a South African thermal-coal miner, spun out of Anglo American.

The tool is Michael Porter's five forces, a 1970s Harvard framework for asking how attractive an industry structurally is, regardless of this year's price: how much power do your customers have, how much power do your suppliers have, how easily can new competitors enter, how easily can customers switch to something else, and how brutal is the rivalry among existing players.

Applied to coal it produces a genuinely mixed answer, which is the point. New entrants: almost impossible, because no bank will finance a new coal mine — good for incumbents. Substitution: very real, from renewables, gas and nuclear — bad. Supplier and stakeholder power: bad, because everyone from the government to the state railway that hauls the coal to the export terminal wants a cut. Running the checklist forces you to state the weaknesses out loud instead of falling for a cheap earnings multiple.