Adam Rozencwajg (founding & managing partner, Goehring & Rozencwajg Associates — a New York natural-resource investment firm, ~$2.5B AUM) · oil, gas, shale geology, refining, precious metals, uranium, coal, agriculture and the natural-resource capital cycle — running synthesis of his appearances, with per-item breakdowns and a stock index.
The four-letter word nobody can finance, into two dated catalysts: Europe missing its winter gas requirement (the 2022 coal-burn playbook) and Indonesia moving to curb the exports that swing the seaborne market.
Phosphate transits Hormuz, ammonia/urea are made from disrupted gas, and sulfur went $300→$1,100/t — exposure increased "a little bit," deliberately short of full conviction.
Qatari LNG through Hormuz is disrupted so the global gas market is very tight; US supply is propped only by the Permian "gas burp," an ageing-field artifact that does not last.
The volatile leg of the oil book: balance sheets wiped clean by COVID bankruptcies, industry consolidating, "a massive bottoming process" — biggest contributor or biggest detractor in any given quarter.
The mispriced asset: ~10 mb/d shut in since March against inventory that is mostly working capital, and the "5 mb/d of lost demand" is really a 6 mb/d refining outage — "where do prices go from here? I suspect higher."
How G&R owns the barrel — "I trade the equities." Still out of favour after six months of the worst physical disruption in decades, into a sector spending ~$550B/yr against the ~$1.5T/yr needed.
The core oil position since the January rotation: "good long-lived assets in a friendly jurisdiction" — decades of steady production without shale's drilling treadmill, in the safest venue left.
The term price — where 90% of the market transacts — hit an all-time nominal $95.50/lb while the equities fell 30-odd percent on no news; utilities under-contracted, no new mine supply. "Just buy it and put it away."
Owned selectively, not as a sector: quality reserves are getting hard to find and depletion is fast, so it is "not enough to build a whole portfolio around" — but the same slowdown is the reason he is long crude.
Infrastructure data point only: Abqaiq hit again and under-reported, though he is less worried than consensus that shut-in Gulf fields fail to restart.
A one-line mention as the rumoured suitor for NexGen — evidence that major diversified miners are circling development-stage uranium while new supply stays years out. No company view offered.
The real bottleneck and the diagnostic: $100 cracks against a normal $10–20 say refining, not demand, is broken — non-OECD product stocks maybe down 500 mb vs the reported 50 mb. He owns the crude leg, not the margin.
Priced to perfection after 15–20 years of exceptional yields with everyone on one side of the boat; a monster El Niño cuts both ways, so a tail risk to monitor rather than a position.
Cited by the host as evidence, not rated: Mosaic cut phosphate output hard, most likely on the tripled sulfur price — a supply cut into an already-tight nutrient market.
Named, not rated: the BHP takeover chatter is beside the point — Rook I is "a number of years away," which is precisely why the uranium term price keeps making highs.
Sold in January on the "silver sell signal" and still out: ~20% down against a typical ~40% over one to two years, ETF length still elevated, gold/oil still in oil's favour — long-term bull, near-term not finished.
Deliberately avoided — record cracks are the symptom, not the prize; the shortage began upstream, so the spread should close by crude rising rather than products holding.
In one line: The oil crisis everyone has declared a false alarm is only halfway through showing up in the data — 10 mb/d shut in since March against inventory that is mostly working capital, with the shortage quietly migrating into refined products, the half of the balance nobody can even measure. Own the crude molecule through the equities — Canadian oil sands, selective US shale, offshore drilling — not the refining margin; stay out of gold until the correction finishes its normal depth and duration; and pick up the two things nobody will finance, uranium and coal.
The deficit is arithmetic, not a forecast. 10 mb/d of upstream production shut in for 100 days is "essentially a billion barrels of oil that never made it to market," and the world never had a billion barrels of drawable stock — most reported inventory is working capital filling pipelines and tankers, "not readily available to be drawn down." ~400M+ barrels are already gone; the running deficit is nearer 1.76 billion at day 176.
The calm is bought, not free. Three temporary things masked it: a 45–50 day lag before missing barrels appear (only three of five reported months have felt it, so even a reopening tomorrow means two more months of draws); a one-off ~100M-barrel flotilla that escaped during the MOU reopening; and Chinese, Russian and Gulf refinery run cuts that transferred the squeeze from crude into products. Reading the absence of a crisis as proof of abundance is "really really really myopic thinking… you could very well be in the eye of the storm."
The tells are the cracks and the flights — not the agency demand number. Record 3-2-1 and a diesel crack near $102 (normal $20–30) price diesel as if oil were $155; genuine demand destruction historically needed ~$200 in real terms. Meanwhile the reported "demand is down 5 mb/d" is an artifact of using refinery runs as a demand input — "the tail wagging the dog" — flatly contradicted by commercial flight traffic up 5–6% y/y. Reassign the residual: ~4 mb/d of product inventory is draining uncounted. By September he has hardened this into the central claim: global refinery throughput is down ~6 mb/d (Gulf idled, China's ~2 mb/d export refining halted, Russia war-damaged ~1 mb/d) against a reported 5 mb/d demand loss — "very, very similar to that same figure." Cracks that normally average $10–20 "hit 100 bucks. And they're staying there. That… is the sign of a refining problem."
The missing barrels are in the one series that cannot be observed. Satellites can read crude tanks because their roofs float on the oil and the wall's shadow encodes the fill level; refined-product tanks have fixed roofs and give a satellite nothing — "but it doesn't stop people from estimating." So the residual lands exactly where measurement is weakest: the IEA carries non-OECD product stocks down ~50M barrels, "and if we're right, it might be down 500 million barrels. So it's an order of magnitude difference" — with fuel-shortage headlines already coming out of Bangladesh, India and the Philippines.
The supply answer is a decade away, and the boardrooms won't start it. Shale peak growth was 2019 (G&R's own neural-net call — "peak growth, not peak production"), crude is now negative year-on-year, and the Permian is down to six counties while gassing out — the "gas burp," an ageing-field signal, not growth. The industry spends ~$550B/yr against the ~$1.5T/yr needed to recapitalize, and with reserves capitalized at half NAV a CEO creates more value buying back stock than drilling. Capex trims and a flopped ANWR auction at $85+ oil are the proof.
Positioning: long the molecule, through equities — and named as of September. "The crude oil molecule was the mispriced asset today. I continue to think that the oil molecule is the mispriced asset," expressed as Canadian oil sands ("good long-lived assets in a friendly jurisdiction"), US shale where quality reserves can still be found ("not enough to build a whole portfolio around the way it was a few years ago"), and offshore drilling — post-bankruptcy balance sheets, consolidation, "a massive bottoming process," and the biggest swing factor in the book each quarter. Explicitly not the US refiners — the implied bet is "the crack spread would come down with oil benefiting." Plus a small, not-yet-full-conviction fertilizer add and a watch-only stance on grains, "priced to perfection."
The $90 print is a risk limit, not a view. Speculative shorts were force-covered at the outbreak (the pod shops "fired their energy teams" — a positioning tell), rebuilt past the January extreme by the second MOU, then cut again by risk committees rather than by conviction. "The only thing that's really affected the price has been what your risk team is allowing you to short. So, that's going to flip… And you're at a $90 price without any of that having happened yet." Both exits are bullish: a product crisis in the global south into Cushing at operational minimums, or a reopening in which idled refineries, drained SPRs and normal activity all bid for crude at once.
Out of gold until the correction does its normal work. G&R sold gold in January on their own "silver sell signal" — silver lags, then stages a violent catch-up rally, which historically marks the top for both, followed by ~40% over one to two years. So far gold is down ~20% in six or seven months, matching only the shallowest instance in 50 years; ETF holdings are "still quite elevated"; the gold-to-oil ratio "is still very much in oil's favor"; and selling gold for oil in January drew "a loud audible groan" while the reverse today "would give me a parade." Long-term bulls — the single override is central banks doubling or tripling their purchases.
The two things nobody will finance: uranium and coal.Uranium — the term price, "where 90% of the market transacts," made an all-time nominal high at $95.50/lb while the equities fell "30-odd percent, on absolutely no news whatsoever"; utilities are "very under covered," and even NexGen's Rook I (BHP rumours notwithstanding) is "a number of years away," so "just buy it and put it away." Coal — "a four-letter word, completely starved for capital, no one cares" — into two dated catalysts: Europe missing its winter gas requirement (the 2022 coal-burn playbook) and Indonesia, the swing exporter, moving to curb exports.
Falsifiable, and dated. A non-reckless SPR tank bottom is ~18 weeks out on the arithmetic he endorses, and Trump and Vance justified the MOU by saying four more weeks of crisis "there was going to be a major problem" — now seven weeks ago, with the strait shut again. His own line: no crisis by November or December and the framework needs revisiting; meanwhile "I don't think we can write down the probability of tank bottoms to zero."
Appearances
One dated page per item — each has its stock table, talking points, and the saved transcript. Newest first.