Tech Explained: Microsoft MicroLED tech targets AI data center efficiency  in Simple Terms

Tech Explained: Here’s a simplified explanation of the latest technology update around Tech Explained: Microsoft MicroLED tech targets AI data center efficiency in Simple Termsand what it means for users..

Microsoft has outlined a new data transmission technology designed to improve how AI infrastructure is built, using MicroLEDs to replace copper wiring and laser-based fiber optics inside data centers.

The system, developed by Microsoft Research in Cambridge, U.K., uses commercially available MicroLED chips and imaging fiber to transmit data using light. The company says the approach could reduce energy use, improve reliability, and ease physical constraints in AI data center design, as demand for compute continues to increase.

Doug Burger, Managing Director of Microsoft Research Core Labs, referenced the announcement in a LinkedIn post, where he said the innovation is “poised to advance the way AI (and other) systems are built.” He added that Microsoft has chosen to make the technology widely available, noting that lower energy consumption “across the industry is good for all of us.”

Limits of current data center networking systems

The development addresses growing constraints in how data is transmitted between servers and GPUs, particularly in AI workloads.

Today, most short-distance data transmission inside data centers relies on copper cables, which use electrical signals. Copper remains widely used because of its low cost, but it has physical limitations. As transmission speeds increase, the maximum effective distance of copper cables decreases, often to around two meters. This has led to higher power density within racks, making cooling more complex and limiting how systems can be designed.

For longer distances, data centers use fiber optic cables powered by lasers. These can transmit data over much greater distances at high bandwidth, but come with trade-offs, including higher power consumption, larger components, and greater sensitivity to environmental conditions such as temperature and dust.

Burger summarized the challenge in his LinkedIn post, explaining that increasing data rates force chips to be placed closer together, creating “a rapid increase in a rack’s power density,” which he described as “unsustainable.”

MicroLED approach introduces alternative architecture

The new system replaces both copper and laser-based transmission with a MicroLED-based approach that sends data through light using thousands of parallel channels.

The concept originated in 2020, when researchers explored whether commercial LED chips, similar to those used in consumer electronics, could be adapted for high-speed data transmission. The team developed a system that uses imaging fiber, a type of cable typically used in medical endoscopy, which contains thousands of internal cores capable of carrying parallel streams of light.

Paolo Costa, Partner Research Manager at Microsoft, explains how this differs from existing approaches: ”Imaging fiber looks like a standard fiber, but inside it has thousands of cores.” He adds: “That was the missing piece. We finally had a way to carry thousands of parallel channels in one cable.”

Unlike traditional fiber optics, which transmit data through a small number of high-speed channels, the MicroLED system distributes data across many lower-speed channels. Costa describes this as a “wide and slow” approach, compared with the “narrow and fast” model used in laser-based systems.

“The early concept of using LEDs to send data more cheaply — and lower power — than both copper and fiber optics seemed like a fantasy,” Burger says. “This breakthrough has the potential to change nearly every aspect of computing infrastructure … starting with high-bandwidth optical cables.”

Performance, energy, and reliability considerations

Microsoft’s internal testing suggests the MicroLED system could use around 50 percent less energy than current laser-based optical systems. The company also expects improvements in reliability, as LEDs are less sensitive to environmental factors that can affect laser-based components.

The system is designed to operate over distances of tens of meters, positioning it between copper and traditional fiber optics in terms of range, while aiming to combine the advantages of both.

To move the technology toward deployment, Microsoft worked with partners including MediaTek to develop a proof-of-concept system that can be integrated into existing data center infrastructure. The team has miniaturized the system into a transceiver device that is compatible with current server hardware.

Complementary networking developments in Azure infrastructure

The MicroLED system is being developed alongside other networking technologies within Microsoft’s infrastructure, including Hollow Core Fiber (HCF), which is already being deployed in some Azure regions.

HCF differs from traditional fiber by transmitting light through air within a hollow core, rather than glass, enabling faster data transmission and lower latency over longer distances. According to Microsoft, HCF can deliver up to 47 percent faster data transmission and around 33 percent lower latency compared with conventional fiber.

Frank Rey, General Manager of Azure Hyperscale Networking, explains how the two technologies fit together. “With MicroLED, you have the pure efficiency of LED over a laser,” he says. “That has a pure bottom-line impact to power usage at any given datacenter.”

He adds that Hollow Core Fiber extends the reach of data centers: “If you can go a much greater distance before you need to do any signal amplification, that means less buildings, less power, less generators, less energy.”

Together, the technologies are designed to support different parts of the network, with MicroLED focused on connections within data centers and HCF enabling faster, lower-latency connections across longer distances.

From research concept to industry deployment

The MicroLED system reflects a broader shift in how AI infrastructure is being designed, as compute demands increase and existing networking approaches reach physical and operational limits.

The project involved collaboration across Microsoft Research, Azure engineering teams, and external partners, with work spanning optical engineering, signal processing, and hardware design. A working prototype has been developed in the Cambridge lab, and the system has been reduced to a compact transceiver form factor for integration into servers.

Microsoft expects the technology to be commercialized with industry partners from late 2027, positioning it as part of the next generation of data center networking.

Burger highlights the longer-term implications in his LinkedIn post, noting that the approach could “open additional design options to make our systems much better,” while addressing energy and scalability challenges across AI infrastructure.

ETIH Innovation Awards 2026

The ETIH Innovation Awards 2026 are now open and recognize education technology organizations delivering measurable impact across K–12, higher education, and lifelong learning. The awards are open to entries from the UK, the Americas, and internationally, with submissions assessed on evidence of outcomes and real-world application.