The Aluminum Shuffle – Goldman’s Warehouses of Manufactured Scarcity

In the summer of 2011, Coca-Cola quietly complained to the London Metal Exchange. Not about the price of aluminum futures – there was plenty of metal sitting around. The problem was getting any of it out. The queue to pull aluminum out of a Detroit warehouse had stretched to six months or more. “It takes two weeks to put aluminum in, and six months to get it out,” Coke’s head of strategic procurement, Dave Smith, told an industry conference. The warehouse strangling the supply belonged to Goldman Sachs. And by design.

The play: buy the chokepoint

In 2010, with commodity prices still low, Goldman Sachs paid roughly half a billion dollars for Metro International Trade Services, a metal-storage operator based in Romulus, Michigan. Metro already held about 900,000 tons of aluminum in the Detroit corridor when Goldman bought it. Within a few years that hoard had grown to a million and a half tons – roughly a quarter of the entire national supply of aluminum in the US – parked across a network of 27 Metro warehouses around Detroit.

Owning the warehouse was the whole point. Every day the metal sat idle, Metro charged rent, roughly 40 to 48 cents per ton per day, which by 2013 was generating around $200 million a year in revenue on the Detroit stores alone. The longer metal stayed, the more Goldman made. So Goldman made sure metal stayed.

How you manufacture scarcity in a commodity glut

Here is the clever, rotten core of the scheme. The London Metal Exchange, which sets the global benchmark price for aluminum, had a rule meant to stop hoarding: any warehouse storing deliverable metal had to ship at least 3,000 tons out every day. A simple, sensible anti-cornering rule.

Goldman did not violate it. Goldman gamed it. Metro loaded the aluminum onto trucks and train cars – and moved it from one Metro warehouse to another, in a circuit around Detroit. That counted as metal “moved out,” satisfying the LME load-out requirement, while the metal never actually left Goldman’s system and never reached a single can, car, or window frame. Buyers at the back of the line stayed in line. Delivery waits went from an average of six weeks before the acquisition to more than sixteen months after it.

On top of the rent, Metro paid incentive fees of $100 to $230 per ton per year to hedge funds and commodity traders who agreed to park their metal there. In exchange, those traders could use the warrants as a cheap way to hold an aluminum position, and the pile kept growing. A Metro executive put it plainly to CNBC’s Kate Kelly: “The ownership of metal is a control game.” There were so few units in circulation, he explained, that the market was easy to squeeze.

The bill lands on the beer can

The squeeze did not show up in the headline futures price, which stayed fairly calm. It showed up in the premium – the extra cost buyers pay on top of the benchmark to actually get physical metal delivered. The US Midwest premium roughly doubled, from about 6.5 cents a pound in 2010 to a record near 12 cents by mid-2013. Analysts blamed the Metro delays and estimated the manufactured scarcity cost American consumers more than $5 billion between 2010 and 2013 – a surcharge hidden inside the price of every soda can, beer can, car body, and strip of siding.

Coke, a longtime Goldman client that had just hired the bank to advise on a $12 billion acquisition, got the full runaround. Goldman’s answer was that the LME system itself was inefficient and the firm was simply following the rules. That is true, and it is also the entire point: the enshittification here is not a conspiracy so much as a set of incentives. As Bloomberg’s Matt Levine later put it, the “conspiracy” was the rational outcome of an irrational, inefficient system – one where the party controlling the warehouse was paid to make the line longer.

Nobody wins, everyone pays

In August 2013, the Commodity Futures Trading Commission subpoenaed Goldman over complaints that its warehouses had “intentionally created delays and inflated the price of aluminum.” Lawmakers grilled the firm on live television. Antitrust lawsuits piled up. JPMorgan (via its Henry Bath warehouse unit) and Glencore (in Vlissingen, the Netherlands) had quietly built the same play. In December 2014, Goldman sold the aluminum warehousing business to Reuben Brothers and largely pulled out of physical commodities. The antitrust cases were dismissed in 2015, revived by an appeals court in 2019, dismissed again in 2021 – and the direct aluminum purchasers quietly settled with Goldman and JPMorgan in 2022.

The mechanism did not die; it just found new metal

Selling the warehouses did not kill the playbook. Commodity financialization kept right on squeezing – the 2021-2022 aluminum price spike and backwardation, the 2025-2026 copper squeeze where front-month spreads hit their widest backwardation in years as funds and trading houses hoarded metal and drove AI-linked supply anxiety. When the exchange’s warehouse is owned by the house, scarcity is a revenue stream. Every time a metal price jumps and your factory, your utility bill, or your grocery total moves with it, ask who is renting the chokepoint. It is usually a bank.

Status: Still rotting – Goldman exited, but the cornering-and-rent-extraction model lives on across base metals.

---

Sources: The New York Times (“A Shuffle of Aluminum, but to Banks, Pure Gold,” July 2013), The Washington Post Wonkblog, CNBC (“How aluminum became a cash cow for Goldman,” excerpting Kate Kelly’s The Secret Club That Runs the World), Reuters, and Wikipedia’s Aluminium price-fixing conspiracy.

Watch the full segment: Nightly Business Report – Goldman Sachs gets grilled on aluminum prices