Get the daily leads by email

Every lead on this site, in your inbox each morning. Free.

Mon, Sept 14th, 2026

A fuel spread that has never been this wide, and a rainy season that decides how much copper Zambia mines next year.

01Score82

Diesel has never been this expensive relative to gasoline, and the people it bankrupts own the trucks

Summary

US diesel now costs $1.81 a gallon more than gasoline, the widest gap in the twenty-three years the government has published the series — and the market has answered by selling trucking companies down to the lowest prices of the year. The overlooked part is that the carriers a fuel shock actually destroys are mostly not the listed ones.

The US Energy Information Administration put the national average retail price of diesel at $5.967 a gallon for the week ending September 7, 2026, up from $4.578 on July 6, 2026 — a rise of 30% in nine weeks. Regular gasoline over the same week averaged $4.157.

The resulting gap of $1.81 a gallon is the widest in the EIA’s weekly national series, which begins in September 2003. The previous record was $1.61, set in the week ending November 28, 2022. What has moved is not crude but distillate: the shortage is in refining and in refined-product trade flows, not in the barrel.

That fuel shock is landing on a trucking industry whose driver supply was already being reduced by rule. The Federal Motor Carrier Safety Administration’s non-domiciled commercial driver’s licence rule took effect on March 16, 2026, restricting eligibility to a narrow set of visa categories; the agency says roughly 194,000 current non-domiciled CDL holders are affected as their licences come up for renewal, and about 13,000 drivers had already been removed. Trade reporting puts net carrier revocations up 31% year over year through the first half of 2026, and spot rates, although down more than 17% from a July 4 peak, still ran roughly 34-38% above 2025 levels through August.

Prices verified at the September 14, 2026 close: Old Dominion Freight Line at $182.72 and XPO at $179.42 were both at six-month lows. Knight-Swift closed at $68.13, 17% below its June 11, 2026 high; Werner at $37.45, 21% below its July 21 high; Heartland Express at $12.35, 24% below its June 11 high. Over the same stretch Valero closed at $381.11 and Marathon Petroleum at $392.48, each within 3% of a six-month high and each roughly 70-85% above its April 17, 2026 low.

Opportunity

The obvious reading is that expensive diesel is a cost problem for trucking companies, and the share prices have followed that logic down. It is incomplete, because the large listed truckload and less-than-truckload carriers recover fuel through contractual surcharges indexed to the same EIA weekly number, typically with a one-to-two-week lag. The operator who recovers nothing is the small fleet and the owner-operator, who buys diesel at retail, hauls on the spot market, and has no surcharge mechanism at all.

Small carriers are the majority of US truckload capacity, and they were already being removed for reasons that have nothing to do with fuel — a licensing rule that disqualifies a category of driver at renewal, and enforcement-driven revocations of operating authority. A 30% increase in the single largest variable cost, arriving in nine weeks on top of that, is the standard mechanism by which a freight cycle turns. Capacity leaves faster than freight does, and contract rates reset upward at the January bid season for whoever is still holding a licence.

Hypothesis: the market is pricing a cost shock and not a supply shock. That is why carriers with surcharge protection and compliant driver rosters are trading at the lows of the year at the same moment their un-surcharged competition is being eliminated. If that is right, the variable worth measuring is not fuel expense but the rate of carrier exit, and the point at which it shows up in earnings is the winter contract season rather than the next quarter.

The record spread printed on September 7, 2026 and appears so far to have been discussed as a refining-margin story rather than a freight-capacity one. That is an inference about attention, not a fact about positioning.

How it could play out

Diesel stays expensive relative to gasoline because the constraint sits in distillate refining rather than in crude supply. Small carriers without surcharges run down cash, hand back trucks and let authorities lapse. Capacity leaves the spot market faster than freight volumes fall. Spot rates firm, and contract rates then reset at the January 2027 bid season in favour of the carriers that survived, with the benefit landing in 2027 earnings rather than 2026. Separately, a sustained diesel premium changes the arithmetic of moving freight by rail, which burns a fraction of the fuel per ton-mile, making intermodal conversion economic again for lanes that had gone back to the road.

Questions worth asking

How fast are carriers actually leaving? FMCSA publishes net revocations and operating-authority counts monthly, and that series decides this idea — everything else here is inference from a fuel price.
Which listed carriers have the tightest fuel surcharge pass-through and the shortest indexing lag, and which are exposed on empty miles and deadhead, which surcharges have never covered?
Is the distillate problem structural rather than cyclical? Russian refining capacity has been under sustained attack and Gulf refined-product exports are constrained, either of which would hold the diesel premium wide for longer than a normal cycle.
Who else pays this bill with no surcharge at all — farmers, construction contractors, marine operators, waste haulers, regional less-than-truckload — and is any of them still priced for a normal diesel cost?
If truck-to-rail conversion becomes economic at this spread, which intermodal franchise gains volume first, and has anyone put a modal-shift assumption into rail estimates?
Does an unusually warm North American winter, which a strong El Niño makes more likely, cut heating-oil demand enough to narrow the diesel premium before the contract season opens?

Where to look

Knight-Swift Transportation, Werner Enterprises, Heartland Express, Schneider National — asset-based truckload carriers with contractual fuel surcharges, all trading well below their 2026 highs
Old Dominion Freight Line and XPO — less-than-truckload operators, both at six-month lows, in a subsector where pricing discipline has historically held through downturns
J.B. Hunt Transport Services — the largest domestic intermodal franchise, and the most direct expression of any truck-to-rail conversion
Union Pacific, CSX, Norfolk Southern — the railroads that would carry converted freight, several times more fuel-efficient per ton-mile than a truck
Ryder System — truck leasing and fleet management, where equipment handed back by failing carriers physically ends up
Valero Energy, Marathon Petroleum, Phillips 66 — the refiners capturing the distillate margin, and the crowded side of the same fact

Thesis check

The fuel number is a government series that can be checked every week, the record is unambiguous, and the mechanism by which a fuel spike removes un-surcharged trucking capacity has a close precedent in 2022 and 2023. The weakness is demand: if freight volumes are contracting at the same time capacity is, rates need not rise at all, and Knight-Swift, Werner and Old Dominion can sit at these prices for a long while the capacity arithmetic grinds through — the monthly carrier-exit data, not the diesel price, is what would settle the argument.

Timing

Winter contract-rate bids open January 2027; EIA fuel prices publish weekly

Sources

US Energy Information Administration, Gasoline and Diesel Fuel Update, Sep 7 2026 · FMCSA, Non-Domiciled CDL 2026 Final Rule FAQs · Federal Register, Feb 13 2026 · FreightWaves, carrier revocations · Commercial Carrier Journal, Aug 2026

Open on its own pageFound Sep 14trucking

02Score77

The reservoir that powers Africa's copper is 42% full going into the strongest El Niño ever forecast

Summary

US government forecasters now put three-in-four odds on this El Niño exceeding every event since records began in 1950, and El Niño reliably dries southern Africa in exactly the months Lake Kariba has to refill. Kariba's water is what runs Zambia's copper mines, and the last drought there took the lake under 3% and forced mines to cut power use by 40%.

NOAA’s Climate Prediction Center issued its monthly ENSO diagnostic discussion on September 10, 2026 under an El Niño Advisory. It puts a greater than 90% chance on a very strong event through the northern-hemisphere autumn and winter of 2026-27, and a 75% chance that the October-to-December season produces a historic event exceeding the strength of every El Niño on record back to 1950.

The shape of it matters as much as the size. August anomalies ran +3.4°C in the Niño-1+2 region, +2.5°C in Niño-3 and +1.8°C in Niño-3.4, while Niño-4 fell to +0.1°C. That is an eastern-Pacific event rather than a central-Pacific one — the pattern of 1982-83 and 1997-98, and the pattern most strongly associated with failed rainy seasons over southern Africa. The centre’s next update is scheduled for October 8, 2026.

The Zambezi River Authority reported Lake Kariba at 481.39 metres on September 7, 2026, equal to 42.25% of usable storage and in recession, against 478.19 metres and 18.74% on the same date in 2025. The lake is designed to operate between 475.50 and 488.50 metres for hydropower, and it refills only from the rains that run November to April.

Zambia mined a record 890,346 tonnes of copper in 2025, and its finance ministry expects output to pass one million tonnes in 2026 and reach 1.2 million in 2027, helped by First Quantum’s completed $1.25bn Kansanshi and Enterprise expansion and Barrick’s $2bn programme to double Lumwana. During the 2023-24 El Niño drought Kariba fell below 3% of usable storage, load shedding reached 21 hours a day, and the state utility ZESCO asked mines to cut consumption by 40%. Zambia has since commissioned roughly 286 MW of new solar capacity.

Prices verified at the September 14, 2026 close: copper at $6.41 a pound, 5.8% below its September 9 six-month high of $6.80; First Quantum Minerals at C$42.49, 12.7% below its August 25 high; Barrick Mining at $42.79, 12.4% below its August 25 high; Ivanhoe Mines at C$12.52.

Opportunity

El Niño is read as an agricultural event — grain, palm oil, coffee, sugar — and that is where the forecast gets discussed and traded. The consequence with the shortest and hardest causal chain is electrical rather than agricultural. A strong eastern-Pacific El Niño suppresses the November-to-March rains over the Zambezi catchment, and Kariba is a storage reservoir whose entire following year of generation depends on that single wet season.

Zambia’s copper industry is, in operating terms, an electricity business with a mine attached. Comminution and smelting are continuous loads that cannot simply be shed without damaging equipment and losing recovery, which is why the 2024 response ran to a 40% curtailment request rather than something gentler. The country is simultaneously trying to lift output by roughly a third across two years, which raises the load on that grid at precisely the moment the water behind it is put at risk.

Hypothesis: copper is priced on demand and on Chilean and Peruvian supply, and southern African hydrology is not a variable most copper supply models contain. If a record El Niño delivers even a fraction of the 2023-24 rainfall deficit over the Zambezi, the lost tonnes land in a market with very little visible inventory, and they are the same tonnes the 2027 supply forecasts are already counting on.

The honest counterweight is the starting level. The lake enters this season at 42% of usable storage rather than the 19% it held a year ago, so this is a risk with a buffer in front of it rather than a repeat that is already under way.

How it could play out

The eastern-Pacific El Niño peaks over October to December 2026 and suppresses the southern African rainy season. Kariba’s inflows come in below normal from late November and the lake fails to refill through April 2027. ZESCO’s generation deficit widens, imports and emergency generation are bought at cost, and mines are asked to curtail again. Zambian output misses the one-million-tonne mark and the 2027 ramp slips. Those tonnes are subtracted from a copper balance that already assumes them — the miners concentrated on that one grid carry the earnings risk, and the copper price carries the other side of the same event.

Questions worth asking

How much Zambian and Congolese copper output actually depends on Kariba, as opposed to imports, coal, or captive solar and gas? Answering that sizes the entire idea; everything else is inference from a lake level.
Which listed miner has the largest share of group production on that one grid, and has any published model assigned it a probability of curtailment rather than treating power as a given?
Is the forecast symmetrical? An eastern-Pacific El Niño brings heavy rain to coastal Peru and northern Chile, where much of the rest of world copper supply and its tailings infrastructure sits — does the same event create flood and disruption risk at the other end of the cost curve?
Who sells the substitute? A southern African power deficit is historically met with imported power, diesel generation and temporary rental plant, and that demand arrives fast and at bad prices for the buyer.
Zambia’s fiscal path is built on copper volumes and royalties. What happens to the kwacha and to its restructured sovereign debt if the volumes do not arrive?
Does the same rainfall deficit reach further than power — southern African maize, South African agriculture, and the Zimbabwean half of Kariba’s generation?

Where to look

First Quantum Minerals — Kansanshi and Sentinel in Zambia are the core of the company, and it has just completed a $1.25bn expansion there
Barrick Mining — Lumwana in Zambia, mid-way through a $2bn programme to double the mine
Ivanhoe Mines — Kamoa-Kakula in the Democratic Republic of Congo, drawing on the same interconnected southern African power system
Copper itself, and the diversified copper miners priced primarily off Chilean and Peruvian supply, as the expression if the loss proves systemic rather than company-specific
Temporary and rental power providers and large engine-genset builders such as Wärtsilä, which is what a regional grid deficit has historically bought in a hurry; note the purest operators in this niche are privately held
Zambian sovereign debt and the kwacha, where mining royalties are the fiscal base

Thesis check

The two numbers underneath this are published by the agencies themselves and can be watched weekly — NOAA’s ENSO forecast and the Zambezi River Authority’s lake level — and the 2023-24 drought is a precedent close enough to serve as a template rather than an analogy. The weakness is that Lake Kariba begins this season at 42.25% of usable storage rather than the 18.74% it held a year earlier, so a poor rainy season degrades that buffer without necessarily exhausting it; a record El Niño in the Pacific also does not guarantee a record rainfall deficit over the Zambezi specifically, and Zambia has added several hundred megawatts of solar since the last crisis.

Timing

Zambezi rains due from late November 2026; next NOAA forecast update October 8, 2026

Sources

NOAA Climate Prediction Center, ENSO Diagnostic Discussion, Sep 10 2026 · Zambezi River Authority, Lake Kariba weekly levels, Sep 7 2026 · The Africa Report, 2026 · Pulitzer Center, Zambia power shortages · Energy Transition Africa, 2026

Open on its own pageFound Sep 14commodities

Also worth knowing

Washington took preferred equity and warrants in a tungsten maker and handed it a $2bn stockpile contract — The War Department announced on September 14, 2026 a $450 million committed investment in The Elmet Group, starting with a $200 million drawdown at closing, in exchange for redeemable preferred equity and warrants for up to 19.9% of the common stock. More than $165 million goes to plants in Lewiston, Maine, Coldwater, Michigan and Euclid, Ohio, and about $150 million to a majority-owned joint venture at the Springer Tungsten Complex in Nevada alongside EQ Resources. Separately the Defense Logistics Agency awarded an indefinite-delivery contract worth up to $2 billion, with a guaranteed minimum of $150 million through at least August 2031, to supply tungsten for the National Defense Stockpile. Elmet closed at $21.70, up from $16.78 five sessions earlier.

The instrument is the interesting part rather than the metal. The fiscal 2027 budget request carries $18.0 billion for the National Defense Stockpile, inside roughly $48.8 billion across stockpile, industrial-base and Defense Production Act lines — a scale change that turns the government into a contractually committed, price-insensitive buyer. What that does to the cost of capital for Western mine projects, which have failed for fifteen years because nobody would underwrite a price China sets, is the part nobody has costed.
SpaceX’s weight in the Nasdaq-100 more than doubles this week — Nasdaq’s pro-forma data has SpaceX rising to roughly 2.82% of the index from about 1.28%, with the change effective before the September 21, 2026 open and passive funds expected to trade it at the September 18 close. JPMorgan’s estimate of the resulting net index-fund buying is about $15.5 billion. The driver is mechanical: SpaceX’s free float has climbed from under 10% to nearly 30% as lock-ups expire in tranches, which is what the index rules respond to.

The flow itself is well covered. The part that is not is that the same mechanism repeats: each further tranche of lock-up expiry raises the float, raises the index weight, and forces another round of passive buying, while simultaneously releasing insider supply. Mapping that forward schedule against the remaining lock-up calendar is a calculable exercise that would tell you the size and timing of every future episode.
Russia is now importing refined fuel, and most of it comes from a refinery it part-owns in India — Russia imported a record 172,000 tonnes of oil products in August 2026, with India supplying 70% of it, including 120,000 tonnes of gasoline worth €78 million. All of the gasoline loaded at the Vadinar refinery, in which Rosneft holds 49.13%, and which sourced all of its crude from Russia over the first eight months of 2026. Ukrainian drone strikes hit a record 14 refineries in August; 21 of Russia’s 38 large refineries have been struck since January 2025.

A net fuel exporter paying to have its own crude refined halfway around the world and shipped back is the clearest available measure of how much global refining capacity has been destroyed rather than merely disrupted. It is also the supply-side explanation for why distillate, not crude, is where the price damage sits.
The Navy’s first shore-based nuclear microreactor goes to Crane, Indiana, with no vendor named — The War Department announced on September 9, 2026 that a privately owned and operated advanced microreactor will be installed at Naval Weapons Station Crane no later than September 2028, executed through the Army’s Janus programme. It will be Indiana’s first power-generating reactor. Eight companies remain qualified to compete, and officials said the selection for Crane would be announced later.

A dated, funded order without a named winner is the condition in which the small listed nuclear names trade on speculation rather than on contracts. The check worth doing is which of the qualified vendors can actually deliver fuelled hardware by 2028, because enriched fuel availability, not reactor design, is the binding constraint on every one of these programmes.
The Democratic Republic of Congo will end its cobalt export ban on October 16 and replace it with quotas — Kinshasa will lift the eight-month export ban and move to annual limits of 96,600 tonnes for each of 2026 and 2027. The ban had removed the world’s dominant supplier from the market entirely; the quota restores flow but caps it well below prior export levels.

A quota is a different animal from a ban, because it converts a binary supply question into a permanent administered price floor and gives Kinshasa an annually renegotiable lever. Who holds allocation under the quota, and whether it is distributed by producer or by buyer, is what determines which listed miners and which battery chains actually get metal.
Sweden’s election left the centre-left bloc ahead by three seats with counting unfinished — With votes from 6,195 of 6,312 districts after the September 13, 2026 vote, the left bloc led by Magdalena Andersson was on 176 seats in the 349-seat Riksdag against 173 for Ulf Kristersson’s governing right-wing bloc. Final results were expected on Wednesday at the earliest.

The incumbent government legislated state risk-sharing for new nuclear construction and Vattenfall has been running a reactor procurement on the back of it. A three-seat swing that puts a coalition with unresolved internal divisions on nuclear into office is a live policy risk for a programme whose suppliers are listed elsewhere in Europe and North America.
The IMO’s global shipping carbon price returns for a vote in October 2026 — The Marine Environment Protection Committee adjourned in October 2025 without adopting the Net-Zero Framework, on a 57-49 vote to postpone for a year after the United States threatened sanctions and tariffs against supporting states. The framework would apply a carbon price of about $100 per tonne of CO2-equivalent to shipping from 2028, with the earliest entry into force March 1, 2028 if adopted this October.

Shipping is being valued entirely on Hormuz freight rates at the moment, and a binary vote that would reset fuel economics, newbuild specification and scrapping decisions for the whole world fleet is four to six weeks away. If it fails a second time, the EU’s own emissions regime becomes the de facto standard and the market fragments by trade lane rather than converging.
A permanent $103,265 fee on new cap-subject H-1B petitions is in its final comment window — The Department of Homeland Security’s notice of proposed rulemaking, published August 25, 2026, would add $103,265 on top of existing fees for every one of the 85,000 annual cap-subject H-1B petitions, including advanced-degree cases. Comments close September 24, 2026.

The first-order losers are obvious and have been discussed for a year. The unexamined question is where the work goes if entry-level foreign technical labour is priced out of the United States: offshore delivery centres and captive global capability centres in India, nearshore operations in Canada and Mexico, and automation budgets — each of which has a listed real-estate and services chain attached that is not priced for a step-change in demand.
Memory price increases are decelerating without the shortage ending — Contract DRAM prices are projected up 13-18% quarter over quarter in the third quarter of 2026 and NAND up 10-15%, against roughly 60% quarterly jumps in the second quarter. DRAM rose about 170% across 2025. Server DRAM is still expected to be undersupplied, with high-bandwidth memory absorbing the capacity.

A decelerating price increase in a market that is still short is the point at which the buyers, not the sellers, become the interesting question. The companies to identify are those with memory-heavy bills of materials and sticky output prices — automotive electronics, networking hardware, set-top and consumer devices — where two years of input inflation has to show up in a gross margin eventually.
A third of companies say AI coding agents stopped them buying software — McKinsey research published around September 13, 2026 found that nearly one in three surveyed organisations had decided against purchasing at least one software product or feature because they could build the functionality internally with AI-powered coding agents.

This is the first hard survey number attached to an argument that has been made rhetorically for two years. It is a single data point and a self-reported one, but it is the sort of measurement that, if replicated, changes how seat-based software businesses are valued — and the useful work is finding which categories were named, because “at least one feature” and “the core platform” are very different claims.

All leads54 leads · Aug 31 – Oct 4

← Back to today

all 54 energy 9 power 6 commodities 5 defense 4 policy 4 agriculture 3 biotech 2 insurance 2 semiconductors 2 shipping 2 space 2 trade 2 trucking 2 utilities 2 ai 1 automotive 1 coal 1 pharma 1 rates 1 turkey 1 uranium 1