Fabric.AI said it aims to demonstrate its Neural I/o MicroLED-based optical interconnect platform by the end of 2026 as the company positions itself in the rapidly expanding AI infrastructure networking market. The company also disclosed that it currently holds approximately $30 million in cash and cash equivalents, which it believes will fund operations through the planned demonstration milestone.
The Neural I/o platform is being developed in partnership with Kopin Corporation and targets a growing challenge inside AI clusters: the movement of data between GPUs, accelerators, memory systems, and servers. Fabric.AI said the architecture uses MicroLED-based optical interconnects as an alternative to traditional copper and laser-based approaches. The company argues the technology could improve power efficiency, thermal characteristics, density, latency, and scalability for hyperscale AI systems.
Fabric.AI also disclosed that it has signed two non-disclosure agreements with unnamed chipmakers to explore potential integration opportunities for Neural I/o in next-generation AI platforms. The company framed the initiative around the increasing importance of networking and interconnect technologies in AI factories, where bandwidth, latency, and power efficiency increasingly constrain overall system performance as GPU cluster sizes scale into hundreds of thousands of accelerators.
- Neural I/o demonstration targeted by late 2026
- Platform developed with Kopin Corporation
- Focus on MicroLED-based optical interconnects for AI infrastructure
- Target applications include GPU, accelerator, memory, and server interconnects
- Fabric.AI reports approximately $30 million in cash
- Company signed two NDAs with leading chipmakers
- Fabric.AI estimates future AI interconnect market opportunity could exceed $100 billion annually
“The industry increasingly understands that AI performance is no longer determined solely by compute. It is determined by how efficiently data can move throughout the system,” said Josh Silverman, CEO of Fabric.AI.
🌐 Analysis: Fabric.AI enters an increasingly crowded AI interconnect market where established players including NVIDIA, Broadcom, Marvell Technology, and multiple silicon photonics startups are racing to solve scaling constraints in AI clusters. The industry is simultaneously exploring co-packaged optics, linear pluggable optics, optical circuit switching, and advanced electrical fabrics as GPU cluster sizes continue expanding.
MicroLED-based optical interconnects remain an emerging category relative to conventional silicon photonics and laser-based optical links. Fabric.AI’s partnership with Kopin suggests an effort to leverage display-oriented MicroLED manufacturing and integration expertise for high-density interconnect applications inside AI systems, where power consumption and thermal efficiency increasingly dominate data center architecture decisions.
