Michael Burry disclosed a position in Ero Copper, betting that AI data centers' copper demand will outpace new mine supply -- Ero Copper shares jumped nearly 8% on the news.
At a glance
- Michael Burry disclosed a new position in Ero Copper (TSX:ERO), a Brazilian copper and gold miner, in a September 21, 2026 post on his Cassandra Unchained Substack newsletter.
- Ero Copper's share price jumped nearly 8% on the disclosure.
- Burry said he is largely ignoring AI-stock enthusiasm in favor of a copper-supply thesis, citing data showing major copper discoveries have fallen sharply since the 1990s.
- JPMorgan estimates a single large AI data center can require up to 50,000 metric tons of copper, and that data centers as a category will consume roughly 475,000 metric tons in 2026.
What happened
Michael Burry, the investor known for his early bet against the US housing market before the 2008 financial crisis, disclosed a new position in Brazilian copper and gold miner Ero Copper (TSX:ERO) in a 'Trading Post' entry on his Cassandra Unchained Substack newsletter dated September 21, 2026, saying he is largely ignoring the market's enthusiasm for artificial intelligence stocks in favor of a copper-supply thesis. 'All those back at the house are going to be needing a lot of copper,' Burry wrote, framing the position as an indirect way to play AI's growth rather than a bet against it. His case rests on a specific supply-demand mismatch: JPMorgan estimates a single large AI data center can require up to 50,000 metric tons of copper, and that data centers as a category will consume roughly 475,000 metric tons of copper in 2026 alone, demand that can materialize within two to three years, far faster than the roughly 18 years it typically takes a new copper mine to reach production. Ero Copper's share price jumped nearly 8% on the disclosure, and the news landed the same week COMEX copper futures touched a record intraday high of $6.95 a pound on September 22, before closing at a record $6.92.
The details
Burry's copper thesis is built around a timing mismatch rather than a simple 'copper is going up' call, and that distinction matters for how seriously to weigh it. The demand side of his argument moves fast: JPMorgan's estimate that a single large AI data center can require up to 50,000 metric tons of copper, with data centers as a category consuming roughly 475,000 metric tons in 2026, describes demand that shows up within the two-to-three-year window it takes to plan and build a data center. The supply side moves on an entirely different clock, new mine discoveries have fallen sharply since the 1990s, according to the data Burry cited, and even a discovered deposit typically takes around 18 years to reach production once permitting, financing and construction are factored in.
That gap, fast-arriving demand against slow-arriving supply, is the actual investment case, and it's why Burry's position is in a specific miner rather than a broad commodity bet: if the mismatch is real, producers already extracting copper, rather than companies still years from a permitting decision, are positioned to capture the resulting price tension most directly. It's also why Burry frames this as 'an indirect way to play AI' rather than a departure from the AI trade, his bet isn't against AI's growth, it's on a real, physical input that AI's growth requires and that can't be manufactured or substituted on short notice.
The market's immediate reaction adds weight to the argument without proving it: Ero Copper's shares jumped nearly 8% on the disclosure, and COMEX copper hit a record intraday high of $6.95 a pound the same week, before closing at a record $6.92. That's consistent with a market already pricing in some version of the supply tightness Burry describes, but it doesn't confirm his specific mine-versus-data-center growth-rate framing is what's actually driving the broader copper move day to day, as opposed to the mine-specific and currency factors covered separately.
Why it matters
A high-profile investor publicly naming a specific supply-demand mismatch, backed by named third-party estimates from JPMorgan, gives retail and institutional readers alike a concrete, checkable framework for thinking about copper's structural demand story, distinct from the day-to-day price moves driven by dollar strength or mine-specific disruptions covered elsewhere on this site.
Our read
Outlook: bullish. A well-known investor's public, data-backed copper-supply thesis is a supportive sentiment signal for a metal already near record prices, though it's an interpretation of existing supply-demand data rather than a new catalyst in itself.
What to watch
- Ero Copper's production updates and stock performance following Burry's disclosed position.
- Further data-center copper-demand estimates from JPMorgan or other banks as AI infrastructure buildout continues.
- New mine discovery and permitting rates, the supply-side variable Burry's thesis identifies as structurally slow.
- Whether other prominent investors publicly disclose similar copper-supply theses.
For information only, not investment advice.
Copper price in India
metalscost.com India reference price as of 2026-10-03.
Detailed analysis
Timeline
- 2026-09-21: Michael Burry discloses a new position in Ero Copper in a Cassandra Unchained Substack post.
- 2026-09-22: COMEX copper touches a record intraday high of $6.95/lb and closes at a record $6.92/lb.
Demand Drivers
JPMorgan estimates a single large AI data center can require up to 50,000 metric tons of copper, with data centers as a category consuming roughly 475,000 metric tons in 2026, demand that can materialize within the two-to-three-year window typical for planning and building a data center.
Supply Drivers
Major copper discoveries have fallen sharply since the 1990s, according to the data Burry cited, and a newly discovered deposit typically takes around 18 years to reach production once permitting, financing and construction are factored in, a far slower response time than the demand side of the equation.
What could lift prices
- JPMorgan's specific data-center copper-consumption estimates (50,000 tonnes per large facility, roughly 475,000 tonnes for the category in 2026) describe a large, fast-arriving demand source.
- The roughly 18-year lag between a new discovery and production means even a strong price signal today can't quickly bring meaningfully more supply online.
- Copper hit fresh records on both an intraday and closing basis the same week as Burry's disclosure, and Ero Copper's own shares jumped nearly 8%, consistent with the tightness his thesis describes.
What could weigh on prices
- A single high-profile investor's public thesis is not itself new supply-demand data, it's an interpretation of existing figures, and other investors may weigh the same data differently.
- Copper prices already near record highs mean much of a well-known supply-demand story may already be reflected in the price, leaving less room for the thesis to play out as a fresh catalyst.
Country impact
| Country | Impact | Reason |
|---|---|---|
| Brazil | Medium | Ero Copper, the specific miner Burry took a position in, operates copper and gold mines in Brazil. |
| United States | Medium | The demand side of Burry's thesis is driven by US-centric AI data-center buildout, and the underlying JPMorgan research is US-based. |
Industry impact
| Industry | Effect | Reason |
|---|---|---|
| Data Centers | Neutral | Data centers are the demand-side subject of Burry's thesis, consuming large volumes of copper for power and cooling infrastructure, though the article doesn't suggest a direct cost or supply risk to data-center construction itself. |
| Copper Mining | Positive | A prominent investor's public copper-supply thesis, backed by named data, is a positive sentiment data point for copper miners generally, not just Ero Copper specifically. |
Who gains, who loses
- Ero Copper and other established copper producers: A supply-demand mismatch favoring already-producing miners over years-away development projects directly benefits companies with existing output, and Ero Copper's own shares already jumped nearly 8% on the disclosure.
- Copper-intensive AI data-center developers: A structurally tight copper market raises input costs for the data-center buildout that Burry's own thesis identifies as the demand driver.
Other ways this could play out
- If AI data-center buildout slows or copper-substitution alternatives gain traction, the demand side of the mismatch could narrow faster than Burry's thesis assumes.
- A wave of new mine permitting approvals or faster-tracked projects, spurred by high prices, could shorten the effective supply-response time below the historical 18-year average.
- If broader AI capital expenditure cools due to a shift in investor sentiment, the specific data-center demand driver underpinning the thesis would weaken.
Price risks
- A slowdown in AI-related capital expenditure would directly weaken the demand side of the thesis.
- Since copper is already near record prices, a shift in broader risk sentiment or dollar strength could pressure prices independent of the underlying supply-demand mismatch.
Technical view
Price is mixed relative to its 20-period and 50-period moving averages, showing no clear trend alignment.
Computed from metalscost.com's own stored price history.
Brazil
United States