After storage, copper and optical fiber face an AI-era scarcity crunch?

Following the storage sector, copper and optical fiber might be the next markets set to explode due to AI.

Charlie, a metals strategist from Citigroup, has been eager to execute a specific trade in recent weeks: buying a digital call option on LME copper with a strike price of $15,250 and an August expiration.

He believes that since 2022, almost all of the growth in copper demand has originated from energy transition and AI-related sources.

As of May 14, 2026, three-month LME copper is nearing $14,000 per ton, with COMEX copper opening at $6.63 per pound. Over the past 12 months, copper has risen 41%. In the last 4 weeks, it has gained 10%. This is a price near an all-time high.

Over the past two years, the market has framed the entire AI narrative as a chip story, centered on Nvidia's market cap, TSMC's capacity, HBM yields, and CoWoS packaging bottlenecks. Almost all discussions about "AI infrastructure" have focused on those few square centimeters of silicon.

But from a perspective many are unaware of, AI demand is cascading down from the silicon wafer, sinking into copper mines, and seeping into glass.

Copper is the Oil of the AI Era

Supply and demand determine price. The view that "copper demand is robust" can be seen more directly in the market.

Three-month LME copper closed at $13,943 per ton on May 11, 2026, an LME closing record high, rising 2.7% in a single day. COMEX copper hit an intraday high of $6.58 per pound on May 12. Copper has risen 41% over the past 12 months and 10% over the past four weeks.

At the beginning of 2025, copper was still around the $9,000 mark. It broke through $12,000 mid-year, ending the year with a 43% gain — copper's best year since 2009. In January 2026, copper broke $13,000 intraday for the first time. Then, four months later, $14,000 is almost here. The shape of this curve looks more like a rediscovered asset being priced under a new logic.

Trafigura is the world's second-largest metals trader. Its head of metals analysis, Graeme Train, provided a very concise breakdown of demand: of the additional 10 million tons of copper consumption expected over the next decade, one-third will come from electric vehicles, one-third from power generation, transmission, and distribution, and the remaining third from automation, manufacturing capital expenditure, and data center cooling systems.

Goldman Sachs, in a report titled "AI and Defense Place the Power Grid at the Center of Energy Security," offered an even sharper assertion: copper will become the oil of the AI era. Goldman Sachs estimates that by 2030, global grid and power infrastructure construction will contribute over 60% of incremental copper demand.

This might sound somewhat exaggerated, but upon reflection, it makes sense.

Copper's conductivity is 100% IACS, second only to silver among all metals. But silver is too expensive. For almost all large-scale industrial electrical conduction scenarios, copper is the only answer. The closest substitute is aluminum, but aluminum's conductivity is only 61% of copper's, meaning that for transmitting the same megawatt of power, aluminum wires need a larger cross-sectional area — they are heavier, take up more space, and have greater thermal loss. In the centimeter-scale space of data center server racks, this difference is almost unacceptable.

The same applies to thermal conductivity. Copper's thermal conductivity is 401 W/(m·K), five times that of iron and eight times that of stainless steel. A single NVIDIA GB200 card consumes 1,200W, and a standard rack with 72 cards has a total power consumption exceeding 130kW. At this level of thermal density, air cooling is no longer sufficient; liquid cooling is necessary. And in liquid cooling systems, almost every component that deals with "heat" is made of copper — cold plates, pipes, and water blocks.

In other words, copper is not the "preferred material" for AI data centers; it is the "only physical choice."

AI's power consumption is disruptive. Delivering electricity to data centers is highly copper-intensive.

A 1GW AI data center requires roughly 27,000 tons of copper just for power distribution and cabling. Meta's Hyperion data center in Louisiana has a scale of 5GW. This single project alone implies a copper demand of nearly 135,000 tons, and this doesn't even account for the high-voltage transmission lines, substations, and grid upgrades needed to bring power to the data center's doorstep.

Historically, we might have thought of copper as an easily obtainable metal, but recent data suggests this perception may need adjustment.

Starting in March 2026, the US-Iran conflict cut off sulfur and sulfuric acid exports from the Middle East. Sulfuric acid is a key input for heap-leaching copper refining, forcing Chilean smelters to reduce output. This was also the trigger for the recent surge in 2026.

The more structural and macro issue is this: the world hasn't discovered a new supergiant copper deposit in the last decade. John Meyer, an analyst at UK sell-side firm SP Angel, believes the breakeven price for developing the next generation of new copper mines is $13,000 per ton, which is already above current copper prices. Estimates from Wang Jiechao's team at CSC Financial suggest a global refined copper shortage exceeding 100,000 tons in 2026; Citi's forecast is more aggressive, at 308,000 tons.

The 2026 Optical Fiber "Famine"

The copper story, told so far, presents a clear bullish narrative. But if you zoom into the interior of an AI data center, you'll notice something quite subtle: some of copper's demand is being replaced.

"Next-generation AI infrastructure will require massive amounts of optical connectivity because computing demands are growing so rapidly that copper wires can no longer meet the needs." This was the view expressed by Jensen Huang in an interview this month.

As Huang stated, the data transmission demands of AI clusters are breaking through the physical limits of copper cables.

High-speed signal transmission over copper cables has two fundamental constraints: first, signal loss increases sharply with frequency; second, at high frequencies, the volume and weight of copper cables become untenable. Interconnect bandwidths between GPU clusters are advancing from 200G, 400G towards 800G and 1.6T, and the distance copper can effectively bridge shrinks from meters to tens of centimeters. AI clusters are at the scale of tens of thousands of cards, spanning multiple racks, and sometimes multiple data centers. Copper physically cannot do it.

But optical fiber can.

This is why the surge in optical fiber is more dramatic, purer, and more irreversible than that of copper. How extreme is the price increase for fiber this time around?

Data from CRU: the price of bare G.652D fiber in China surged over 80% in three months between November 2025 and January 2026. The average price in January was 31.5 RMB per fiber-km, with some actual transactions in the 40 to 50 RMB range, representing a cumulative increase of 94% to 144%.

Optical fiber, an industrial commodity whose price barely stirred for years, more than doubled in three months.

By February 2026, high-end fiber categories rose even more sharply. For instance, G.657.A bend-insensitive fiber jumped from above 30 RMB per fiber-km to over 50 RMB within a single month. Sun Telecom directly proclaimed a "2026 optical fiber famine." Its G.652D fiber, priced at $2.20 per kilometer in 2024, rose to $3.00 by December 2025, and then to $4.10 a month later. Fiber prices across Asia surged 75%, hitting a 7-year high.

The demand from AI data centers for optical fiber represents an order-of-magnitude disruption.

Rahul Puri, CEO of STL's optical network business, mentioned a figure that even made our editors pause: an AI data center requires 36 times the amount of optical fiber compared to a traditional CPU-based rack. This is a cliff-edge leap.

GPU clusters operate fundamentally differently from CPUs. A training cluster with tens of thousands of cards requires non-blocking, high-speed interconnects between all GPUs. This network structure, known as a Scale-out architecture, demands bandwidth unimaginable in the CPU era. Additionally, data centers need DCI links to stitch together computing clusters distributed across different geographic locations into a single supercomputer. Meta's Hyperion data center project alone requires 8 million miles of optical fiber.

Returning to the economic principle that price is determined by supply and demand: with demand in this state, what about the supply side?

Light Reading reported that at least one major fiber manufacturer has sold out its entire 2026 inventory. Data Center Dynamics reported that lead times for major customers have extended to 20 weeks, while smaller customers face waits of nearly a year.

Why can't production capacity be expanded quickly? Because the expansion cycle for optical fiber preforms — the core raw material for fiber — is 18 to 24 months, involving extremely complex processes. Even if all manufacturers decided to expand capacity today, new output wouldn't be available until the second half of 2027 at the earliest. In the meantime, demand will only continue to rise.

Anis Khemakhem, Chief Commercial Officer at Clearfield, provided an even more macro figure: by 2029, the United States alone will need an additional 213.3 million miles of optical fiber, nearly doubling the existing 159.6 million miles to 372.9 million miles. The nation's entire fiber inventory doubling in just 6 years.

The biggest winner in this story is Corning.

This is a glass company founded in 1851. It made the glass bulbs for Edison's light bulbs, the glass for television picture tubes, and the Gorilla Glass for iPhones. Many people don't even know this company still exists. But it is now a core fiber optic supplier for Meta, Nvidia, OpenAI, Google, AWS, and Microsoft. Its stock price has surged over 75% in the past year. The story of Corning is one our editors might explore in more detail in a future article, so we'll stop here for now.

The story of copper and optical fiber seems just to be starting to gain attention from the broader market. However, our editors believe this could become the next explosive sector following the storage track.