{"data":{"slug":"tensorrt-llm","etype":"software","name":"TensorRT-LLM & Dynamo","cat":"Inference","subtitle":"NVIDIA's tuned inference stack — peak tokens from its own silicon.","summary":"TensorRT-LLM compiles models into aggressively optimized NVIDIA kernels (FP8/FP4, speculative decoding, in-flight batching); Dynamo (2025) orchestrates disaggregated prefill/decode serving across racks. Together they are how NVIDIA converts hardware advantage into benchmark-leading tokens per second.","attrs":null,"kf":{"maker":"NVIDIA","license":"Apache 2.0"},"aliases":["TensorRT","Dynamo"],"ticker":null,"exchange":null,"hq":null,"founded":null,"website":null,"status":"Active","region":null,"asof":"2026-01-15","sources":[{"url":"https://github.com/NVIDIA/TensorRT-LLM","label":"TensorRT-LLM"}]},"meta":{"found":true,"count":1,"served_from":"site cache (no live upstream calls)","as_of":"2026-10-14T08:41:21+00:00","disclaimer":"Research and information only — not investment advice. Structured records carry basis labels (vendor spec / reported / estimate); market quotes are delayed/indicative.","attribution":"Records compiled from vendor documentation, official filings and reputable public reporting; market quotes via financial data providers. See /data-sources."}}