← October 13, 2026 briefing · added in the 20:00 KST update
NVIDIA plans to support Google's 'Falcon' transport protocol in next-gen ConnectX-10 SuperNIC and BlueField DPU (report)
Announced October 12, 2026
Verified (fact)
According to AI Times Korea (인공지능신문), NVIDIA plans to support Falcon Transport, a transport protocol developed by Google, in its next-generation network cards, the ConnectX-10 SuperNIC and BlueField DPU. The announcement was made at OCP Global Summit 2026, held October 12–15 (local time). The focus is on addressing network congestion and packet loss in large-scale AI training and inference environments and improving bandwidth utilization in Ethernet-based infrastructure.
Why it matters (interpretation)
NVIDIA, the leader in the GPU market, will support a Google-built data center transport protocol in its own networking hardware. This could improve interoperability between different vendors' equipment in Ethernet-based AI clusters. It is separate from the previously reported release of the Vera Rubin design through OCP.
Confidence medium · official source pending
Sources
- Press 엔비디아·구글, 팔콘 트랜스포트 지원…차세대 AI 네트워크 상호운용성 강화 AI News Korea
More about NVIDIA
-
Microsoft to deploy its in-house Maia AI chips at multi-GW scale from next year; Maia 300 volume deployment in 2027 (reported)
On October 12 (local time), CEO Satya Nadella said on X and the official blog that Microsoft is making its largest-ever investment in...
-
NVIDIA introduces ReplaySSM, a state-storage method that speeds up decoding in hybrid SSM-attention models
In a developer blog post, NVIDIA introduced ReplaySSM, a redesign of how state is stored to speed up decoding in hybrid models. Hybrid...
-
Lablup open-sources GLM-5.2-NVFP4-DFlash-ko, a draft model to speed up GLM-5.2 inference in Korean (report)
Lablup said on the 12th that it had open-sourced GLM-5.2-NVFP4-DFlash-ko, a draft model that speeds up LLM inference for Korean-language...
-
Microsoft announces Maia 300, its third-generation in-house AI chip, as the 'primary inference system' for its services (report)
According to The Verge, Microsoft announced Maia 300, its third-generation in-house AI chip. The chip is designed as the 'primary AI...
-
NVIDIA opens Vera Rubin architecture as an open standard at OCP Global Summit, announces AI factory partnerships
At the Open Compute Project (OCP) Global Summit in San Jose, California, NVIDIA announced it will open the Vera Rubin architecture to...