Rick Rule — These Commodities Are Mind-Bogglingly Underpriced
"How much of this AI value enhancement is priced into these companies right now? Zero. They're selling at a discount to the net present value of their ore bodies at today's prices and an 8% discount — so you get the back years, the exploration upside, and the AI-efficiency warrant for free."
One-line take: A macro/thematic episode framing the AI build-out as a physical-resource problem, not a digital one. Rule's core argument: if you believe the AI build-out, you must be bullish commodities, because the physical capacity (power, water, copper) to build all the slated data centers doesn't exist — so the most-bullish AI case can't happen on schedule, and the resources will be rationed by price. Three "sins" compound into a nominal price floor: 30 years of underinvestment (the sins of the past), future AI/electrification/development demand, and a US dollar in "inexorable decline" (it fell 75% in real terms in the 1970s). Copper is the must-own big market — Robert Friedland's claim that the world needs more copper 2026–2050 than in all recorded history, plus a 16–17-year supply-response lag, makes rationing-by-price unavoidable short of a synchronized global depression. Uranium is his "surprise winner" — 24/7 non-carbon power for AI, and the energy-density/energy-security thesis (the 1973 Arab oil embargo birthed the French & Japanese fleets); probability uranium is "the fuel of the AI business" = 100%. Oil: a war-driven $55→$95–100 spike likely falls back toward $60–65 if the Strait reopens, but ~$1bn/day of deferred sustaining capital makes the structural shortage worse, not better — only Exxon is reinvesting; the rest cannibalize via buybacks. The cleanest idea is the NPV "free warrant": resource companies trade below the NPV of proven reserves at an 8% discount (cash flow past year 11–12 is worth zero), so you get the long-tail reserve years, the commodity-price upside, and the coming AI-efficiency gains all for free. On gold he's a price-insensitive saver; he reiterates selling silver at $75 at the hockey-stick top ("when I see a hyperbolic chart, up or down, I bet against it"). Timestamps link into the video.
1. Stocks & names mentioned
| Ticker | Name | Research | View | What he said | At |
|---|
| COPX | Global X Copper Miners ETF | SA · STK | Positive | The big market you must own for the AI/demographics build-out — "copper, which is a great big market, will do extraordinarily well." 30 years of underinvestment + a 16–17-year supply-response lag mean nothing can alter supply in a reasonable timeframe; demand can only be cut by a depression. Friedland: more copper needed 2026–2050 than in all recorded history. Rationing-by-price is locked in. | 39:47 |
| URA | Global X Uranium ETF | SA · STK | Positive | His "surprise winner." AI needs prodigious 24/7, non-carbon power — uranium delivers both; the Strait-of-Hormuz conflict revives energy security after a 50-year absence and uranium is its "unsung beneficiary" (the 1973 oil embargo built the French & Japanese fleets). "The probability that uranium is the fuel of the AI business is 100%." Manifests over 5–10 years, not 2026, absent a Chernobyl/Fukushima. | 40:20 |
| CCJ | Cameco | QT · SA · STK · FA | Positive | "The company that seems to be progressing the fastest in terms of nuclear technology is Canadian — it's Cameco, having bought Westinghouse." Named as the best-positioned name on the nuclear-reactor-build side of the uranium/AI-power thesis. | 46:31 |
| XOM | Exxon Mobil | QT · SA · STK · FA | Positive | The lone disciplined allocator: "it's not occurring except at places like Exxon" — i.e. cutting buybacks/dividends to reinvest free cash flow into future production (3–5 years out) instead of cannibalizing the company to subsidize shareholders today. Also cited (with Shell) as the furthest along in applying AI to its drilling/production data (10,000 West Texas well logs). | 1:01:06 |
| GLD | SPDR Gold Shares | SA · STK | Positive | "I'm a saver in gold" — fairly price-insensitive; he saves systematically whenever he has a liquidity event (will put conference proceeds into gold). Near-term he won't call the low: higher nominal rates lift the dollar and gold is dollar-denominated, so gold can falter first — but eventually, like end-1975, the political class loses its nerve, sacrifices the dollar to domestic politics, and gold benefits. "Will it occur? Absolutely." | 55:25 |
| SHEL | Shell | QT · SA · STK · FA | Neutral | Illustrative, not rated — named with Exxon as the two oil majors he knows have advanced AI far enough to mine their own well-log/seismic/completion data for correlations a human geophysicist could never compute, an example of the AI-efficiency "free warrant" being largest in oil & gas where the data is richest. | 31:02 |
| SLV | iShares Silver Trust | SA · STK | Neutral | Referenced as the call he got right — he sold his silver at $75 an ounce into the hyperbolic top, taking "slings and arrows" for it (alongside Quartermain at VRIC). "When I see a hyperbolic chart, up or down, I bet against it." Not re-rated here. | 52:33 |
| TerraPower | TerraPower (private — Bill Gates SMR) | — | Neutral | Illustrative — the US is building a commercial-scale SMR right now at Kemmerer, Wyoming ("Bill Gates is doing it"), ironically on top of an old coal mine; evidence that SMR build-at-scale is real and will drive down unit costs over the next 7–8 years. | 44:43 |
| China General Nuclear | China General Nuclear (CGN — private/state, China) | — | Neutral | Illustrative — he knows CGN well; "they build like Ford used to build — you can have any color you want as long as it's black." The standardized, repeatable model that gets a plant built in China for ~$5bn in ~3 years vs ~$20bn / 26–27 years in the US; the cost-of-scale lesson the US is finally absorbing. | 46:02 |
"View" is Rick Rule's stance in this conversation (Positive / Neutral / Negative), not a price rating. This was a macro/thematic appearance — the AI-vs-physical-resources collision, copper, uranium, oil supply, gold/USD and the NPV "free warrant" framework are carried in the talking points and the macro view, not as a basket of rated single names. The metals/uranium are tracked via their ETF proxies (COPX / URA / GLD / SLV); CCJ and XOM are the two single names he named with a clear positive stance. Shell is an illustrative AI-data example, and TerraPower (Bill Gates' Kemmerer SMR) and CGN are private/illustrative nuclear references. Westinghouse (owned by Cameco/Brookfield), PDVSA/PEMEX (Venezuela), titanium/vanadium and fusion were named only in passing and are kept in the talking points. Research: QT Qualtrim · SA Seeking Alpha · STK Stock Analysis.
2. Talking points
0:00 The hook — AI value priced into resource companies is "zero"
- Cold open: asked how much of the AI-driven efficiency/value enhancement is priced into resource companies right now, Rule answers "Zero." Sets the thesis that the AI overlay is a free option on top of already-cheap natural-resource stocks.
5:39 If you believe AI, you must be bullish commodities — the physical capacity isn't there
- Responding to a Jesse Felder point: we don't have the physical capacity (power, water, critical materials — particularly copper) to build all the slated data centers. So the most-bullish AI case "will not happen," or happens 20–30 years late after the physical capacity is built. Either way, overwhelming near-term commodity demand.
6:56 Friedland's copper claim + the 16–17-year supply lag
- Robert Friedland: the world needs more copper mined 2026–2050 than in all recorded history — and that's before data centers. Start exploring today and first success is statistically ~10 years out, +3 to drill it off, +3 to permit (forever in California): you don't change copper supply for 16–17 years. So demand can't be met at a price the market will pay.
9:45 The three sins — past underinvestment, future demand, and a falling dollar
- Beyond the "sins of the past" (underinvestment) and the demand of the future (growth + lifting the poor + AI), a third sin: commodities are priced nominally in US dollars, and the dollar fell ~75% in real terms over the 1970s. A debasing unit of measurement alone pushes nominal commodity prices up even in a static demand environment — a "super boom / ultra boom."
11:29 Markets cure the sin — but over 15–20 years; nothing fixes copper in 5
- The 1970s oil boom caused the 1980s oil bust — high prices destroy demand and incent supply, and when the curves cross the move is precipitous. But the cure is slow: "nothing whatsoever… short of a synchronized global depression… will cause the copper price not to have to rise over five years." Supply will fall; you needed to invest 15 years ago.
13:06 Oil — the late-2025 "ration by price" call, accelerated by war
- He and Taggart argued end-2025 that ~$1bn/day of deferred oil & gas sustaining capital would force rationing-by-price by 2029–2030. He didn't forecast the war; the shortage arrived sooner and for different reasons — the $55→$95–100 move is rationing by price. If the Strait reopens, high prices destroy demand (India, Sri Lanka) and price could fall to ~$65 — but the systemic deferral problem is now worse (Iran/Saudi/UAE/Russia all not reinvesting, plus war damage to rebuild).
16:08 The limiter on AI is physical (and human) resources, not dollars
- The market prices as if resources show up alongside the dollars; hyperscalers say $800bn this year, $1.1tn next. If the resources (and skilled technicians/engineers) don't show up on that scale, earnings forecasts come down, and because the AI complex is "the tail that wags the dog," the broad market comes down with it.
17:34 "I'm a rockhound, not a student of the market" — no view on the correction's size
- He won't predict the scope of any AI-driven correction — the market is "a facility for buying and selling fractional ownership of businesses," not his subject. Honesty over a manufactured opinion.
19:34 A liquidity squeeze hits everything first — even gold stocks
- In a sell-off "margin clerks, not speculators" make the sell decision, and they sell whatever has a bid — gold famously has a bid, so it gets sold (1987: gold stocks held up for 24 hours). Holders who expect resource/gold stocks to save them in a collapse will be "thunderstruck" by the immediate aftermath.
20:55 Two paths after a crash — NPV postponement vs. QE counterfeiting
- If a market collapse presages a real economic collapse, demand for resources is postponed (not destroyed) — and a 3–5-year postponement at an 8% discount "is difficult." But the other path is the government injecting liquidity — lower rates, easier credit, QE ("something you'd describe as counterfeiting") — which is "artificially but extremely bullish for natural resources."
24:54 AI transforms mining & oil — but only with the right questions
- AI's impact on exploration is "enormous and happening today" — it assimilates and synthesizes huge databases for coincident anomalies (alteration, structure, geochemistry). But "if you don't know the questions to ask, AI stands for artificial ignorance." He uses Claude to do in two minutes the six-week comparative-financials work he did by hand as a young analyst; jokes Albert Lu said AI could run "the Rule Classroom without Rule."
29:41 The AI overlay is an unpriced "warrant" on mining companies
- As exploration/extraction get more efficient, the value of already-owned deposits rises automatically — "possession is everything." That future AI efficiency "isn't really priced into the stocks" — you get it for free — yet another reason resource companies trade at discounts to true value.
31:02 Oil & gas is furthest along — Exxon and Shell mining 10,000 well logs
- The free warrant is biggest in oil & gas where the data is richest: Exxon and Shell ("the only two I know about") can throw 10,000 West Texas well logs at AI and ask what correlations between logs, completions, seismic and production they're missing. AI teaches itself to answer questions humans aren't yet smart enough to ask — and "doesn't get tired."
32:35 The free-warrant framework — NPV at 8%, the back years for free
- Resource companies are valued on NPV of proven reserves at current/forecast prices and an 8% discount — at which anything past year 11–12 has zero present value. So a 30-year reserve life means "you're getting the last 18 years for free"; rerun it in 5 years and you get the same answer having banked 5 years of cash flow for nothing. You get exploration upside, commodity-price upside, the time value of money — and now AI efficiency — all for free.
36:56 "A lot of my speculative success has been free warrants" — but be patient
- The free-warrant stack (reserve-life + outyear commodity price + AI efficiency) is real but won't manifest in 2026; collecting a warrant "isn't measured in months." Position before the world wakes up and the repricing happens.
38:55 Pick a big market — copper, not a small one that "cracks"
- Avoid small commodities (titanium, vanadium) where one new deposit cracks the market and "you get your ass kicked." Play big markets — energy or copper. Given the AI scenario + 20-year demographics, copper "will do extraordinarily well." Those who fear a depression should avoid copper; those who think we muddle through, "that's a no-brainer."
40:20 Uranium — the surprise winner; energy density & the 1973 analog
- AI needs 24/7, non-carbon power — uranium delivers both. The Strait conflict's "unsung beneficiary" is uranium: the 1973 Arab oil embargo built the world's #3 and #4 nuclear fleets (Japan, France) out of energy insecurity. Japan's diet noted uranium's energy density lets one warehouse store 5 years of national power. Energy security is back after a 50-year absence; probability uranium is the fuel of AI = 100% (over 5–10 years, absent a catastrophe).
43:05 SMRs & build-at-scale — the US Navy already proved it
- SMRs aren't new — the US Navy has run them for decades (a submarine). US conventional nuclear stalled after Three Mile Island and was built as costly one-offs ($20bn / 26–27 years) with no standardization; China builds the same plant for ~$5bn in ~3 years because it builds at scale "like Ford" (CGN). The US is building a commercial SMR now at Kemmerer, WY (TerraPower / Bill Gates) on an old coal mine. Build more, get cheaper.
47:52 Nuclear waste is fuel — reprocessing closes the cycle
- New reactors can use existing nuclear "waste" as feed — a utility fuel-supply veteran told him 15 years ago to store waste "where you can get it back." A pound of enriched uranium is barely utilized; reprocessing (conventional or new tech) is astonishing, and rising mine-supply deficits otherwise threaten the industry. Fusion would be transformative if ever commercialized, but nothing's "on the doorstep."
52:13 Precious metals — sold silver at $75, "bet against the hyperbolic chart"
- He won't call the precious-metals bottom. He reiterates selling his silver at $75/oz into the top (with Quartermain at VRIC, taking "slings and arrows"). His rule: "when I see a hyperbolic chart, up or down, I bet against it." Vertical moves don't end by going sideways — the back of the hockey stick is as steep as the front.
54:23 Gold, rates and the dollar — numerator vs. denominator
- Higher nominal rates lift the dollar; gold is dollar-denominated, so gold can falter in the face of higher rates. But the US "can't afford real interest rates" — they'd decimate the long bond at refinancing, then equities/housing — so eventually, like end-1975, the political class loses its nerve and sacrifices the dollar to domestic politics. When that happens, gold is the beneficiary. "Will it occur in 2026? No idea. Will it occur? Absolutely."
55:25 What Rule is doing now — save in gold, take more risk in the riskiest miners
- He saves systematically in gold (price-insensitive). On gold stocks he's taking more risk: in a risk-off market the names down most are the perceived-riskiest, while after 2½ years of exploration spend he's seeing spectacular drill results — so he's deploying new money "in the riskiest part of the sector," reversing his year-ago "emphasize the biggest and best" advice. He won't name them — that's reserved for the symposium.
57:37 Oil supply response — only Exxon is reinvesting; the rest cannibalize
- The reflexive response is starting — $90 oil makes tier-2 Permian/Delaware locations economic and stacked rigs are turning. But the US still isn't making the sustaining/new investment (nor is Venezuela). You'll know the message landed when analysts tell companies to cut buybacks/dividends and reinvest free cash flow into production 3–5 years out — "it's not occurring except at places like Exxon." Shareholders are "always rearview."
1:01:37 Defers to Jeff Currie on inventories — but "the cure for high prices is high prices"
- On Jeffrey Currie's low-global-inventories warning, he defers ("he's a better pundit; I'm a rockhound"). His own view: high oil barely deters US demand but "obliterates" it where people can't afford it — in Reno you curse and drive; in Colombo, Sri Lanka the taxi driver parks his cab. The marginal (poorest) buyer sets the price.
1:03:26 The Rule Symposium & Battle Bank — the riskless transaction
- Symposium (Boca Raton, July 6–10; livestream available, in-person sold out): every exhibitor interviewed beforehand and posted free on YouTube (~70 interviews); unconditional money-back guarantee (refunds <0.1% of tuitions in 30 years); 46 hours over 4 days plus recordings for the balance of 2026 — "probably the only riskless transaction in finance." Battle Bank is growing "very well" off a 20,000–23,000-person wait list (he built the 28bn EverBank from zero wait list over 14 years).
3. In plain English
A jargon-free summary of the thesis behind each pick — what it actually is and why he holds that view. (Plain-language companion to the table above; renders on each ticker's consolidated page.) This was a macro/thematic talk; the substantive single-name/ETF theses are below — passing illustrations (Shell, TerraPower, CGN) carry no callout.
COPX — Global X Copper Miners ETF Positive
This fund owns a basket of copper-mining companies, so it's a way to bet on copper without picking one miner. Rule's case is mechanical, not hopeful: building all the AI data centers people expect needs enormous amounts of copper, and the mining industry simply can't dig it up fast enough. New copper mines take 16–17 years from first exploration to production, and the world skipped 30 years of building them. Mining pioneer Robert Friedland says the world will need more copper between now and 2050 than has been mined in all of recorded history — before you even add data centers. When demand outruns what mines can supply and you can't fix supply for over a decade, the only way to balance the market is "rationing by price" — the price climbs until enough buyers drop out. The one thing that breaks the thesis is a global depression that crushes demand, so this is a bet for people who think the world muddles through.
URA — Global X Uranium ETF Positive
Uranium is the fuel for nuclear reactors, and this fund holds the companies that mine and process it. Rule calls it the "surprise winner" of the AI build-out for two reasons. First, AI data centers need power that runs 24 hours a day (wind and solar don't) and that doesn't emit carbon — nuclear is the only source that does both at scale. Second, energy security is suddenly back on everyone's mind after a 50-year gap: when the 1973 Arab oil embargo scared countries that import their energy, France and Japan responded by building huge nuclear fleets, because a tiny amount of uranium stores an enormous amount of energy (Japan noted one warehouse of it could power the country for five years). He puts the odds that uranium becomes "the fuel of the AI business" at 100% — but stresses this plays out over 5–10 years, not in 2026, and only if there's no nuclear-accident scare.
CCJ — Cameco Positive
Cameco is a large Canadian uranium miner, and it now also owns part of Westinghouse, which designs and services nuclear reactors. Rule singles it out as the company "progressing the fastest in terms of nuclear technology" — so it captures both sides of his nuclear thesis: digging up the fuel and building the plants that burn it. In a world where the AI build-out forces a nuclear revival, owning the best-positioned name across the whole chain is his point here.
XOM — Exxon Mobil Positive
Exxon is the giant integrated oil company, and Rule holds it up as the one disciplined manager in the patch. His worry about oil companies is that they "cannibalize" themselves — handing all their spare cash back to shareholders through dividends and buybacks instead of reinvesting to keep production up in three-to-five years. He says almost nobody is reinvesting properly "except at places like Exxon." On top of that, Exxon is one of only two majors he knows of (Shell is the other) using AI to mine its own drilling and production data at scale — the kind of "free efficiency upside" he says isn't priced into oil stocks. So Exxon is both the well-run allocator and an AI-efficiency early mover.
GLD — SPDR Gold Shares Positive
Gold is where Rule saves rather than where he tries to make money, so the day-to-day price barely matters to him — he simply buys more whenever he has cash (he'll put conference profits into it). Near term he won't guess the bottom: when interest rates rise, the dollar strengthens, and because gold is priced in dollars a stronger dollar can push gold down. But his bigger point is that the US government can't actually afford high real interest rates — they'd wreck the bond market when the government refinances its debt, and then stocks and housing — so eventually, just like in the mid-1970s, politicians will force rates down and let the dollar weaken to protect the economy and the voters. When that happens, gold wins. His timing answer: "Will it occur in 2026? No idea. Will it occur? Absolutely."
SLV — iShares Silver Trust Neutral
Silver sits in Rule's speculation bucket, not his savings bucket. Here he revisits it as a case study in his selling discipline: he sold his silver at $75 an ounce into a near-vertical price spike — a "hockey-stick" chart — and took plenty of criticism for calling the top early. His rule is simple and symmetric: "when I see a hyperbolic chart, up or down, I bet against it," because vertical moves never resolve by drifting sideways. He didn't give silver a fresh rating in this conversation; the takeaway is the method, not a new stance.
Summary & timestamps derived from the public YouTube video (transcript in transcript.txt) for personal study. Not investment advice. © the host / Rule Investment Media for source material.