AI BriefingPeking University Develops Optical AI Chip 100 Times Faster
Key point
Researchers at Peking University have developed an optical chip system that boosts speed by 100 times over existing methods while cutting power consumption to 1/9.
Details
Researchers at Peking University have built an Optical Chip System that succeeded in improving AI Inference speed by more than 100 times compared to existing methods.
Instead of adding more computing units as in the conventional GPU approach, this technology uses a completely new approach that connects chips using Light. This dramatically increases computation speed while also cutting power consumption to about 1/9 of the existing method.
If this technology is commercialized and scaled up, it is expected to become a fundamental solution to the data center expansion and power supply issues that are currently major challenges for the AI industry.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.
Related
Hierarchical Memory Architecture Needed to Resolve AI Inference Bottlenecks
2026 AI Inference Hardware Innovations: New Architectures for Prefill/Decode Disaggregation and Solving Memory Bottlenecks