How Physical Infrastructure Is Being Repriced

Copper and Fiber in the AI Era: How Physical Infrastructure Is Being Repriced

A UCS Perspective on the Structural Shifts Reshaping Structured Cabling in 2026

Introduction: A Market Repriced

The structured cabling industry entered 2026 with expectations of steady, incremental growth. What it got instead was a structural repricing.

Two signals tell the story. First, BSRIA’s Structured Cabling Worldwide 2026 report revealed that the global market grew 21% in 2025 to $9.08 billion, with the data center segment surging 54% year-on-year-3-8. Data centers now account for more than 41% of all cabling installed globally—nearly double the 21% share they held between 2015 and 2018-5-8.

Second, copper prices hit record highs, with London Metal Exchange futures touching $14,635 per metric ton in September 2026, up more than 17% year-to-date-15. The copper rally has been driven by a confluence of factors: a 1.1% decline in global mine output in the first half of 2026, robust demand from AI data centers and electrification, and tariff-driven stockpiling that has drained supply from global markets-2-11.

For anyone who has spent years treating structured cabling as a mature, commoditized layer of the network stack, these numbers demand a second look. The physical layer is no longer a passive backdrop. It is becoming a strategic variable—one that determines whether AI clusters perform at their potential or choke on their own connectivity.

For UCS, this moment carries a particular resonance. The company’s 1999 launch of a Category 6 zero-bit-error structured cabling system helped define what “high-performance copper” meant for a generation of networks. Today, the question has shifted: in a world where copper and fiber are being asked to do fundamentally different jobs, what does “high-performance” even mean?

I. Data Centers: The New Center of Gravity

The most consequential shift in the structured cabling market is geographic and architectural at once.

The United States alone accounted for 89% of the total global value increase in 2025-5-8. US data center cabling sales grew 44% year-on-year, and the country now holds 69% of the global data center cabling market—fourteen times the size of second-placed China-3-8. Germany, the UK, Australia, and India follow as the next largest markets, but the gap between the US and the rest of the world has widened dramatically-3.

This concentration matters for suppliers everywhere. When nearly nine in ten dollars of new market value flow through a single country’s hyperscale ecosystem, supply chains, product roadmaps, and even patent strategies tend to follow. The companies that win US hyperscale business gain volume, scale, and technological leadership that compounds globally.

But the US dominance is not monolithic. Within it, a quiet divergence is taking shape.

II. The Copper Paradox: Why the “Legacy” Medium Refuses to Die

Conventional wisdom held that AI would accelerate the long-predicted “fiber-to-the-chip” transition, displacing copper from every layer of the data center hierarchy.

The reality is more nuanced.

Copper cabling continues to hold its ground—and in some segments, is gaining. BSRIA notes that Category 6A has now overtaken Category 6 in global sales, a shift driven by the need for higher bandwidth in LAN environments and the availability of cost-effective shielded solutions-8. In data centers, Direct Attach Copper (DAC) cables remain the interconnect of choice for rack-level and top-of-rack connections, prized for their near-zero power consumption, minimal latency, and lower cost compared to optical alternatives-1.

The copper price surge has introduced a new variable. With raw material costs up sharply, the Producer Price Index for copper wire and cable rose approximately 18% year-over-year in July 2026-2. This cost pressure has shifted the commercial conversation as much as the technical one. Specifiers who once defaulted to the lowest-priced compliant product are now forced to weigh upfront savings against the risk of premature replacement—a calculation that favors higher-grade, longer-lifecycle solutions even at a premium.

There is an irony here that the industry would do well to acknowledge: copper’s cost crisis may ultimately strengthen the case for copper. When a material becomes expensive, the penalty for wasting it—through underspecified designs, premature replacements, or avoidable failures—becomes proportionally larger. The cheapest cable is no longer the cheapest infrastructure.

III. The Fiber Frontier: Density, Density, Density

If copper’s story in 2026 is about resilience under cost pressure, fiber’s story is about density under architectural pressure.

AI clusters are pushing fiber counts to levels that would have seemed implausible even five years ago. Corning has estimated that a single AI server rack may require more than 1,000 fibers, with future designs potentially reaching 5,000 per rack. At these densities, the physical connector—not the cable—becomes the bottleneck.

This is where the market’s most interesting technical development is unfolding. Traditional MPO/MTP connectors, based on 8-fiber or 16-fiber ferrules, have served high-density applications well for two decades. But AI’s insatiable demand for more channels in less space is exposing their limits. An 8-fiber MPO connector occupies a fixed panel footprint; doubling fiber count means doubling the number of connectors and the space they consume.

The industry’s response has been the emergence of Very Small Form Factor (VSFF) connectors—MDC, MMC, SN-MT, and their kin. These connectors dramatically reduce the physical footprint per fiber, enabling higher port densities in the same rack unit. The global VSFF connector market is projected to reach $182.5 million by 2031, growing at a 22% compound annual rate from 2024-12.

The strategic stakes of this transition became visible in early 2026 when US Conec, the pioneer of the MTP connector platform, announced a multi-party collaboration with Sanwa Technologies and Hakusan to expand supply of MMC connectors and TMT ferrules-4-9. The partnership was framed explicitly in terms of “supply chain infrastructure” for the rapidly growing high-density connectivity market. Sanwa, with over 75 years of optical manufacturing experience, will produce MMC connectors; Hakusan will leverage its 35-plus years in precision MT ferrule technology to supply x12 and x16 TMT ferrules-9.

The subtext is clear: the AI infrastructure buildout has reached a scale where no single supplier can meet demand alone. Multi-sourcing, once a procurement afterthought, has become a strategic imperative.

IV. China’s Divergent Path: A Cautionary Tale for Suppliers

No discussion of the global structured cabling market is complete without addressing China—and in 2026, that discussion is shaped by divergence.

While the US market surged 44% and captured nearly all the global value growth, China’s non-residential construction fell 17%, dragging down commercial LAN cabling demand across Asia-Pacific-5-8. The regional LAN market contracted 1.9%, largely attributable to China’s decline-8.

More significantly, China’s AI data centers are taking a fundamentally different architectural route. According to BSRIA, Direct Attach Copper cables account for 90–95% of connections in Chinese AI data centers, chosen for their lower cost and shorter lead times-3-8. Structured cabling—the organized, standards-based, high-performance system that defines Western hyperscale deployments—plays a substantially smaller role.

This has commercial implications that extend beyond China’s borders. Suppliers planning Asia-Pacific strategies must confront the reality that the region’s AI buildout will generate less structured cabling demand per unit of compute than the US or Europe. The “AI infrastructure boom” is not a uniform global phenomenon; it is a set of regional markets with distinct cost structures, supply chains, and architectural preferences.

V. Standards and the Compliance Layer

Beneath these macro shifts, the standards framework that governs structured cabling continues to evolve—and 2026 has brought meaningful updates.

China’s GB/T 18015 series for digital communication cables saw eight parts of the standard revised and implemented on May 1, 2026, covering horizontal, work area, and vertical cabling products. The industry standard YD/T 3296.1-2026 for polyolefin-insulated outdoor paired cables took effect June 1, 2026, replacing the 2017 version and covering outdoor cabling for communication systems, including support for low-voltage, low-power Ethernet power supply applications.

For specifiers and integrators, the layering of mandatory codes, recommended product standards, and design/acceptance standards requires careful navigation. The line between “must comply” and “should consider” is not always obvious, and the consequences of conflating the two can be significant—in design reviews, tender specifications, and project acceptance.

VI. The LAN Side: Steady, Essential, Underappreciated

While data centers dominate the headlines, the LAN market continues to do the essential work of connecting offices, powering desks, supporting WLAN, and keeping security systems running. BSRIA describes LAN demand as “steady rather than spectacular”-8.

LAN sales grew 5% in 2025, following 2% growth in 2024. The Americas now hold approximately 45% of the global LAN market and grew 9% in 2025, driven by the US and Canada-8.

For UCS and other suppliers with deep roots in commercial and enterprise cabling, the LAN market remains foundational. It is also where the copper price surge has its most immediate impact. Every office retrofit, every campus upgrade, every structured cabling project in a school or hospital or government building is now subject to a cost calculus that did not exist two years ago.

The response, for those who understand the lifecycle economics of cabling, is straightforward: lower-grade products that require replacement in five to seven years are a false economy. The true cost of cabling is not the invoice at installation; it is the sum of initial cost, replacement cost, downtime cost, and the operational drag of a network that cannot keep pace with demand. In a high-copper-price environment, that equation tilts decisively toward quality.

VII. What This Means for UCS

UCS’s position in this market is shaped by three decades of experience navigating precisely the kinds of transitions now underway.

The company’s heritage—from its 1974 origins as Abscissa Connect, through the 1990 merger that created UCS, to the 1999 launch of a Category 6 zero-bit-error system—reflects a consistent focus on the physical layer as an engineered system rather than a commodity input. The 2024 national patent for a twisted-pair production system addressing crosstalk challenges is a recent expression of that focus.

In a market defined by copper price volatility, fiber density challenges, and standards evolution, UCS’s value proposition rests on several pillars:

Lifecycle economics. When copper is expensive, the cost of premature replacement becomes intolerable. UCS products are designed for service lives measured in decades, not product cycles.

Technical depth. The transition to higher-density fiber connectivity and the continued evolution of copper performance standards require suppliers who understand the physics, not just the catalog. UCS’s investment in production technology and standards participation positions it to navigate these transitions.

Supply chain resilience. The US Conec–Sanwa–Hakusan collaboration is a reminder that even the most established supply chains require active management in a high-growth, high-stakes market. For UCS, the ability to source copper and fiber components reliably—and to offer alternatives when single-source risk becomes unacceptable—is a competitive advantage.

Long-term relationships. The structured cabling industry is not a fast-moving consumer market. Its customers—integrators, consultants, end-users—value consistency, support, and the confidence that a specification made today will be serviceable and compliant for years to come.

Conclusion: The Physical Layer Repriced

The AI boom has not made structured cabling obsolete. It has made it more consequential.

When compute clusters are measured in thousands of GPUs and network fabrics span hundreds of thousands of fiber strands, the physical layer stops being an afterthought and becomes a determinant of system performance. When copper costs $14,000 per ton, the economics of cabling shift from “how little can we spend?” to “how long will this last?”

For UCS, the answer to that question has not changed. Cable once. Cable right. Cable for the long run.

In a market defined by volatility—copper prices, AI architectures, standards, supply chains—the companies that endure will be those that understand something fundamental: the physical layer is where reliability is built, or lost. No amount of software abstraction can compensate for a cabling system that was specified to fail.

The repricing is underway. The question is not whether copper or fiber will “win.” The question is whether the industry will build infrastructure worthy of the networks it supports.

UCS intends to be part of the answer.