Chinese semiconductor design firm Biren Technology has unveiled next-generation "supernode" solutions at the 2026 World Artificial Intelligence Conference (WAIC) in Shanghai — systems designed to link over 1,000 AI chips across a single cluster using optical data transmission.
Speaking at WAIC on Friday, Biren vice-president Ding Yunfan said single-GPU performance was no longer sufficient for modern AI workloads. The core challenge, he stressed, was transforming massive GPU clusters into integrated, seamless systems.
Traditional copper-wired electrical connections are hitting a physical wall, restricting current architectures to 128 GPUs per server. To break through, Biren is betting on near-packaged optics (NPO), which integrates optical fibers closer to chips to boost data speeds. The layout could allow Biren to scale up to 1,024 cards across a cluster.
"Optical interconnect has become an inevitable choice to break through this bottleneck," Ding said.
The launch underscores how highly connected server systems have become the latest battleground in AI infrastructure. The industry is racing to scale raw computing power as AI models advance to trillions of parameters and AI agent applications gain widespread adoption.
Biren, which went public on the Hong Kong Stock Exchange earlier this year in a highly anticipated IPO that saw shares double on debut, is positioning itself as a domestic alternative to Nvidia amid the ongoing US-China chip war.
