vLLM adds an AFD plugin for disaggregated MoE serving on GPU and Ascend
vLLM Blog · rss · 2026-07-23
vLLM has announced its AFD Plugin, a new addition to its serving stack for flexible MoE inference.
What it adds
- Attention–FFN disaggregation for MoE serving
- Support for GPU and Ascend NPU backends
- Connector-based execution
- Graph execution and ubatching support
The update is aimed at making MoE serving more flexible and efficient inside the vLLM ecosystem.
More from Infra
- DeepSpeed shows AutoEP/AutoTP/AutoSP parallelism gains, hosts office hours — PyTorch · 2026-10-27
- Cloudflare Durable Objects validated as top architecture for AI agent WebSocket servers — tristanbob · 2026-10-27
- Gary Marcus: With Q3 data added, Anthropic's compute-efficiency growth shows deceleration — GaryMarcus · 2026-10-27
- SpaceX locks in $40B GPU financing deal with Apollo, the largest ever committed — TechCrunch AI · 2026-10-27
- Inference is just Kubernetes with a multi-trillion-dollar market, and the communities overlap — matt_slotnick · 2026-10-27
- vLLM eyes Rust migration as inference engines race to rewrite the stack — AccBalanced · 2026-10-27