Omaha. The last race is over.

The grandstand at Ak-Sar-Ben is emptying. Men in hats walk to their cars. Some of them had a few too many beers. The floor is covered with paper. Thousands of tickets, torn, crumpled, stepped on.

A boy is bent over, picking them up.

He is not picking up all of them. He wants a specific kind. Tickets thrown away by men who only understood the word “win.” Place tickets. Show tickets. Tickets held by someone too drunk to read his own bet.

Some of those tickets paid.

The track would not cash them for a kid. His Aunt Alice did.

They called it stooping. The boy was Warren Buffett.

Ten thousand pages, twice

Ten years later he was still stooping. Only the tickets had changed.

At twenty one he sat down with the Moody’s Manuals. Thousands of pages of tiny print, one company after another, numbers out to the margins. He went through them page by page. Every business. Then he did it again.

That is where he found Western Insurance Securities. Price in the manual, twelve to twenty dollars. Earnings, sixteen dollars per share. One year of profit bought the whole stock. Everyone else had stepped over it.

Decades later a student asked him how to find ideas today. Start with the A’s, he said.

Nobody wants that answer. There is no formula in it. There is a boy in an empty grandstand and a young man turning ten thousand pages, and both are doing the same thing. Looking at every ticket, because the winners are lying on the floor with the losers.

I have three children, a day job and this newsletter. I do my research in a café with a pen and a piece of paper. I do not have ten thousand pages in me. I checked. 🙂

So I did the only honest thing a man with less time than Buffett can do. I wrote down exactly what a winning ticket looks like, in one line. Then I let a stock screener do the bending for me. Which screener, I will get to in a minute. The line matters more.

One line

A company passes if its diluted share count today is at least 20 percent below five years ago, or at least 30 percent below ten years ago. One of the two is enough.

Nothing else. No market cap. No sector. No profit. No debt limit. No P/E. No requirement that the company bought a single share this year. I need you to understand why.

A press release says “we repurchased 500 million dollars of stock.” It does not say that options, restricted stock and an acquisition put most of those shares straight back. Diluted shares are what is left after all of that. It is the only number that tells you how much of the business your share owns. Everything else is management talking.

I do not ask whether they bought this year, because buybacks are lumpy. A good allocator buys hard when the stock is cheap, then stops when the price gets silly. A screen that demands a buyback every year rewards the company that buys mechanically at any price and punishes the one that paused for a reason. I want the second company. Twenty percent over five years is about 4.4 percent a year. Thirty over ten is about 3.5 percent a year. Both leave room for a pause.

And there is no P/E, because a cyclical at the bottom of its cycle has no earnings at all, and that is exactly when I want it on the list. In 2025 Alpha Metallurgical had no earnings. A P/E filter would have thrown it out. Earnings are the easiest number in a report to move. The share count is the hardest.

Five hundred and sixty three tickets

I run that line in Stock Rover. I started on the free trial in July, before there was any deal. Since August I have been on a complimentary Ultimate account they gave me for this work.

Explore Stock Rover’s Screening Tools

What it does for me is boring and rare at the same time. It lets me write an equation against historical fundamentals, in this case diluted share counts going back five and ten years, and run it across roughly 8,500 North American stocks in seconds. Most screeners give you dropdown menus. This one lets you write the sentence yourself.

Here is the sentence, exactly as it sits in my account.

On September 3 the line returned 564 companies. When I finalized the Library version the next day, the live data had moved to 563. That is the number behind the cuts below.

Five hundred and sixty three tickets on the floor. Now the part no software review shows you. What happens to them.

Five cuts

I added five filters, one at a time, and wrote down the count after each one.

Cash. Average free cash flow over five years must be positive. A company that shrank its share count while burning cash bought those shares with a lender’s money. 95 gone. 468 left. That cut is blunt. It also throws out General Motors and AerCap, because a finance arm makes free cash flow look like a bank’s. Those I read by hand.

Debt. Net debt below three times EBITDA. A buyback financed by a bond due in 2027 is a bet, not a return of capital. 156 gone. 312 left.

The leak. Stock compensation below 1.5 percent of market value. Stock Rover puts compensation yield in the column right next to the share count, which is the fastest way I know to catch a company that buys back three percent a year and hands two of it to management. 46 gone. 266 left.

Then I asked whether the pattern was still alive. How many of the 266 had turned into net issuers over the last twelve months?

Two.

Long term cannibals almost never reverse. 264 left.

Price. Free cash flow yield above six percent. 101 gone. 163 left.

From 563 to 163 in one afternoon. Buffett needed a year of evenings and Aunt Alice. I needed one line and a Thursday.

Two tickets that lied

I tried to kill the line. Two tickets nearly did.

Alpha Metallurgical. 18.4 million shares in 2021, 12.6 million this summer. Down 30+ percent. Net cash. It passes everything, then dies on the last cut, because free cash flow is slightly below zero. A met coal miner at the bottom of its cycle with a storm damaged terminal. The machine throws it away exactly where I want to be looking. That is why the cuts are questions, not verdicts.

Hertz. 490 million diluted shares when it left bankruptcy in 2021, 314 million this spring. Down 36 percent. The line loves it. Hertz spent 3.5 billion dollars on its own stock in 2021, 2022 and 2023, then stopped, and by July the count was back above 356 million. Five year average free cash flow, minus eight billion dollars a year by the screen’s definition, because a rental fleet is bought with borrowed money. The first cut kills it.

The machine found both. Only one is a cannibal. The machine cannot tell the difference. I can.

The watchlist

One hundred and sixty three rows. Eleven are the same company listed twice. Forty one are insurers, banks and closed end funds, where free cash flow means nothing and I skip them by hand. That leaves 111 operating businesses. That is my cannibal watchlist. Not a buy list. A reading list.

I did not know most of them. That is the point. A few that stood up off the page, with the numbers as they stand today.

• Marathon Petroleum. 649 million shares in 2020. 290 million this summer. Down 55 percent, 40.8 billion dollars retired. A refiner. The biggest cannibal on the list by a mile.

• Dillard’s. 22.7 million shares in 2020. 15.6 million now. They bought hard at 175 a share, kept buying at 255 and 307, and stopped at 650. Paid a 30 dollar special dividend instead. Exactly the behaviour the line is built to keep and every other screen throws out.

  

• CNX Resources. Appalachian gas, the one industry more hated than my coal. 224 million shares in 2020. 148 million in July. Still retiring 9.5 percent of the company a year at a P/E of six.

• H&R Block. 189 million shares five years ago. 129 million now. 2.5 billion dollars of buybacks for a company worth 6.3 billion.

• Seneca Foods. Canned vegetables. 9.9 million shares in 2016. 6.9 million now. In 2021 they offered 75 million dollars for their own stock in a tender. Shareholders sent in 531 shares. Nobody would sell. So they bought in the open market at 50 dollars. The stock is 207.

• Garrett Motion. Turbochargers. 318 million diluted shares when it left bankruptcy in 2021. 191 million now. Down 40 percent, in an industry the world has declared dead.

Not one of them has announced a “transformational” anything. 🙂

Each of these deserves its own post. Some will get one. Not today. Today is about the line.

The honest part

Stock Rover is not simple. It has more than 800 metrics and it is happy to show you a lot of them at once. The Table looks like a cockpit.

My wife walked past on the first evening, looked at the screen, and asked if I was landing a plane. 🙂

The first hour is ugly and it takes an evening before the thing looks like yours. It gets good after that. It is not good before that.

The data has bad days. A few foreign ADRs show share reductions I could not reconcile with their filings. Petrobras appears with a 64 percent reduction. Its share count has barely moved in ten years. Transcontinental in Toronto shows an 81 percent reduction and a buyback yield above 100 percent. Wrong.

Two more things before you pay for anything. Equation screening and ten years of history start at Premium Plus. The cheapest plan cannot run the line you just read. The plan page shows what sits where. My run used the “Most US & Canada” universe. Stock Rover also offers a broader “All Exchanges” setting, but that is not the same screen—and broader coverage does not automatically make the underlying data reliable. Gulf Keystone trades in London. TAV Airports trades in Istanbul. Neither appeared in the universe I used here. If your research is global, verify the available listings before relying on the screen.

Time to drop the storyteller voice.

Screener, 9 out of 10. Why not higher. It shows how many stocks pass each criterion on its own. I wanted the count after each cut, in sequence. That I built by hand.

Data, 7 out of 10. Why not higher. Petrobras. Read the ADR rows twice.

Interface, 6 out of 10. Why not higher. Ask my wife. 🙂

Coverage, for what I do, 6 out of 10. Why not higher. London. Istanbul. I still need a separate tool for direct listings.

What you get

I did the work for you. Open the trial, go to the Screener Library, and import Cannibal Stocks. The share count line and all five cuts are already built in. Run it and you are looking at the same screen I used. On September 4, it returned 163 rows—the snapshot behind the breakdown and watchlist in this article. When I checked again on September 8, it showed 158. Your live count may differ as prices and underlying data update. Remove the duplicate listings and separate the financial companies, and you have the starting point for my watchlist. The trial is 14 days, every feature, no credit card. That is more time than you need.

Start Free Trial

Coming soon

Next on my desk is a department store.

Yes. A department store. In Southern malls. The industry Wall Street buried twenty years ago.

Ten years ago it had 39 million shares. Today, 15.6 million. It bought its own stock at four times earnings and stopped at fifteen. Then the dividends started. Thirty dollars a share last January, the largest in its history.

If you own Alpha, read this one. This is what the endgame looks like when a cannibal runs out of cheap shares to eat.

The track would not cash a winning ticket for a kid. He needed his aunt.

The machine picks up every ticket on the floor. It cannot read them.

That part was never for sale. That part is still you.

Neither of us had to be the smartest man in the grandstand. We just had to bend down where everybody else walked past.

Cheers, Sandro

The screener found 163 survivors. My marriage was not one of them. 🙂

Like and Restack.

“Be assured, my young friend, that there is a great deal of ruin in a nation.”

— Adam Smith, letter to Sir John Sinclair, 1777

In 2021 Ukraine dug up seven million tonnes of coking coal.

Six of them came out of one mine.

Pokrovske, in Donetsk oblast. Two hundred million tonnes still in the ground. Every other coking coal mine in the country, added together, did not reach a tenth of it.

On January 13, 2025, Reuters reported Russian troops less than two kilometres from one of its shafts. The next day the owner, Metinvest, shut it down. The miners came up. The coal did not.

Since then Metinvest has written off $1.4 billion against the mine. What is left on its books for the shafts, the buildings and the machines is four million dollars. That is the owner telling its auditors what it thinks is coming back.

Then the coke plants. Coking coal is not burned. It is baked into coke, and coke is what a blast furnace eats. Ukraine had seven coke plants before the war. It has five. The biggest one in Europe, Avdiivka, is gone. Coke production fell from ten million tonnes to under three.

Then the steel. Twenty one million tonnes a year before the war. Seven last year. Mariupol alone took forty percent of the country’s capacity with it. What is left is five blast furnaces, three in Zaporizhzhia and two in Kamianske. On August 10 a Russian strike on Zaporizhstal killed seven people and stopped production.

That is the inventory. One mine, two coke plants, two steel cities, two thirds of the output.

Give a gift subscription

Steel without coal

And yet Ukraine made seven million tonnes of steel last year with almost no coal of its own.

When Pokrovske stopped, every forecast had Ukrainian steel falling to two or three million tonnes. It did not happen. The mills bought the coal abroad. Coal imports tripled in 2025. Coke came in from Poland by rail, most of it from JSW, the Polish state miner.

And in April 2025 a ship called the Bison docked in Ukraine with eighty thousand tonnes of coking coal from West Virginia. Metinvest owns a mine there too. When its Ukrainian mine died, it started feeding its Ukrainian coke ovens with its American coal, one ship a month.

Remember that ship.

The trade everyone has ready

On September 5, Steve Witkoff and Jared Kushner spent three hours with Putin in the Kremlin. Nothing was signed. But there is a process again, and every trader I know has the same reflex loaded.

Peace means sanctions come off. Sanctions off means Russian coal floods back. More coal means lower prices. Sell the coal stocks on the headline.

One fact against that reflex.

In 2025 Russian coal companies were offered thirty million tonnes of railway capacity they had already agreed to use. They did not show up. The trains were there. The slots were booked. The coal stayed in the ground.

Below the line, why that happened, what Russia actually sells, how much coal Ukraine really needs, and what a ceasefire does to the price.

Read more

“How did you go bankrupt?” “Two ways. Gradually, then suddenly.”

— Ernest Hemingway, The Sun Also Rises

Updated as of September 1, 2026. Sources are Xinhua, Bloomberg, Reuters, CRU, Sxcoal, Argus, Kallanish and the Australian Resources and Energy Quarterly.

May 22. Shanxi province. 19:29 local time.

Two hundred and forty seven men were underground at the Liushenyu mine when the gas ignited. Eighty two did not come up.

Within days, Beijing dispatched twenty four inspection teams into all thirty one provinces with orders to find every mine digging past its permit. Shanxi produces a third of China’s coal. More than 130 million tonnes of annual coking coal capacity went dark, almost overnight.

To understand what happens next, you need to understand the strangest commodity market on Earth. It will take a few minutes, and at the end of it you will know why this market is currently holding its breath.

A small market with a civilization-sized job

Metallurgical coal is not the coal you burn for electricity. It is the coal you bake into coke, and coke is what strips the oxygen out of iron ore inside a blast furnace. 770 kilograms of it go into every tonne of steel made the traditional way, and about seventy percent of the world’s steel is still made the traditional way.

There is no substitute. Hydrogen steel lives in pilot plants and press releases. In the physical world of 2026, no coke means no iron, and no iron means no steel.

Now look at how little of this stuff actually moves. The seaborne trade is around 300 million tonnes a year. The world pours two billion tonnes of steel. In dollar terms, the whole seaborne met coal market is worth on the order of $60 billion a year. Apple spends more than that just buying back its own stock, a comparison I enjoy for reasons regular readers will understand. 🙂 A market this small carrying a job this large is a structural absurdity, and it gets better.

Only four countries export it in size. Australia, the United States, Canada, Russia. Australia alone is about half of the seaborne trade, and around 85 percent of Australia’s share comes out of one state, Queensland, moving down a handful of rail lines into three ports. India, Japan, Korea and Europe import nearly everything they use. China mostly supplies itself, until the day it doesn’t.

And none of this can be fixed quickly. A new coking coal mine takes five to ten years from decision to first tonne, the best geology is already taken, and banks stopped financing coal years ago. Higher prices cannot summon new supply. They can only ration demand, and you are about to see why demand refuses to be rationed.

One more detail, and it is the one that loads the spring. Most tonnes move on long-term contracts. The freely traded spot market is a thin sliver on top. So when anything goes wrong anywhere, every desperate buyer on the planet converges on the same few spare cargoes.

Economists would say that both supply and demand are inelastic. I use a different word.

Nervous.

Share Cannibal Stocks

The tape

Do not take my word for it. Take the tape’s.

• In 2008, the annual benchmark roughly tripled, to about $300 a tonne.

• In 2011, Cyclone Yasi and a La Niña summer flooded Queensland’s mines. The quarterly contract hit $330, a record at the time. One weather system.

• In 2016, Beijing cut the number of days its mines were allowed to work from 330 to 276. Not a war. Not a disaster. A calendar change. The price went from $75 to over $300 in six months. Beijing relaxed the rule in November and the price came back down almost as fast.

• In 2020, China banned Australian coal for political reasons. By October 2021 the Chinese import price hit $410, an all-time high, while the Australian export price sat far below it. The same rock briefly carried two prices, split by a border and a grudge.

• In 2022, Russia invaded Ukraine. The benchmark went above $650.

At the peak of that last panic, the Australian government’s own commodity desk observed that met coal was moving more than $15 per day, “well above the typical level of under $1 per day.”

Fifteen years. Five detonations. The triggers had nothing in common. Weather, a calendar, politics, war. And notice that every spike collapsed nearly as fast as it formed. The 2022 record halved within a year. Nervous cuts in both directions, which is exactly why nobody can model this market. The structure never changed. Small market. Four sellers. No substitute. Thin spot. When this market is asked a question, the price does all the talking.

Why demand cannot blink

It is my favorite piece of industrial logic anywhere.

A blast furnace runs continuously for fifteen to twenty years. Inside it sits molten iron at fifteen hundred degrees. If the furnace cools, the iron freezes solid in the hearth. Steelmakers call the frozen mass a salamander, and they remove it with explosives. Then they rebuild the furnace. A full reline costs $300 to $400 million and months of zero production.

So the steel mill does not stop buying coal. It cannot afford to.

Walk the math with me. At $150 coal, the coal inside a tonne of steel costs about $115. At $600 coal, it costs $460, which is more than the entire normal cost of making the steel. In 2022 that exact scenario played out, and the result is the punchline of this whole post. European furnaces did go dark that year, because gas and electricity prices went up tenfold. Coal at four times its normal price did not cause a wave of permanent blast-furnace closures. Anywhere. American steel hit $1,955 a short ton that cycle, a record, so the mills passed the coal bill straight to their customers and the furnaces kept breathing.

Demand for this commodity did not flinch at a fivefold price increase. Which means when supply fails, there is no cushion. Nobody steps back. The price simply travels until it finds the number where the panic stops.

That is what nervous means, mechanically.

August 2026

Which brings us back to Shanxi.

Of the capacity switched off in May, 50 to 60 million tonnes was still offline in mid August. The consultancy CRU estimates China is missing a tenth of its domestic supply, “currently 10% of domestic supply” in their words, and expects the shortage to run to year end. Mines that reopened are running at seventy or eighty percent, because it is hard to set records with an inspector standing next to the longwall.

Coking coal futures in Dalian rose 46 percent in August. That is the largest monthly move since the contract began trading in 2013. The previous record, 38 percent, was set in July 2025. The two biggest months in the history of the contract have now happened within fourteen months of each other. The nervous market is getting more nervous.

And yet. The seaborne market moved only half as much. Australian premium coal fell ten percent to $215 on August 7 even as Chinese domestic coal sat near $296. By month-end, seaborne premium had recovered to around $270, up roughly 25 percent for August against Dalian’s 46 percent surge. Chinese low-sulfur coal ended the month near $356 ex-washplant. The raw gap narrowed, but remained close to $85, in a market whose entire history says gaps do not sit open.

How? Because China plugged the hole overland. In July, Mongolia and Russia together supplied 87 percent of China’s coking coal imports. Mongolia shipped 60 million tonnes last year and might reach 80 this year. But Mongolian coal is blending-grade material, not the premium coal that makes premium coke. China has also been draining its port stockpiles and simply making less steel, output is down three percent this year.

Every one of those cushions is finite.

And one cushion is already leaking. In June, Chinese imports of Australian coking coal rose 46 percent year on year. The base is tiny, Chinese purchases of Australian coal had collapsed by a third the year before, but the direction has changed. China is back on the water, selectively, for the premium tonnes Mongolia cannot replace.

Meanwhile the other buyer is standing at the shoreline checking its watch. India imports 95 percent of its coking coal, coal is 40 percent of its cost of making steel, and its imports rose 32 percent last year and another 15 percent through May. Indian mills have spent the monsoon season drawing down stockpiles instead of buying, because Chinese steel exports are crushing their margins and premium coal already averages $236 this year, a quarter higher than last. Priced out of Australian premium, Indian mills have been shopping the discount shelf instead, Russian cargoes, Mozambican cargoes, and American high-vol coal, which has spent all year trading at an unusually deep discount to the Australian benchmark. The discount shelf is where India planned to wait this out. The monsoon ends in September/October. Every year, on schedule, India returns to the spot market to restock.

On August 12, the discount shelf caught fire. Literally. At 2:15 in the morning, carbon monoxide alarms went off at the Longview mine in West Virginia, one of the larger American high-vol coking coal operations. Every miner got out. The coal did not. The mine has been sealed shut to starve the fire of oxygen, roughly three million tonnes of annual supply behind it, with no timetable for reopening. It is the second fire at the same mine in two years, and it burns in exactly the corner of the market where India shops.

One of Australia’s great premium mines, Moranbah North, suffered a gas ignition in March 2025 and has not cut coal since. Five million tonnes a year, dark for seventeen months. A hole in the premium shelf and a fire on the discount shelf, at the same time.

Even BHP, the most sober voice in the industry, wrote in August that strong Indian demand and supply disruptions have “tightened an otherwise balanced seaborne market.”

Gradually, then suddenly

The gap between Chinese and seaborne prices closes one of two ways.

Either the inspectors go home, Shanxi restarts, and the Chinese price falls back to meet the sea. That is the calm ending.

Or the restarts keep stalling, China steps onto the water for premium tonnes at the same moment India comes back from the monsoon, and two buyers ask a thin market one question at once, with five million tonnes of premium supply still missing and the American discount shelf sealed shut.

I do not need to predict which. The tape at the top of this post already told you what this market does when it is asked a question. A calendar change quadrupled it. A flood set a record. A ban created two prices for one rock. A war took it above $650.

Eighty two men died in Shanxi in May, and I will not dress that up as a trading opportunity. It is a tragedy, full stop. But it set in motion the oldest sequence in this market, the one that always starts slowly and never ends that way.

The seaborne price has moved. But next to China, it has barely twitched.

Gradually, then suddenly.

Cheers, Sandro

Get 20% off forever

Like and Restack. Apparently this is the only coal Substack won’t promote.🙂

“Only when the tide goes out do you discover who’s been swimming naked.”

— Warren Buffett

Quintin Kneen has watched the tide go out twice. The first time it took his company. The second time, he was already out of the water. This post is about what happened in between, and what he did with his own money near the top and after the first collapse.

May 17, 2017

Picture the federal bankruptcy court in Delaware on a Wednesday in May 2017. Two offshore supply vessel companies file for Chapter 11 that day. Not that week. That day, in that courthouse, seven docket numbers apart.

Case 17-11132 is Tidewater, the company that invented this industry, filing for the first time in 62 years. You read that story last week.

Case 17-11125 is GulfMark Offshore. Its shareholders will be left with 0.75 percent of the company and some warrants struck so far out of the money they might as well be lottery tickets. Roughly $430 million of bonds, borrowed while a man named Quintin Kneen was the CFO, get converted into ownership of the company he had been running as CEO since 2013.

Most executives do not survive a wipeout like that. Kneen kept his seat, collected a $510,000 retention bonus approved shortly before the filing, and steered GulfMark through the exit. Eighteen months later he merged it into its courthouse neighbor, Tidewater, and took the CFO job. He later called stepping down from CEO one of his greatest risks, because he believed the combined company would be the greatest franchise in offshore energy.

Ten months after that, Tidewater’s board moved on from its own CEO, paying him a without cause package on the way out, and gave Kneen the keys to the company that invented the industry.

What happened next is the part nobody argues about. Five year total shareholder return of 485 percent. Two completed acquisitions that look better with age, and a third pending. A balance sheet with near-zero net debt in a sector famous for dying of leverage.

In seven years as an officer of this company, Kneen has traded his own stock in the open market during exactly one calendar year. 2024. What he did in those eight months, and what the rest of the boardroom did in the same six week window, is the most complete portrait of insider psychology I have ever assembled. The full ledger is below, dated to the day, alongside the five tests and the grades.

Read more

“That was a new scheme that people kind of laughed about when they first saw it. It looked kind of strange, but it worked much better and it became the standard.”

— Alden “Doc” Laborde, on the boat that built this company

In the early 1950s, supplying an offshore rig in the Gulf of Mexico was a comedy performed with leftover war equipment. Companies used converted shrimp boats and surplus Navy landing craft. The pilot house sat at the stern, the cargo up front, which meant a captain had to hold the bow, the part of a boat hardest to control, steady under a swinging crane in open water. Men got hurt. Cargo went swimming.

A Navy trained engineer from Marksville, Louisiana looked at this and had a thought so simple it took an outsider to have it. Flip the boat. Put the wheelhouse all the way forward. Leave the entire deck behind it open and flat, back the stern up to the rig, and hold position with power.

His name was Alden Laborde. Everyone called him Doc. And this was already his second act. A few years earlier, as a drilling superintendent, he had designed the first transportable offshore drilling rig. His employer passed on the idea, so he quit, knocked on doors until Murphy Oil funded half, and built it anyway. The rig was called Mr. Charlie, and it went to work for Shell at what Laborde himself called the “incredible sum of $6,000 a day.”

Now he wanted to build the boat that would feed rigs like it. He gathered ten men, including his brothers, at $10,000 each. The boat cost $225,000 at a New Orleans yard, 120 feet long, 90 of them open deck. They named it Ebb Tide and launched it in 1955. The industry laughed at the strange backwards boat. Remember the laughter. A few hundred words from now it becomes an investment thesis.

Shell stopped laughing first. Before the Ebb Tide finished its first charter, Shell ordered an identical one. Phillips ordered the third. By the end of the first year the little partnership ran eleven boats. The Louisiana oil association would later call Laborde the offshore industry’s equivalent of the Wright brothers, because one man had now invented both the mobile rig and the vessel that supplies it. Years later a young Texan named George Bush came by his office asking how to get into offshore drilling, and Doc helped him start Zapata.

The backwards boat became the standard for the entire planet. Roughly three thousand working vessels today, on every offshore field on earth, carry Doc Laborde’s layout.

Share Cannibal Stocks

Seventy years later

The company those ten men founded is still here. It is called Tidewater, it counts its seventieth birthday this year, and it is still what it was in 1956, the largest owner of offshore support vessels in the world. Just over 200 ships, in more than 30 countries, average age 13 years, the youngest large fleet in the industry.

It survived every oil bust since Eisenhower. Then 2014 broke something even Laborde’s company could not absorb. After 62 years without a bankruptcy, Tidewater filed Chapter 11 in May 2017. The stock went from $59 to 68 cents. Shareholders kept 5 percent and some worthless warrants. Three thousand people lost their jobs.

So why am I writing about a 70 year old shipping company with a bankruptcy scar?

Because of one number. Four percent.

That is today’s global orderbook for these vessels, as a share of the existing fleet. At this point of the last cycle it was 34 percent. The industry that Doc Laborde invented has, for all practical purposes, stopped building his machine. And the company that owns more of these machines than anyone is priced at 40 percent of what its steel would cost to replace.

Below the line, the full arithmetic. Why nobody builds, with the three numbers that prove it cannot change before 2029. The math of what this fleet earns as day rates climb, worked out to the dollar per share. The one company that finally dared to order new boats, and what happened to their confidence since. Who is buying the stock. What kills the thesis. And one specific day rate number that would restart the world’s shipyards, which is also, and here is the twist, the number that tells you exactly when to sell.

Read more

“The real job of the board of directors is to come up with the right CEO and to prevent him or her from overreaching; if they do that job well, the rest will take care of itself.”

— Warren Buffett, 2004 Berkshire Hathaway Annual Meeting

Simon Johnson was there at the very beginning. He joined Seadrill in 2005, the year Fredriksen founded it, and spent five years selling its rigs while the empire swelled toward $22 billion. He left before the fall and watched both bankruptcies from other companies’ windows. Then, in March 2022, one month after the second wipeout, this board fired the CEO who had steered it through the courtroom, Stuart Jackson, effective immediately, and handed the wreck to Johnson.

He sold the junk. He shrank the share count by a fifth. He signed the billion dollar Petrobras contracts that are paying for this year’s recovery. On February 25, he hit his own guidance and told the market Seadrill enters 2026 from a position of strength.

Fifteen days later, the same board did it to him. Same month of the year. Same two words. Same chair, Julie Robertson, who has run this boardroom since the day the new Seadrill was born. No transition. No reason.

The severance agreement runs for pages. Salary continuation. An earned bonus. Forfeited equity. And near the bottom, one line I have not been able to forget. The company that paid Simon Johnson $27 million over three years agreed to ship his household goods back to Australia and fly his family home. In economy.

Most companies change CEOs the way diplomats change governments, with six month transitions and flattering press releases. Seadrill changes CEOs the way you change a lock.

A few days ago, I showed you a company that can no longer die the way it used to.

Today I answer the question that ending left open. Who actually runs this thing. Same five tests I ran on Copart’s two CEOs. How the men are paid. What they did with their own shares. What they did with your money. What they said when it went wrong. And what they did when they thought nobody was reading. Full scorecard at the bottom. No mercy.

One number before the line, free of charge. Seadrill has been a public company for four and a half years. In that time, the number of shares bought on the open market by its own insiders, any insider, any amount, is zero. Not one CEO, not one director, has ever paid cash for a share of the company they run.

Below the line, it gets stranger. This June, days before the company started buying back its own stock, two of its top executives were selling theirs.

And further down, inside the new CEO’s employment agreement, sits a clause I think explains everything on this page, including why Johnson had to go. It triples his successor’s severance in exactly one scenario. Before you scroll past the line, try to guess which one.

Read more

“I am confident that Seadrill will emerge from this downturn even stronger and that we will resume our distributions in the future.”

— John Fredriksen, November 2014

Less than three years later, Seadrill was bankrupt.

“Having restructured our business, we are now well positioned to capitalize on the recovery.”

— Anton Dibowitz, Seadrill CEO, November 2018

Just over two years later, Seadrill was bankrupt again.

Rewind to 2013, when promising was easy.

John Fredriksen owned the most aggressive drilling company on earth.

Sixty nine rigs, twenty one of them still sitting in shipyards. A $19 billion backlog. A market value around $22 billion. The Norwegian tanker billionaire told investors Seadrill would remain a high growth, high dividend stock. He meant it. Both halves. At the same time.

The company paid $1.4 billion in dividends in 2013. Another $1.5 billion in 2014. While it still owed the shipyards another $7 billion for rigs that did not exist yet. I need you to hold those two numbers next to each other. Please. Three billion out the door to shareholders. Seven billion in invoices still in the mail.

Oil had been falling for five months when, on November 26, 2014, the board suspended the dividend. The stock lost 23 percent in a day. Fredriksen’s promise at the top of this page comes from that exact announcement. The distributions never resumed. Not to this day.

By April 2017, Fredriksen was telling a Norwegian newspaper he was working 18 hours a day and did not have a minute to waste. Five months later, on September 12, 2017, Seadrill filed Chapter 11 in Houston. An $895 million bond was due that same week. Twelve point seven billion dollars of debt sat on the table.

The old shareholders walked out of that courtroom with 1.9 percent of the new company.

Some of them stayed. New investors came in behind them, and who can blame them. The new CEO was saying it out loud. Restructured. Well positioned to capitalize on the recovery. That is the second promise at the top of this page.

Less than four years later, Seadrill was back in the same courthouse.

The second restructuring cut $5.6 billion of bank debt down to $683 million. When the company emerged in February 2022, everyone who had held the stock through it, including every investor who bought after the first bankruptcy believing lightning does not strike twice, owned 0.24 percent of the new Seadrill.

Not 24 percent. Zero point two four.

Twice in five years, Seadrill turned its owners into a rounding error.

Share

The third Seadrill

The rigs never sank. Through both bankruptcies, Seadrill’s ships kept drilling, kept getting hired, kept being some of the best assets on water. What died, twice, was the balance sheet wrapped around them.

The company that exists today is the third attempt, and it is built like a survivor of two heart attacks. About 62.5 million shares. Gross debt of $750 million against $360 million of cash. Net debt of $390 million, which is less than one year of the EBITDA it just guided to. A $2.9 billion contract backlog. A share count that has shrunk 22 percent since September 2023 because the company has been eating itself.

Regular readers know where my own offshore money sits. I own Transocean, the leveraged bet on this cycle. My wife asked when we sell. I said when the thesis plays out. We have not spoken since July. 🙂 I wrote about the cycle itself that month.

Seadrill is the opposite corner of the same trade. Transocean is the biggest fleet with the biggest debt. Seadrill is the cleanest balance sheet in the industry attached to seven of the youngest drillships on earth.

And right now it trades at roughly 7.5 times this year’s EBITDA guidance and about a third of what its fleet would cost to build.

Below the line, I will show you the seven ships that matter, the supply math that cannot unwind before 2030, the 2027 earnings arithmetic step by step, who is quietly buying, and the two things that can still kill this. One of them involves Petrobras and a $213 million invoice from a ghost.

Read more

“I can live with a slightly lower return on equity and lower earnings-per-share growth, but at least I will live.”

— Pulak Prasad

Look at the chart before you read a word.

Alpha sold 3.5 million tons of coal in Q2 for $493 million of revenue. Warrior sold 3.7 million tons for $510 million. Then the paths split. Warrior kept $157 million of EBITDA and $103 million of free cash flow. Alpha kept $26 million of EBITDA and lost money.

Same coal. Opposite outcomes. This post explains where the money goes.

Get 20% off forever

One note on timing. I wrote this last week, with Alpha near $157. It closed at $194 on August 19, up almost 15 percent in a single session. Every price in this post is from the time of writing, so the discount is narrower than the arithmetic below suggests.

Warrior is the better business

This is not close, and pretending otherwise would insult you.

Start underground. Warrior mines with longwalls, giant shearing machines that cut coal at a fraction of the labor per ton.

Alpha runs mostly room and pillar sections, which need more people and more machines for every ton.

That single difference puts Warrior in the bottom quartile of the global seaborne cost curve. In Q2 the gap was $92.53 against $103.07 per ton, and it persists through every cycle.

Then the path to water. Warrior reaches the port of Mobile by rail or by barge. Two options mean pricing power against the railroad. Alpha depends on a single railroad to reach the East Coast. Warrior’s inland freight at $15 to $30 per ton against $25 to $60 for Central Appalachian producers. Cost is measured when the coal touches the water, and Warrior touches it cheaper.

Then the growth. Warrior just brought Blue Creek online, a brand new longwall mine, and raised 2026 guidance to 13 to 14 million tons because customers are taking the coal as fast as it comes. At benchmark prices near the cycle floor, Warrior’s own March math pointed to more than half a billion dollars of annual cash flow. In the worst met coal market in years, Warrior prints money. Alpha, in the same market, idles mines and defends a balance sheet.

Case closed. Better mines, better logistics, better costs, better cash flow. Warrior is the wonderful business.

Now the twist that binds these two companies. Blue Creek, Warrior’s crown jewel, is a high vol A mine. It belongs to the roughly 11 million tons of new longwall supply that broke the price of exactly the coal Alpha sells most of. The spread between premium low vol and high vol A went from $5 per ton in early 2025 to over $30 today. Warrior’s own realization collapsed to 66 percent of the benchmark under the weight of it, down from 80 percent a year ago.

The best business in the sector built the flood its neighbor is drowning in. Warrior can afford the flood. Alpha cannot, which is why it trades like a company in trouble.

So buy Warrior and move on?

Not so fast. A stock is not a company. A stock is future cash flow divided by a share count.

And the moment you divide, the picture inverts.

Read more

On June 14, a storm hit Newport News, Virginia. At the Dominion Terminal, where Alpha loads its coal onto ships, the wind bent one of the two giant stacker reclaimers into scrap.

Six weeks later Alpha reported its third quarterly loss in a row. The stock sits around $160, down from $255 in the past year. And this week the 13F filings landed, showing that Mohnish Pabrai trimmed his AMR position by 11 percent.

Is Alpha in danger, or does this just take time?

I went through the 10-Q, the annual report and the technical reports to answer exactly that.

Give a gift subscription

How much cash is Alpha burning?

In the first half of 2026 the business generated $68.9 million from operations. It spent $85.8 million on mines and equipment and $23.3 million on its share of the damaged terminal. So the operation burned about $40 million in six months. Management spent another $36.7 million on buybacks, but that part is voluntary.

For the full year, the burn depends on one variable. Alpha still has 4.1 million tons committed to customers but not yet priced. Every $10 per ton on those tons is about $41 million of revenue, and most of it falls straight through to cash. Price them near $115 and the full year burn approaches $90 million. Price them near $135 and the second half is roughly cash neutral.

Can they afford either outcome? Alpha holds $338.5 million in cash and short term investments against total debt of $11.4 million. The bank agreement requires a minimum of $75 million, so the usable cushion is about $263 million, with another $184 million of untouched credit behind it until 2029. The worst scenario above costs a third of one year of that cushion.

The legacy obligations people fear with coal companies are also visible and small. Alpha projects reclamation and pension outflows of $44.6 million in 2026 and $49.4 million in 2027, then roughly $25 million and $22 million in the two years after.

So there is no solvency question. The question is how fast the burn goes to zero.

What Alpha needs to do now

Everything in the next 18 months should serve one target. Bring the cash burn to zero without selling coal at giveaway prices. And most of the moves are already in motion.

Keep cutting the losing tons. Alpha has already idled the Elk Run complex and the Long Branch surface mine and reduced output at Jerry Fork and Black Eagle.

Management ranks mines by margin, not by cost. If high vol prices stay here, expect more cuts. A ton sold at a loss is worse than a ton left in the ground, and the ground stores it for free. Keep routing coal to the best outlet. Their weaker high vol B currently earns roughly the same money in the thermal market, so it goes there.

Ramp Wildcat. The new mine reached production this year and builds through the second half. It mines low vol coal from the Sewell seam, and in the independent engineering models it carries the highest margin of any underground mine Alpha owns, over $40 of EBITDA per ton. Every Wildcat ton that replaces a high vol B ton upgrades the whole mix.

Collect the money already owed. Coal inventory swelled from $193 million to $262 million in six months, largely because the broken terminal slowed shipments. That is $69 million of cash sitting in coal piles, coming back as logistics normalize. The terminal insurance claim is filed too. No amount is disclosed, but Core Natural Resources, Alpha’s partner at the same terminal, just settled a comparable storm claim for its full $154.5 million policy limit, in cash, within months. I assume zero insurance money everywhere in this post. Whatever arrives is upside.

And bank the tax credit. Met coal was added to the federal list of critical minerals, which gives Alpha a refundable production credit worth 2.5 percent of production costs through 2029. That was $14.3 million in the first half alone, $28 million a year, paid even in loss years.

Add it up. Cuts shrink the losses, Wildcat upgrades the mix, inventory and insurance return cash, the credit covers most of the rest. A company burning $40 million per half year has a realistic path to roughly zero burn during 2027 even if coal prices do nothing.

And if the burn is near zero, the balance sheet answers the duration question. Alpha can operate in this mode more or less indefinitely. Not comfortably. Not profitably. But indefinitely. For contrast, Coronado Global Resources sits at $98 million of liquidity against roughly $700 million of debt, with its newest loan at 14 percent interest.

Alpha’s problem is patience. Coronado’s problem is oxygen.

Read more

"Commodity businesses have risk unless you're the low-cost producer."

– Warren Buffett

In part one I showed you the thesis. A modeled gas shortage arriving 2028, roughly 70 percent odds of a serious squeeze, and a game you can win on both sides of the coin.

Three companies passed my screen. Lowest costs. Cash returned to shareholders. Open exposure to 2028.

Today they lie on the table.

Gloves on.

Read more

June 2003. Alan Greenspan sits before Congress.

The most powerful economist on Earth has come with a warning. America is running out of natural gas. Storage is low. Prices have tripled. Fertilizer plants are fleeing to Trinidad.

His solution was to build import terminals and bring gas in from Qatar on ships.

America listened. Billions flowed to the Gulf Coast. Sabine Pass, Freeport, Cameron. Giant facilities rose along the water, built for one purpose.

To receive.

Then a stubborn son of a Greek immigrant named George Mitchell cracked the code on shale rock in Texas, and American gas production did something no forecaster had on paper. It doubled. Prices collapsed and stayed collapsed for fifteen years.

Every one of those import terminals was reversed. The pipes built to flow in now flow out. Sabine Pass, constructed because America was running out of gas, is today the largest gas export facility in America.

The men who bet on the shortage had Greenspan. They had the data. They had storage charts that looked terrifying.

They were wrong anyway.

Hold that story in your head. Because a man has appeared with the most detailed gas shortage call I have ever read.

And this time, I cannot fully dismiss it.

Share Cannibal Stocks

The letter

Matthew Smith runs a fund called Chronometer Partners. In June he sent a letter to a small circle of investors. It found its way to me, and I have now read it more times than I want to admit.

Matthew Smith’s June 2026 letter laying out Chronometer Partners’ natural gas shortage thesis.

Smith spent fifteen months building a model of the entire American natural gas system. Every producing well, fit to its own decline curve. Every pipeline. Every storage facility. Every LNG terminal. Every power plant and every data center. All of it in one database, connected the way the real system is connected.

Here is what it said, argument by argument.

Argument one. Supply has a ceiling.

America produces about 112 billion cubic feet of gas per day. Smith’s model says the absolute maximum the country can reach by the end of 2030 is 132.

“That is the maximum we see.”

Why a ceiling? Because the myth of infinite shale belongs to the past. The best rock has been drilled. What remains is deeper, worse, and more expensive. Many producers, he claims, have far less quality drilling inventory left than they publicly admit. The gas that does remain needs much higher prices to be worth drilling. And even where it exists, there are no pipelines planned to carry it.

Argument two. The exports are already sold.

America currently exports about 15 billion cubic feet per day as LNG, liquefied gas on ships. Approved projects take that to 35 by the end of 2030. More than doubling. These terminals are financed, under construction, and their gas is already sold to foreign buyers on contracts that run twenty years. Japan, Korea, Europe. The contracts were used to finance the construction itself.

That gas is leaving the country. It has, in effect, already left.

Argument three. AI eats gas.

Every data center needs power around the clock. The sun sets and the wind stops, but a chatbot answering questions at 3 a.m. does not. Only two sources deliver that at scale. Nuclear, which takes fifteen years to build. And gas, which takes three.

Smith counts more than 5 billion cubic feet per day of new gas demand from approved power plants alone, most of them built to feed AI. Two to three times more have been proposed. He did not even include those.

Argument four. The infrastructure cannot be built in time.

Suppose the gas exists somewhere. You still need pipes to move it. Building major gas infrastructure in America takes five to eight years. The country has completed exactly one large interstate gas pipeline in the past decade, and that one needed almost ten years and an act of Congress.

If the pipes needed to solve a 2028 problem were coming, they would be under construction today.

They are not.

Argument five. The cushion is gone.

Storage is the system’s savings account. When winter hits, America lives off gas injected underground during summer. In the last fifteen years, that savings account grew 7 percent.

Demand grew 54 percent.

In 2010, storage could cover 72 days of key demand. Today, 50. By 2030, on Smith’s math, 37. The system got bigger. The cushion got smaller. And nobody is building more.

Put it together. A supply ceiling. Exports that doubled and cannot be recalled. An AI power boom. No new pipes. A shrinking savings account. Smith’s model shows a deficit of more than 5 billion cubic feet per day by 2030, storage draining in a way that has no precedent starting in 2028, and working storage potentially hitting zero by 2030.

Not a price increase. A shortage.

Follow the buyers

Smith ends his letter with a challenge I respect. Do not take my word for it, he says. Ignore me. Watch what the smartest money is doing.

Shell paid 16.4 billion dollars for ARC Resources, one of Canada’s top gas producers. Mitsubishi paid 7.5 billion for Texas and Louisiana shale gas. Williams is paying up to 5.5 billion for Momentum Midstream, a gas pipeline system. Devon received an 8 billion dollar offer for its Marcellus gas position. Antero bought HG Energy for 2.8 billion. EQT bought Olympus for 1.8 billion. And the Financial Times ran a piece on how Citadel, the most profitable hedge fund in history, has built itself into an energy trading giant.

All of this in roughly the last twelve months. All of it gas.

One more data point, and it is my favorite. In January, one winter storm hit the system. Weekly gas prices brushed thirteen dollars. EQT, largely unhedged, made about a billion dollars in a single month.

Today gas trades at 2.88.

Thirteen dollars and 2.88 in the same year. That is what a market without a cushion looks like. The crisis Smith predicts for 2028 already sent a postcard in January.

Give a gift subscription

I tried to kill it

Every good investment starts with the risks first. So before I let this thesis anywhere near my money, I want to kill it.

The first weapon is history, and you already met it. It was the opening of this post. The 2003 panic was so convincing that America built import terminals, and shale turned them into the punchline of the century.

It was not the only time. High prices from 2004 to 2008 summoned the shale revolution that crushed gas for fifteen years. The 2022 spike above nine dollars died within nine months. Every structural gas shortage call of the last twenty five years has ended the same way, with new supply showing up faster than anyone modeled.

Betting on American energy scarcity has been the fastest way to lose money for a generation.

The second weapon is geology. Independent data from Enverus counts over 800 trillion cubic feet of American gas resource with breakevens below three dollars, and hundreds more above that. The gas is not gone. It is waiting for a better offer. Smith’s ceiling of 132 is real, but it is a price wall, not a rock wall. Give this industry five dollar gas and two years, and it will embarrass every model ever built. It always has.

The third weapon is the escape hatches. About 145 gigawatts of coal plants will still be standing in 2028, and when gas spikes, utilities switch back to coal and demand falls fast. And those sacred LNG contracts? Contracts are sacred until a government watches households pay triple for electricity in an election year while a quarter of the country’s gas sails away on ships. Some of those exports would be slowed or renegotiated long before storage hits zero.

So no. I do not believe in empty storage and twenty dollar gas.

Where I land

Numbers, because opinions are cheap.

I put the probability at roughly 70 percent that Smith is directionally right. That the market tightens hard around 2028 and gas prices must jump, meaningfully, for a window of about two years before the machine catches up.

Why 70? Three reasons.

First, this demand is different in kind from 2003. Back then the shortage lived in spreadsheets. This time it lives in poured concrete, financed terminals, twenty year contracts, and turbine factories that are sold out through 2030. You cannot cancel demand that has already been built.

Second, the cushion that saved every previous scare is gone. Storage grew 7 percent while the market grew 54. January showed you what one cold week does now.

Third, and this is the almost poetic part, today’s low price is manufacturing tomorrow’s squeeze. At 2.88, producers are shutting wells and starving investment. The weaker this year gets, the tighter 2028 becomes. The cure for low prices is low prices. It just needs time.

And why only two years? Because that is the other side of the same coin. It takes the industry roughly two years to answer a price signal with rigs, wells, and coal switching. The squeeze is real, and so is its expiration date. That is why every spike in gas history has been violent and temporary, and I expect this one to be the same. Just bigger, because the cushion is smaller.

The market is slowly waking up. When Smith wrote his letter, 2028 gas was priced around 3.60. As I write this, January 2028 trades near 4.67.

The curve is walking toward him.

The game

“Heads, I win. Tails, I win.” – Mohnish Pabrai

You do not need to bet on the shortage. That is the whole trick.

You need companies that win both sides of the coin. They exist, and the test for finding them is three questions.

• Do they make real money today, at 2.88, at the bottom of the cycle?

• Do they hand that money back through buybacks and dividends instead of drilling it away?

• And is their 2028 production unhedged, unsold, and waiting?

I ran every meaningful gas producer in North America through those three questions. Most failed. But three passed.

Heads. Smith is right. Gas jumps in 2028 and these three catch it with open books across hundreds of billions of cubic feet. Cash flow doubles or triples, and every share they bought back at today’s prices multiplies the effect. Hallelujah.

Tails. Smith joins Greenspan in the museum of gas shortage calls. Prices stay low. These three keep printing money anyway, because their costs sit near two dollars, and they keep swallowing their own shares at depressed prices, which is exactly when swallowing shares creates the most value. You do not lose. You compound while you wait.

Either way, you get paid. The biggest risk now is picking the wrong one of the three.

The autopsy table

Which is exactly why part two exists.

I will open the hedge books year by year, because the hedge book decides who actually catches the spike and who watches it through a window. I will read the insider filings and show you who sold at the bottom and who went silent. I will dig up a four billion dollar corpse one of these companies buried in Louisiana, and check whether the men who held the shovels still work in the building.

And at the end, one answer. The best bet in the entire gas market.

If you want the answer, upgrade before the autopsy begins.

And bring gloves.

Get 20% off forever

Back to 2003

Greenspan was not wrong about demand. He was wrong about what price does to supply, and the men who built the import terminals lost fortunes because they made a one way bet on scarcity.

I am not making that bet. I am buying companies that get paid at 2.88 and get rich at six.

The coin is weighted. Both sides pay.

Flip it.

Cheers, Sandro

Like button costs less than natural gas. For now. If gas hits $13 again, I’m charging for likes. Act now. 😂⛽

Updated as of August 9, 2026. Sources include EIA, company SEC filings, Enverus, Wood Mackenzie, GE Vernova disclosures, and Bloomberg.

"I'm just not interested in explaining to them that we went broke because there was a 0.01% chance that we would go broke."

- Warren Buffett

Every morning before the market opens, I read Value Investors Club. Most days it is a graveyard of sensible ideas. Banks at eight times earnings. Insurance rollups. I drink my coffee and move on.

A few days ago, I read something different. I read it once. Then again. Then I stood up, told my wife I should go to Kurdistan.

She had a point. Then she asked why I had Kyrgyzstan open in the other tab.

That is not a joke about her. I had spent the first forty five minutes researching the wrong continent myself. Kurdistan. Northern Iraq. Easy mistake. Expensive mistake, if you book the flight. 🙂

Why Kurdistan?

There is a company in London called Gulf Keystone Petroleum. It costs $500 million today. Since 2019 it has mailed shareholders $512 million in dividends and buybacks. The company costs less than the checks it has already sent.

The stock costs $2.30 per share. It has no debt. It has 27 years of reserves in one giant field. In its best year it produced $266 million of free cash flow, more than half its current market cap, in twelve months.

And the state owes it money. $122.8 million for oil delivered in 2022 and 2023. That works out to $0.55 per share still owed to the company — a quarter of today’s share price. On top of that, Gulf Keystone already has $0.30 per share in cash sitting on the balance sheet. And there is more. Since last September, it has been paid only $30 per exported barrel while invoicing $51. The unpaid difference keeps accumulating as additional money the government owes the company.

The field lifts one barrel of oil for $4.30.

Saudi Aramco, the most profitable company ever built, put its lifting cost at $3.50. The North Sea needs mid teens. Oil sands need twenty and up.

A small London stock, in northern Iraq, pulls oil out of the ground at a cost that belongs on the same chart as Saudi Arabia. Not the same economics, their crude is lighter and mine is heavy and sour, but the same chart. At these costs, you can stop caring what oil does. This field makes money at numbers that bankrupt Texas.

So yes. Pack your bags. We are going.

Share

Welcome to Kurdistan

Put your phone away, the roaming is terrible anyway, and look out the window.

That city below us is Erbil. People have lived on that Citadel for six thousand years. It was old when the pyramids were new. Down there you will find snow capped mountains, waterfalls, glasses of sweet tea that never stop coming, and hospitality that will embarrass everything you know about Europe. In spring these hills turn green and half of Iraq drives north just to look at them.

I can talk all day. But the single most important piece of due diligence on this trip is not something I can tell you. Because you are not just buying an oil field. You are putting your money into a place, a culture, a political system, and a civilization with thousands of years of history behind it.

This video tells you where your money is actually going.

Jokes aside 🙂

That post from VIC might be the best pitch anyone has ever written. Which is exactly why I did not trust it. So I try to destroy it. Here is your itinerary for the paid tour.

• Stop one, the hill. A field so good it survived its own company nearly dying.

• Stop two, the border crossing. Where I found the clause the pitch skips.

• Stop three, the five receipts. I built a table to prove that oil prices drive this stock.

• Stop four, we stop pricing the company in pounds and start pricing it in years.

• Stop five, we count craters.

The rest of the tour is for paid passengers only. The ticket costs less than the taxi from Erbil airport.

Read more

Four years later the company was worth zero.

August 2015. Chapter 11. Shareholders wiped out completely. A Fortune 500 company with 13,000 employees, gone.

But the mines did not die. Creditors scooped the best assets out of the wreckage and named the new company Contura. In 2021 Contura did something strange. It took the dead name back.

Alpha Metallurgical Resources.

The old Alpha. In this industry companies die but the coal survives. When you buy AMR you are buying mines that outlived every owner they ever had.

And that graveyard did two things.

Outside the company, capital fled Appalachia and never came back. Banks stopped lending. Supply can barely respond, even at much higher prices. New mines require scarce geology, years of permitting, rail and port capacity, skilled labor, and capital that has largely abandoned Appalachia.

Inside the company, the people who crawled out of that bankruptcy learned the lesson in blood. Never again a 7 billion dollar acquisition. Never again an empire. Since 2022 this management has spent its cash on exactly one asset. Its own shares. Roughly a third of the company, eaten. The float is collapsing. My readers know this very well.

The company that died buying others is getting rich buying itself.

AMR just reported its second quarter. How many shares are left?

Related AMR research

Read the complete AMR long-term investment thesis

Read why AMR is squeezed by the cycle, not broken

Read more

No production that day. Just a maintenance crew, two thousand feet underground in Mine No. 5.

5:20. Evening. A roof collapse released methane. The first explosion injured several miners. One of them, Gaston Adams, was too badly injured to move. Twelve miners moved toward the damaged section to bring him out. Charlie Nail was trying to place his rescue mask on Adams.

6:15. The second explosion. Thirteen men died. Gaston Adams and the twelve men who tried to save him.

I need you to understand where this happened. Please.

It happened in the Blue Creek seam. The same seam Warrior Met Coal mines today. Mines No. 4 and No. 7 sit between 1,400 and 2,100 feet underground, some of the deepest vertical shaft coal mines in North America. Men still ride that same three-minute cage down every morning.

Why does anyone dig that deep for coal? Because this is not coal. Not the kind you burn.

Steel is made by tearing oxygen away from iron ore. The thing that does the tearing is coke. Baked metallurgical coal, almost pure carbon. Inside a blast furnace, coke is not just fuel. It strips the oxygen from the ore and holds the entire column open so the furnace can breathe. Nothing else does all of that at once. Not electricity. Not gas.

And coal that makes strong coke is geologically rare. Blue Creek coal has some of the lowest sulfur on Earth and coking properties that match premium Australian hard coking coal. That is why steel mills in Japan, Brazil and the Netherlands call Alabama.

A new mine like this takes five to ten years to build. Meanwhile the price sits near the floor of the cycle. At these levels, Central Appalachian mines are cash flow negative. Some are already closing.

Warrior mines the same commodity, at the same prices, and just reported its second quarter. I read the full earnings release. I listened to the entire call. The full breakdown is below.

Share Cannibal Stocks

Twelve men went back down that shaft for one of their own. The least we can do is understand what they were digging for.

Read more

Not about the merger. Not about the dayrates. Something much bigger.

For years I preached the same gospel. America has infinite cheap gas. Texas is the well that never runs dry. The offshore bet never needed shale to die. It only needed deepwater molecules to be cheaper. Shale dying was never part of the thesis.

One analyst spent eighteen months mapping every producing gas well in America. Every single well. Its location, its decline rate, its pipes.

The math is that America produces 110 billion cubic feet of gas per day. The physical maximum is 130. The export terminals already under contract will take those 20 by 2030. Every cubic foot of growth is already spoken for.

Then AI shows up asking for more.

His conclusion. By mid 2028 American gas storage breaks below anything in recorded history. By 2029 it drops below all known evidence. His words for what follows. A knife fight for physical gas. And shale burns through its core inventory in four to five years. The weapon that murdered offshore in 2014 is bleeding out.

Now follow the chain. American gas doubles. American LNG loses its edge. The world starts hunting for gas that is not American. That gas sits under the ocean. And the rigs that reach it take five years to build. Nobody is building any.

To be clear. The offshore thesis was built on oil and it still stands on oil. Scarce rigs. A dead shipyard cycle. Dayrates that have to rise. None of that changes. The gas is a second customer knocking on the door of the same fixed fleet. Demand I never counted.

One variable is left to balance that market. The dayrate.

So yes, I was damn wrong. My thesis was too small.

Transocean just reported its second quarter. The market is still pricing Transocean as if Texas never runs dry. Now let’s look at what the results and the earnings call actually revealed.

Read more

His answer separated two things. There is business risk, the chance that the company itself fails. And there is price risk, the damage you inflict on yourself by paying too much for a company that never fails at all. Overpaying for a wonderful business, he said, is usually “a risk of time rather than loss of principal.”

You do not lose your money. You lose your years.

No stock of the last three decades proves the point like Copart. The business compounded at over 21% a year, grew revenue in 27 of 30 years, and never once gave anyone a cheap entry. It punished two kinds of investors the entire time. The ones who paid any price served Buffett’s sentence and lost years waiting for the earnings to catch up. The ones who waited for a discount never got in at all.

A fantastic company. And for long stretches, at 35 and 40 times earnings, a bad investment. Those are two different things, and confusing them is the most expensive mistake in quality investing.

This is the post where I find out which one Copart is at $29.

There is a ritual on Copart earnings calls. An analyst asks what return the company earns on its land. Half a billion dollars a year of shareholder money, going into dirt. Management gives the same answer every time.

“We don’t know. We don’t calculate it.”

They will count every car in every yard on every continent for you. On this one number, the most transparent company in the industry goes silent. Book value of the land, $2.39 billion. Real value, unknown.

So I calculated it. What I found changes the entire question, because at a certain land value, the price risk Buffett described does not shrink. It inverts. Then I reverse engineered the growth rate the market charges you at $29.

In Part 1, a farmer won an auction because he counted two thousand cars while the men in suits bid by hunch.

In Part 2, I graded the men who run the company.

This week, I counted the acres.

Read more

“You can’t make a good deal with a bad person.” — Warren Buffett

In 1989, a nineteen year old walked into Willis Johnson’s office with a proposal.

He was dating Willis’s daughter. He wanted to drop out of college and learn the business instead.

Understand who he was asking. Willis Johnson had crawled through junkyards since he was fourteen. He had survived a year in Vietnam walking point ahead of an entire company. He had sold his house and moved his family into a trailer to buy his first yard. Nothing in his life had ever been handed to him, and nothing would be handed over now.

Willis said fine.

Then he handed him a broom. The kid swept floors, drove forklifts, untangled DMV paperwork, the least glamorous job in the least glamorous industry in America.

His name is Jay Adair.

On July 31, 2026, thirty seven years after the broom, he became chief executive of Copart for the second time. The company is worth $28 billion. He will run it for a salary of one dollar a year.

He replaces Jeff Liaw, the polished Harvard MBA the board spent a decade grooming. Liaw lasted 27 months as sole CEO. The stock fell 38% in his final year. Three days before walking out, he sold another $845,945 of stock.

Two men held the same wheel, at the same company, in the same decade. That means I can run both through the identical five tests I use on every management team I own, and nothing can hide behind different industries or different eras.

I wrote a piece on why Copart is one of the greatest businesses ever built.

This post grades the men driving it. There is a full scorecard at the bottom, four names, five tests, no mercy.

Continue to Part 3: What Is the Copart empire worth now?

Read more

The most important second of your car’s life happens without you.

Your car is hit. The adjuster shows up two days later. He does not walk around the wreck with a clipboard. He opens a laptop.

The software already knows. It knows what 7.6 million comparable vehicles sold for. It knows the parts prices this week, the labour rate in your county, 41,000 municipal tax rates, 53,000 state fees. It has watched every collision repair in America for forty five years.

It runs one comparison. Cost to fix, against what the car is worth.

A number appears. Your car is dead. You had nothing to do with it.

That software is made by a company called CCC. You have never heard of it. Twenty seven of the thirty largest auto insurers in America run on it. So do thirty thousand body shops. A hundred billion dollars of claims cross it every year.

Few days ago I wrote about the company that hauls the corpse away.

This is the company that signs the death certificate. Two businesses, billing in the same second. One computes the verdict. The other sells the body.

Chuck Akre owns both.

And he is bleeding on both. He paid an average of $41.42 for Copart, which trades at $27. He paid an average of $10.09 for CCC, which trades at $5.82. Down on one, down on the other. He funded both by selling Mastercard, Visa and Moody’s near their all time highs.

Last quarter he bought more of both.

Then, on July 9, Reuters broke a story. CCC has hired Morgan Stanley and is calling private equity firms about selling the entire company.

Charlie Munger had a favourite question for moments exactly like this. When you find an extreme case that makes no sense on its face, he said, you stop and ask it out loud.

“What in hell is going on here?”

I ran CCC through its own algorithm. Cost to fix, against what it is worth. Every section below is one field on the claim form.

The verdict prints at the bottom. It is not the one I expected to write.

Read more

Northern California. The mid 1960s.

A wrecking yard with two thousand dead cars on it. The state is putting a highway through, so everything has to be gone in ninety days. The whole lot goes up for sealed bid.

A farmer sees the ad in the paper. He drives out with his fourteen year old son.

They spend the whole weekend crawling through the trees. The boy marks every car with an X. Complete cars in one count. Stripped ones in another. Two thousand cars. They count them all.

Share

Monday morning. The bidding room.

The other men arrive in business suits. They spent the weekend at home. The farmer arrives in overalls with a filthy kid beside him.

Everyone drops a sealed envelope into a hat in the middle of the room.

The farmer wins at fifteen thousand dollars. A brand new house cost twelve thousand that year. The men in suits stare at him.

Before he can leave the room, one of them walks over and offers him quick cash to hand over the deal. Real money, for a weekend of walking through a junkyard. The farmer says no.

Outside, the boy asks his father why he turned down easy profit.

He’ll come back tomorrow, the father says. And he’ll pay more.

The man came back the next day. He paid more.

The boy never forgot what he saw in that room. The men in suits were not stupid. They were rich, connected, experienced. They lost because they decided a yard full of wrecks was not worth a weekend of their time. They bid on a hunch and assumed nobody would outwork them.

The boy’s name was Willis Johnson.

He would spend the next sixty years doing the exact same thing, on an exponentially larger scale, until his junkyard covered more land than Manhattan and was worth twenty five billion dollars.

That company is Copart. And right now, the market is bidding on it.

The stock is down 50%. Wall Street has decided that self driving cars will end car crashes, that salvage is a dying industry, and that a debt free business with 36% operating margins should trade at the same multiple somebody paid to take out its weaker competitor.

The market is bidding by hunch.

But a few people counted the cars.

Chuck Akre built one of the great compounding records in American investing. An average holding period of eight years. A philosophy that has not moved an inch since 1989. In the second half of 2025, his firm started selling the positions that made its name. Mastercard. Moody’s. Visa. O’Reilly.

It used the money to buy 8 million shares of a junkyard, at an average cost of $41.42.

The stock trades at $28 today. Akre is sitting on roughly $107 million of paper losses.

In the first quarter of 2026, the firm bought more.

At the same time, the company itself spent $1.6 billion devouring its own stock. The largest buyback in its 43 year history. And on July 31, a man who has worked for a salary of one dollar since 2020 returns as chief executive.

Three buyers who counted. One market bidding by hunch.

What did they count? A fee structure, written in the 1980s, that quietly manufactures this company’s own supply. A land bank assembled before the cities arrived, at prices that no longer exist anywhere on earth. And a twenty year trend that Wall Street believes is weather, but that one man built by hand.

Let me show you the count.

Continue the Copart series

Part 2: Who Really Runs Copart? The leadership analysis

Part 3: What Is the Copart empire worth now?

Read more

On February 2, 2016, the company filed for bankruptcy. The stock closed at 11 cents.

Somewhere between those two numbers, Mohnish Pabrai, the man who built his entire reputation on a checklist designed to catch other investors’ mistakes before he could repeat them, rode more than six million shares almost all the way to zero.

He lost roughly $100 million. His flagship fund finished 2015 down 30 percent. When he finally sold, he got 28 cents a share.

This was not a fraud. There was no hidden accounting scandal, no overnight collapse, no lie that only insiders knew. Horsehead told the public it was dying, in official filings, every single quarter, for eighteen months straight. Short sellers called the bankruptcy in writing, with a date on it, four months before it happened. The bond market priced it in October 2015.

So the question of this post is not “how could anyone have known.”

The question is how one of the most disciplined investors alive managed not to know. And what does that say about you and me?

Let me tell you the whole story. It’s the best investing lesson I have ever worked through, and by the end of it you will never look at a “great business with some debt” the same way again.

Share

The most beautiful story on the market

To understand why smart money died here, you first need to feel why smart money showed up. Because the thesis wasn’t stupid.

The thesis was gorgeous.

When steel mills melt scrap, they produce a toxic dust that is about 20 percent zinc. By law, someone has to dispose of it. Horsehead was the biggest company in America doing exactly that. Steel mills paid Horsehead to haul the dust away, and Horsehead then extracted the zinc and sold it.

Every other zinc producer on the planet pays to dig its raw material out of the ground. Horsehead got paid to receive its raw material. A commodity producer with negative feedstock cost. A real moat in the most moatless industry that exists. Pabrai said it plainly in interviews:

“Commodity businesses are fine to own, as long as you own the lowest cost producer.“

He had first bought Horsehead back in 2008 as a classic Ben Graham cheap stock. It went up more than 400 percent in about a year. The stock had already made him look like a genius once.

Then came the part that turned a good story into an irresistible one. In 2011, Horsehead announced it would shut its ancient 80-year-old smelter and build a brand-new, state-of-the-art plant in North Carolina. Budget: around $350 to $375 million. Promised profit once running: $90 to $110 million of extra EBITDA per year. Projected return on the project: 45 to 50 percent.

New plant, half the production cost, lowest cost zinc producer in the world. Pabrai didn’t just buy it. By late 2013 his funds owned 11.5 percent of the entire company, one of the biggest positions in a portfolio of only six stocks. His close friend Guy Spier bought it too.

Here is what makes this story immortal, and I want you to hold this thought until the very last line of this post. The thesis was largely correct. The moat was real. Horsehead really did get paid to take the waste.

The company still went to zero.

How is that possible? The answer arrived one quarter at a time, in public, while everyone watched. On July 8, 2014, six weeks after the shiny new plant started up, Horsehead issued a press release admitting the plant had produced at about 23 percent of its designed capacity and had already been shut down once for repairs.

That was the first crack. Pabrai didn’t move. It was early, and every new plant has teething problems. Right?

What follows below the line is the full autopsy of what happened, quarter by quarter.

Read more