Copper has acquired a new identity.

For decades, investors called it “Dr. Copper” because its price was viewed as a rough barometer for global industrial activity.

Now the metal is increasingly being treated as an AI infrastructure commodity.

Copper rose for a third consecutive session Tuesday, climbing as much as 0.4% in London before trading about 0.1% higher at $14,438 a metric ton. The latest advance followed a technology-led rally in Asian equities and continued the metal's rebound after its September record.

The connection to technology may sound surprising.

AI models live in data centers.

Data centers consume enormous amounts of electricity.

Electricity requires power generation, transmission and distribution.

And copper sits throughout that electrical infrastructure.

That is why the long-term copper story is increasingly becoming a story about something much bigger than computers:

the global race to build enough power for the digital economy.

The metal is not starting this rally from a low base.

Three-month copper on the London Metal Exchange reached an all-time high of roughly $14,875 a ton on September 10, after a powerful run fueled by U.S. tariff concerns, stockpiling and tightening availability outside the United States.

Tuesday's price of around $14,438 remains close to that record.

That means copper is behaving like a market where investors are willing to pay a premium for immediate supply even though the longer-term fundamental picture remains complicated.

And that complication is important.

Copper isn't rising purely because the world has suddenly run out of the metal.

A major part of the rally has been caused by where copper is located.

The prospect of U.S. tariffs on refined copper has encouraged traders to move metal into the United States ahead of any potential duties.

Reuters reported that U.S. imports of refined copper cathodes reached almost 885,000 metric tons in the first half of 2026, while U.S. inventories climbed sharply. At the same time, inventories outside the United States have become tighter.

This creates a strange market.

The United States can have a huge inventory.

The rest of the world can simultaneously become more concerned about supply.

That is because copper is fungible physically but not always interchangeable economically.

Shipping costs matter.

Tariffs matter.

Warehouse locations matter.

And the timing of future policy matters.

Traders are therefore pricing not only how much copper exists, but where it will be when manufacturers need it.

The supply story becomes even more interesting.

Analysts had previously expected a global copper surplus in 2026.

Reuters reported that CRU had forecast a surplus of around 639,000 tons before U.S. tariff-driven stockpiling changed the market's effective balance.

So why are prices near records?

Because a nominal global surplus does not necessarily mean abundant metal is available everywhere.

Copper accumulated in U.S. warehouses can effectively disappear from the pool available to manufacturers in Europe or Asia.

That is why tariff policy can influence prices even before any actual duty is imposed.

Markets anticipate future scarcity.

And anticipation changes physical flows.

The structural bullish argument is different.

AI data centers require large quantities of electricity.

That electricity has to reach the facility through transmission lines, substations, transformers and switchgear.

Data centers also need internal electrical distribution systems, backup power and cooling infrastructure.

Copper is present throughout those systems.

Bloomberg's latest market analysis says demand for copper from data centers and power infrastructure is expected to rise in the coming years, linking the metal's current performance directly to the same AI investment boom lifting technology stocks.

That gives copper something it didn't have to the same degree in earlier technology cycles.

A direct connection to the physical infrastructure required to scale computing.

This is one of the most important changes in the AI economy.

At the beginning of the generative-AI boom, the primary bottleneck was computing capacity.

There weren't enough advanced accelerators.

Then memory became a constraint.

Now power is becoming one of the key limitations.

Reuters recently reported that parts of the U.S. power system are struggling with rapidly rising data-center demand. PJM Interconnection, the largest U.S. grid operator, is facing a projected capacity shortfall of roughly 6.8 gigawatts linked to surging large-load demand, including data centers.

That means copper demand is increasingly tied to a much larger physical buildout.

It isn't just about data centers.

It is about the grid required to supply them.

The AI industry often talks about GPUs, models and software.

But a hyperscale data center is fundamentally a giant industrial facility.

It needs land.

Power.

Transmission connections.

Transformers.

Cooling.

Back-up generation.

Networking.

And electrical distribution.

Every additional gigawatt of computing capacity potentially requires enormous amounts of infrastructure around it.

That is why copper producers increasingly attract investor attention when technology spending accelerates.

The metal is sitting at the intersection of computing and electricity.

There is an important counterargument.

AI systems are becoming more efficient.

Data centers can use alternative materials and technologies.

Fiber optics can replace some copper in high-speed networking.

Nvidia has been developing higher-voltage architectures that can reduce copper intensity in portions of data-center systems. Reuters previously noted that this could limit some of the expected copper demand growth from AI.

That doesn't eliminate copper demand.

It means investors should be careful about assuming a simple relationship:

more AI equals proportionally more copper.

Technology changes.

Engineering choices matter.

And manufacturers constantly look for ways to reduce material intensity when prices become expensive.

Even if AI demand grows rapidly, the copper industry still has to produce the metal.

And that is becoming increasingly difficult.

A September market analysis from the Wall Street Journal noted that global copper mine production fell roughly 1.1% in the first half of 2026, with notable declines in Chile, Indonesia and the Democratic Republic of Congo. At the same time, refined production continued to increase, largely because of activity in China and the DRC.

That distinction matters.

Mines provide the raw material.

Smelters turn concentrate into refined copper.

If mining growth slows while refining capacity continues expanding, smelters eventually have to compete aggressively for concentrate.

That can squeeze margins and make the entire supply chain more fragile.

China dominates copper refining.

Reuters recently reported that Chinese smelters process approximately 60% of the world's copper cathodes and are facing a shortage of feedstock, prompting Beijing's antitrust authorities to seek copper-supply commitments from Anglo American as part of the proposed Teck merger.

That development reveals how strategically important copper has become.

China isn't simply a major consumer.

It is a critical part of the processing system.

And if Chinese smelters struggle to secure enough concentrate, the problem can affect copper availability well beyond China's borders.

This is one reason governments are increasingly treating copper as a strategic material alongside other critical minerals.

America has a particularly difficult copper problem.

The country relies heavily on imports and has limited domestic smelting capacity.

The possibility of future tariffs has therefore created a powerful incentive to bring metal into the country before any new policy takes effect.

Reuters has reported that the U.S. currently has only two operational copper smelters, underscoring the difficulty of rapidly expanding domestic refined supply.

Mining a new copper deposit also takes years.

Permitting can be slow.

Construction can be expensive.

And a mine requires enormous upfront capital.

That means a sudden price signal does not automatically produce new supply.

The copper market is therefore highly sensitive to disruptions that might look temporary in another commodity.

There is a paradox here.

High prices encourage investment.

But high prices also encourage stockpiling.

That can make physical availability tighter.

U.S. buyers accumulate copper because they fear tariffs.

Other regions lose inventory.

Prices rise.

Higher prices encourage still more inventory movements.

The cycle can feed itself.

That is exactly why analysts have warned that tariff uncertainty has transformed what should have been a global surplus into a much tighter effective market.

The current rally therefore contains both fundamental and financial components.

Some is genuine demand.

Some is supply anxiety.

Some is trade-policy positioning.

There is a broader macroeconomic implication.

Higher copper prices feed into the cost of electrical equipment, construction materials, industrial machinery and manufactured products.

If the metal stays near record levels, businesses have to decide whether to absorb those higher costs or pass them along.

That matters at a time when U.S. inflation is already above the Federal Reserve's target.

The recent ISM manufacturing report showed factory input prices rising sharply, while services businesses also reported increasing input costs.

Copper alone will not determine inflation.

But it is increasingly part of a broader commodity-cost problem involving oil, electricity, metals and industrial inputs.

The bullish case is compelling.

AI data centers are expanding.

Power grids require investment.

Electric vehicles continue to need copper-intensive infrastructure.

Renewable-energy systems consume large amounts of electrical wiring.

Industrial electrification adds another source of demand.

The bearish case is equally straightforward.

High prices encourage substitution.

Supply can increase.

Tariff distortions can unwind.

U.S. stockpiles can eventually flow back into the global market.

And a slowdown in AI infrastructure spending could reduce demand growth.

That is why copper at $14,000-plus per ton is not a simple “buy because AI” story.

The market is pricing a combination of strong demand and constrained immediate availability.

That may ultimately be copper's most interesting transformation.

Nvidia represents the intelligence layer.

Broadcom represents custom silicon and networking.

Micron represents memory.

Utilities represent electricity.

And copper represents part of the physical infrastructure connecting all of them.

The metal is therefore becoming a proxy for the question:

How much of the AI boom can the physical world actually build?

If data centers continue multiplying, grids need more transmission, and electrification accelerates, copper demand could remain structurally strong.

If technology becomes dramatically more copper-efficient or AI capital expenditure slows, the market's current premium becomes harder to justify.

For now, investors are choosing the bullish side.

Copper has climbed for three straight sessions and remains close to its record.

But the next leg of the rally will probably depend less on another technology-stock surge.

It will depend on whether mines, smelters, power grids and manufacturers can keep up with the physical demands of the AI economy.