Natera, Inc. announced a new collaboration with NVIDIA to accelerate the development of Natera's multimodal AI foundation model platform and expand applications across precision medicine. Under the collaboration, Natera will integrate its unique longitudinal and multi-time point in vivo datasets and proprietary machine learning models with NVIDIA's high-performance computing platforms, accelerated libraries, and AI frameworks. Together, the companies aim to establish the computational infrastructure required to train and deploy large scale AI models on complex medical datasets, enabling more precise diagnostics, deeper biological insight, and new tools to support personalized therapeutic decisions.

Natera's AI solutions, anchored by one of the largest multimodal, longitudinal oncology datasets in the world, are designed to address some of the most complex challenges in oncology. Recent data highlight strong performance in AI-driven immunotherapy response prediction and neoantigen prediction.oSelect?? - Natera's algorithm to prioritize neoantigens - outperformed 25 established neoantigen prediction algorithms. In addition, NeoPredict??

- Natera' algorithm to predict response to immunotherapy - has demonstrated more than a two-fold improvement in immunotherapy response hazard ratios compared with tumor mutational burden, the current clinical standard. These AI models have significant utility for discovery of new drug targets, prediction of therapy response, and identification of new prognostic biomarkers. Since announcing this platform, Natera has seen substantial engagement from pharmaceutical partners, reflected in new data collaborations and growth in the data business, reinforcing the momentum behind this approach.

Natera will utilize several NVIDIA platforms under the collaboration, including: NVIDIA Parabricks for accelerating bioinformatics workflows; NVIDIA BioNeMo for optimizing large-scale model training; and NVIDIA NeMo Agent Toolkit for building multi-agent AI systems designed for clinical and research use. Early engineering milestones show an average reduction of 75.7% in running time and 59.2% in cost using Parabricks for biological analysis, and 45.5% gains in training speed by integrating NVIDIA Transformer Engine. Natera's global testing and research networks generate exceptionally large volumes of data points each year, providing the scale needed to train increasingly sophisticated multimodal AI models.

Together with NVIDIA computing platforms, the companies aim to advance AI systems that deliver more accurate, data-driven insights across precision oncology, organ health and beyond.