NVIDIA AI Updates: July 9, 2026

1. NVIDIA Pairs Nemotron With LangChain Deep Agents for Top Benchmark Scores

NVIDIA. NVIDIA published benchmarks showing its Nemotron 3 Ultra model reaching leading performance on agentic tasks when paired with LangChain’s Deep Agents harness. The company detailed a custom harness profile that raises accuracy while keeping inference costs below proprietary alternatives, framing the configuration as an open-stack option for building agentic systems. Source

2. NVIDIA Runs GPU-Accelerated Presto on GB200 NVL72

NVIDIA. NVIDIA described running GPU-accelerated Presto, the open-source distributed SQL engine, on its GB200 NVL72 platform for low-latency analytical workloads. The technical post outlined how GPU acceleration speeds interactive queries over very large datasets, positioning the approach as a way to accelerate data analytics pipelines. Source