NVIDIA Open-Sources BioNeMo Inference Runtime in Public Beta

NVIDIA Health announced that BioNeMo Inference Runtime has entered public beta: an open, PyTorch-native inference acceleration library for running biomolecular structure prediction faster on NVIDIA GPUs. It open-sources the technology behind its earlier NIM microservice acceleration, covering structure prediction models for proteins, nucleic acids, and ligands.

Confirmed

• Provides an end-to-end pipeline: from parsing FASTA/MSA inputs to structure outputs
• Accelerates inference with specialized GPU kernels and CUDA Graphs
• Supports structure prediction models such as Boltz-2 and OpenFold2
• According to AllThingsApx, Boltz-2 throughput improved nearly 3x
• The team previously saved substantial compute serving the AlphaFold Database with this technology

Why it matters

• Structure prediction is a core computational workload in drug design and biological research; the open-source library lets researchers achieve near-production NIM microservice-level acceleration on their own GPU clusters, lowering the barrier to entry
• Multiple reposts carry consistent information, all recounting the same release event

2026-09-10 ~ 2026-09-11 · 8 related posts

• NVIDIA's BioNeMo Inference Runtime hits public beta, boosting Boltz-2 folding throughput 2.9x — AllThingsApx · 2026-09-10
• [source] NVIDIA open-sources BioNeMo Inference Runtime, the GPU engine behind AlphaFold DB at million-scale — AllThingsApx · 2026-09-10
• NVIDIA launches BioNeMo Inference Runtime: 2.9x throughput for structure prediction, 1.6-3x in production — AllThingsApx · 2026-09-11
• NVIDIA open-sources BioNeMo Inference Runtime, saving 1.35 GWh across 31M protein predictions — AllThingsApx · 2026-09-11

4 near-duplicate retellings: AllThingsApx · AllThingsApx · AllThingsApx · AllThingsApx