Nomi Prins — The Diesel Squeeze: What Refined-Product Tightness Means for the Real Economy and Hard Assets

The signal: a $100 diesel crack spread

• On Monday 17-AUG-2026 the U.S. diesel crack spread — the premium of ultra-low-sulfur diesel over WTI, i.e. the refiner's margin for "cracking" crude into diesel — printed a high of $102.20 (Reuters called it the first time above $100/bbl; Bloomberg logged a record the next day).
• The long-term normal range is $15–$30. Not just spiking but settling above $100 is the event: the NYMEX monthly chart at the 31-AUG close showed $99.98, with the 200-day moving average still rising.
• Caveat carried in the piece: Tracy Shuchart's point that the 17-AUG extreme may partly reflect regional U.S. refining and pipeline shortages rather than a purely global squeeze. The trend, not the single print, is the argument.

Why it's a bottleneck, not a crude shortage

• U.S. distillate inventories (diesel + heating oil) sat near 107 million barrels in early August — the lowest for that point in the year since 1996 — even though refineries are running hard and exporting heavily. Stocks simply are not rebuilding.
• Global refining capacity has been damaged by the Russia–Ukraine and Middle East conflicts. China holds the world's largest pool of spare refining capacity, but export quotas keep most of it off the global market.
• The structural point: you cannot conjure refining capacity overnight. The four drivers the author lists — geopolitical disruption/export bans, very low inventories, seasonal demand peaks (harvest), and the inability to ramp capacity quickly — are all live at once.
• The U.S.–Venezuela long-term oil arrangement may add crude eventually, but takes years and does nothing for the refining bottleneck.

The parallel European problem: gas

• EU natural-gas storage was ~63% full in late August — well below the seasonal norm heading into winter (GIE AGSI data; 2026 estimate via Asymmetric Research / ZeroHedge). The piece calls this a largely self-induced constraint.
• Dutch TTF futures are rising in sympathy. Add AI data-center load and broader electrification, and the whole energy system tightens — which makes diesel more important as backup power and for remote operations, not less.

Transmission #1 — mining: diesel is 15–25% of AISC

• For many open-pit operations, diesel routinely accounts for 15–25% of all-in sustaining costs (transport, drilling, on-site power, explosives). Several miners have already raised AISC guidance.
• The contrarian read: in a supercycle this is less damaging than it sounds. Rising metals prices historically more than offset higher AISC — and higher costs make new projects harder to sanction, stretching already-long lead times and reinforcing the scarcity that drives the cycle in the first place.

Transmission #2 — food and freight

• Diesel is the fuel of planting, harvesting and moving food. Per Bloomberg, cereal prices (wheat, corn, barley, rice) rose 22% year-over-year through July 2026; wheat hit a three-year high in August.
• Sugar jumped roughly 40% in months, prompting India — the world's largest consumer — to import over 1 million tonnes.
• Weather compounds it: wildfires near arable land plus an "unprecedented" El Niño. Too dry, too wet or too cold all end the same way — damaged yields.

Transmission #3 — the pump, and why to distrust CPI

• U.S. retail diesel is back over $5/gallon. Gasoline is no better: state-level increases from 18-FEB-2026 to 20-AUG-2026 ran +22% (California) to +69% (Iowa, Oklahoma) — Iowa $2.45→$4.14, Oklahoma $2.29→$3.87, Colorado $2.76→$4.48, Minnesota $2.57→$4.16. A 12-gallon fill-up costs about $15 more than before the war; California still pays the most at $5.59/gal.
• The conclusion drawn: even if crude softens, a wide crack spread keeps the inflationary impulse alive in the physical economy — so take CPI/PPI/PCE with a pinch of salt.

The monetary trap — cost-push inflation the Fed can't fix

• This is classic cost-push inflation: higher input costs, not excess demand. Conventional tools don't reach it — central banks can raise rates to cool demand, but "they cannot refine more diesel or rebuild inventories by decree."
• Explicit near-term call: no Fed hike in September or October — "the best thing they can do at the moment is to do nothing."
• Bond yields keep grinding higher (10-, 20- and 30-year monthly charts as of the 31-AUG close), reflecting inflation plus fiscal pressure. Persistent energy/logistics cost-push keeps nominal yields higher for longer, which forces the Fed to run the economy hot — the only politically palatable option once you rule out raising taxes, cutting spending, or defaulting.

Why that backdrop favours gold — even with high nominal yields

• Running economies hot raises the probability of QE, balance-sheet expansion and other liquidity support, which accelerates currency debasement. In that world, elevated nominal bond yields do not prevent gold from performing — what matters is real yields falling, plus growing demand for protection against eroding currencies.
• The self-reinforcing loop: physical scarcity → higher metals prices → sticky inflation → a higher chance of a monetary response → which again favours real assets.
• Years of underinvestment in both refining and mining capacity, against structural demand growth from electrification and AI, is what makes this "more than a short-term disruption."

The stated playbook (no securities named)

• The macro frame: higher commodity prices in a supercycle; higher inflation; QE and/or broadening monetary accommodation; continued fiat debasement.
• The action points: clear bad debt; watch the diesel crack spread closely; hold physical gold and silver as insurance; keep at least 6 months of cash/liquid reserves; favour well-capitalised commodity producers in areas of genuine supply constraint where rising prices more than offset higher AISC; don't assume high bond yields restrict gold; and have a Plan B (second residency/passport).
• Note the omission — not a single ticker or company is named anywhere in the piece. It is an asset-class and process note, so there is no securities table on this page by design.

Reinforces the standing hard-asset thesis from a new angle

• It arrives at Prins's own conclusion — debasement and structural scarcity, not the policy rate, set the gold price — by a different route than her 2026-AUG-30 rate-hike rebuttal. Hers was monetary (the 2022–23 back-test, $40T debt, ~1,000 t/yr of central-bank buying); this one is physical and cost-side (diesel cracks → AISC → sticky CPI → hot economy → QE).
• The AISC channel is the practical addition for the miners in the index: a 15–25% diesel share of open-pit cost means the same energy squeeze that supports the metal price also compresses producer margins first — which is why the piece's producer screen is "well-capitalised, in genuinely supply-constrained areas, where price gains outrun AISC."
• The power-demand leg — AI data centres and electrification tightening the whole energy system — is the same one running through the 2026-AUG-19 nuclear piece and the 2026-MAY-28 firm-power recommendation, here reaching diesel as backup generation.