Actionable insights — No ETF, No Futures Market: Why 17% Of His Money Is In This Metal

20:38 1. Hub and spoke — let two questions decide every position size

1. Split the book into a hub and spokes before picking any name. The hub is a small number of large, boring, diversified core holdings; the spokes are the individual bets arranged around it.
2. Build the hub from instruments that cannot go to zero on a single management decision — for a metals book, the physical metal plus a broad producer ETF and a junior ETF.
3. Define the spoke buckets explicitly rather than letting them accumulate. His are three: critical minerals, gold equities, silver equities.
4. Decide hub-versus-spoke with two questions, in this order: (a) how much conviction do I have in the name, and (b) do I have any relationship with the company at all? No relationship caps the position at spoke size regardless of how cheap it looks.
5. Reject equal weighting as a default. His own partner runs ~180 names roughly equal weighted; he calls that a legitimate strategy and declines it — "I don't like equal weighting. I like taking bets."
6. Put a floor under the whole thing before sizing anything: no margin, nothing you cannot afford to lose, nothing that touches retirement savings. "You can't live on margin and cross your fingers. That's not a strategy."

• Any single spoke drifting toward hub weight without the conviction and relationship test having been passed; a hub that has quietly become a collection of individual stock risks.

13:29 2. The four-part junior checklist — people, project, share structure, jurisdiction

1. Management first. Not the deposit, not the metal — the people. "The project can be great. If you have a terrible person running it, it's not going to work." He concedes this ordering is unusual and defends it anyway.
2. Project second. Grade, metal, stage, mine life — assessed only once the people have passed.
3. Share structure and jurisdiction, tied third. Neither is a veto on its own; both are disqualifiers when bad.
4. For share structure, do not screen on the raw count. A large count is acceptable if you know "where the bodies are buried" — see insight 4.
5. For jurisdiction, work top-down from the politics: identify countries whose government has just turned more mining-friendly, then ask "if I like Colombia, who do I like in Colombia?" and research into it. He flags Peru (new, more conservative, more pro-mining elections) and Colombia (a new regime a month old) as the current examples.
6. Accept the cost: "it's a lot of work" — and he runs the full four every single time he makes an investment.

• Elections and permitting-regime changes in mining jurisdictions; a management team you rate arriving at a company you had discarded on the project.

13:57 3. The two-strike honesty rule — and the permanent blacklist that follows

1. Treat honesty as the single screening variable that cannot be modelled or diversified away: "you can't guard against lying. When someone lies to you, they lie to you and then you have to make an adjustment."
2. Give one strike. A first misstatement can be error, optimism or bad timing.
3. On the second, exit permanently — no re-underwriting, no "the new project is different."
4. Extend the ban to the person, not the company. He never again owned anything that CEO, that investor-relations person or that board member did, at any other company.
5. Write the rule down in advance and be binary about it. "I'm cut and dried. I'm not a gray guy. It's either black or white with me. And that's how I'm successful."
6. The retail-investor version: you cannot phone the CEO, but you can keep a dated file of what management said would happen and check it against what did — guidance, timelines, financing promises, resource targets. Two clean misses on things they controlled is the same signal.

• Guidance quietly restated rather than withdrawn; a timeline that slips twice with a new explanation each time; the same executives and directors reappearing at a new shell with a new story.

14:39 4. Interrogate the share count — "where are the bodies buried?"

1. Do not reject a company on share count alone. "If it does have a large share count, that's okay. But I want to know where the bodies are buried in that share count."
2. Ask what proportion is locked up and by whom: "if you have 400 million shares out, I need to do a phone call with you as a CEO to understand: do you have 30% of this company locked up so that we're not going to see it taken under as an investor."
3. Check for institutional anchors by name and percentage — "Does Franklin Templeton have 3%? Does VanEck have 5%?" Their presence is evidence the company will not "do something weird."
4. Reframe a big register as a positive when the holders are structurally non-sellers: 250 million shares is a lot until roughly half of it sits with people who will never sell at this price.
5. Use it as a pre-catalyst risk test, not a valuation input: the question is who dumps into good news, and who does not sell into bad news.
6. Retail version: the same information is in the filings and the register disclosures — insider ownership percentage, above-5% holders, warrant and option overhang, and the last financing's terms.

• Warrant and option overhang relative to shares outstanding; lock-up expiries; a financing priced below market with a full warrant attached; an institutional holder cutting to zero.

10:50 5. The leverage diagnostic — miners should move 2–3× the metal, and if not, ask why

1. Set the expectation numerically before you own anything: "if you buy gold at 4,000 an ounce and it goes to 4,400, you're going to make 10%. But the miners during that period should go up two or 3x to that move in gold."
2. Measure your own holdings against that on up days, not down days. Under-performance on a rally is diagnostic; under-performance in a general liquidation is not.
3. When a name lags the metal on a green tape, ask the specific question rather than a vague one: is the company hedging production? A hedge locks in a selling price, which protects the downside and caps exactly the upside you bought the stock for.
4. Do not treat hedging as fraud — "airlines hedge, it's a normal process" — treat it as a mismatch with the reason you own the stock. He wants "companies that are going for it, that are unhedged."
5. Rule out the alternatives before concluding: dilution since your entry, a jurisdiction event, a cost problem, or a permitting delay will all produce the same lag.
6. Use the sector-wide version as a bull-market confirmation: juniors rising ~3× gold was his evidence through August that the framework was intact.

• Hedge books and forward-sales disclosure in the quarterly reports; a producer whose realised price is persistently below spot; a name that stops participating in sector rallies for two or three of them in a row.

12:19 6. Buy artificial dislocations — the price fell for a non-fundamental reason

1. Separate two causes of a large drop: something changed at the company, or something changed in the plumbing. Only the second is an opportunity.
2. Screen the plumbing events specifically — index additions and deletions, forced fund selling, merger-driven reconstitution, tax-loss selling, a delisting or listing-venue change.
3. Verify the fundamentals are literally unchanged over the window. In an index deletion, the ounces in the ground on the day after are identical to the day before.
4. Do the work to confirm the mechanism rather than assuming it — he brought the CEO onto his own show to have him explain what had happened. The retail equivalent is the index provider's own reconstitution notices and the company's press releases.
5. Only buy the dislocation inside a bull framework. He is explicit that the rule is conditional: "you look for these artificial things happening in the market within a bull framework and just say, 'Hey, I'm going to buy these dips.'"
6. Size it as a real position, not a token. This is one of the two names he holds at over 2%.

• Russell reconstitution and index-deletion notices; mergers that change a listing's index eligibility; a stock down 50% on volume with no operational news; December tax-loss selling in beaten-up juniors.

24:06 7. Trade the milestone calendar — MRE, PEA, PFS

1. Know the sequence a development company walks through, because each step converts a story into a number: MRE (mineral resource estimate — how much metal is actually there), PEA (preliminary economic assessment — a first pass at whether it can be mined profitably), PFS (pre-feasibility study — the engineered version with capex and payback).
2. Maintain a dated calendar of upcoming milestones across the names you follow. "I look for catalysts like MREs, PEAs, PFSs. These are milestones for a company. They're really important to look for."
3. Position ahead of an announced milestone in names where the register will not sell into a disappointment (insight 4) — that is what converts a binary event into an asymmetric one.
4. Read the market's reaction as separate from the result. A strong MRE that spikes and fully retraces on unrelated macro news is a fact that has not been paid for, not a rejected result.
5. On a PFS, go straight to payback period and capex, not the headline NPV: "outstanding numbers, payback in less than a year" is the sentence he leads with.
6. Check volume around the event — a milestone that draws massive volume has produced the liquidity that lets a larger holder build.

• Company news releases announcing a study is "in progress" or "expected in Q_"; the gap between a milestone landing and the share price giving it back; capex and payback figures in the study rather than the press-release headline.

33:48 8. "No ETF, no futures market" — the metals where the equity is the only vehicle

1. Screen commodities by how they can be owned, not just by their supply-and-demand story. Gold, silver and copper have futures, ETFs and physical markets; tungsten, antimony and rhenium have none.
2. Recognise what that does to the flow. When generalist money decides it wants exposure and there is no fund and no contract, the only available buy is a handful of small equities — a very large demand curve meeting a very small float.
3. Verify the physical tightness independently of the price move, because the price move alone will always look like it has gone too far. He answers the "$920 to $2,800 is a big move" objection with a sourced deficit rather than a chart.
4. Source it from people who have to know: three 2026 Zoom calls with the US government's SAFE division and four staff, all giving the same answer — a year-and-a-half to two-year deficit.
5. Test the irreplaceability. Tungsten is used in tanks and Tomahawk missiles and heavily in technology; there is no substitute. A metal with a substitute has a ceiling; one without does not.
6. Then size it as a theme rather than a stock: 17%+ of the portfolio spread across producers, developers and explorers rather than concentrated in one.

• The first ETF or futures contract launching on a metal you already own the equities of (the flow arrives, and the equity premium may compress); US permitting reform and refinery/smelter announcements; offtake agreements being signed rather than discussed.

34:46 9. Follow the money, not the mine — funding and offtakes as the real de-risking

1. For any pre-production company, ask the two questions that actually decide the outcome: is the construction funded, and is the output sold? Geology rarely kills a junior; financing does.
2. Compare committed capital against expected capex as a ratio. Committed money exceeding capex means dilution risk is off the table — a rarity worth paying for.
3. Weight the identity of the funder. A development bank or a sovereign entity is a different signal from a retail placing: it implies a strategic buyer of the output as well as a source of money.
4. Track offtake agreements — contracts to buy future production — as a proxy for demand certainty. "It's tying up future production. It gives you as an investor much more comfort that there is a buyer on the other end."
5. Distinguish a conversation from a contract, out loud, the way he does: "those are considered to be like offtake conversations. They're not definitive. Nothing's been done yet."
6. Use funding as the sanity check on a hot sector: "show me a gold stock that has $1.6 billion lined up right now… You're not seeing this in other parts of the mining area, but you are seeing it in critical."

• EXIM / development-bank and sovereign funding announcements in critical minerals; offtakes converting from MoU to binding; capex estimates rising toward the committed funding line.

21:22 10. The relationship sell rule — up 100% with no relationship, sell half or all

1. Classify every holding by whether you have real access to management. He owns cheap names in tungsten, antimony and rhenium where "I don't have a relationship with the company, but the stock looks so cheap I can't help myself."
2. Attach a pre-committed exit to the no-relationship bucket: up 100%, sell half or all. The rule is set before the gain exists, so it is not a judgement call in the moment.
3. State the actual reason, which is information rather than valuation: "I don't know what's next for the company… are they going to do a financing out of left field? I don't want to get caught like that."
4. Note the asymmetry a surprise financing creates for a small holder: dilution at a discount, usually announced after the stock has run — precisely when you are most exposed.
5. For names where you do have the relationship — "we've talked to these CEOs 30 or 40 times" — the rule does not apply, and those are the ones that earn hub-size weight.
6. Build the relationship deliberately if you want the bigger position: he cold-called the Stillwater CEO with no prior connection, and it became both a friendship and a multi-year holding. Retail version: attend the calls, ask questions, read every release, and demand more evidence in lieu of access.

• A no-relationship name doubling; a company that has not raised in over a year with cash covering less than four quarters; shelf registrations and at-the-market programmes being filed.

22:04 11. Triage a losing position — was it the thesis or was it you?

1. Ask the honest first question, which is about your own entry rather than the company: "was that me getting caught up in a euphoric moment… and then the rug got pulled because of the war?"
2. If the answer is yes, the position was never underwritten — "take a loss, or reduce your position." Acknowledging that the entry was emotional is the whole test.
3. If the entry was sound, switch to the catalyst question: is the company still going into production, still getting its permits, still delivering "some positive catalyst"? If yes, "then you have to stick with it."
4. Separate the two causes of the drawdown explicitly. He attributes this summer's selling to the Fed and the war — external, temporary, and not information about any individual deposit.
5. Add on dips only in names where you already had conviction, never to rescue an average: the discipline is "we start a position in something we have conviction in and then we add to it on dips", which presumes the conviction preceded the loss.
6. Fold in the tax calendar before acting — hold past a year where possible so the outcome is a long-term gain or a long-term loss, with an advisor.

• Whether the catalyst schedule has slipped or only the price has fallen; cash runway at the current burn; whether you can state the original entry thesis without reference to the price.

4:09 12. State a macro call with its falsifying condition attached

1. When you make a directional call, name in advance the one thing that would break it. Three weeks earlier his bottom call carried exactly one condition: Trump doubling down on Iran before November.
2. When that condition appears to trigger, re-examine the reasoning rather than reflexively honouring the rule — the US struck Iran twice in three days and he still concluded the timing had changed, not the thesis.
3. Anchor the re-examination on a constraint rather than an intention. His is electoral: Republicans need a better midterm result, so sustained escalation is against the actor's own interest until November.
4. Pair the call with price levels that make it checkable — gold holding 3,900 (a double test would not surprise him; a break would change things), silver testing but not breaking 50.
5. Read the central bank the same way: not by its stated first priority but by the full list. "Warsh at the Fed is not just worried about inflation… number two on the list is the labor market. And the labor market is crap right now."
6. Convert that into a falsifiable consequence: with payrolls at −23,000 against a +80,000 estimate, "I don't think they have the ability to raise rates at the pace that they're talking about." The next print is the test.

• The named levels (gold 3,900, silver 50) and whether they hold on a retest; non-farm payrolls and ADP prints; Fed projections at the September meeting; any escalation after the November vote, which he expects to be a separate regime.