Hippocratic AI's Polaris system has handled 10 million patient calls with a reported 99.9% clinical safety score [1], [2].
This milestone suggests that large-scale generative AI can be deployed in safety-critical medical environments without compromising patient security. By scaling these interactions, the system aims to reduce the burden on healthcare providers while maintaining evidence-based standards.
The system operates on DigitalOcean’s AI-Native Cloud platform [1]. It utilizes NVIDIA HGX B300 Blackwell Ultra GPUs to process the high volume of data required for real-time medical interactions [1], [2]. This hardware integration allows the Polaris system to maintain stability and speed during patient engagements.
The deployment focuses on providing evidence-based AI for healthcare [5]. The company said the goal is to create a system capable of handling large volumes of patient interactions while ensuring high clinical safety [4]. This approach is intended to outperform existing frontier models on critical medical tasks [5].
DigitalOcean and NVIDIA provide the underlying infrastructure that enables this scale [1]. The partnership combines cloud orchestration with specialized GPU hardware to support the computational demands of medical AI. The 10 million call mark serves as a benchmark for the system's ability to operate at scale in a professional health context [3].
“10 million patient calls with a reported 99.9% clinical safety score”
The integration of specialized Blackwell Ultra GPUs with a dedicated AI-native cloud indicates a shift toward 'vertical' AI stacks in healthcare. By prioritizing a clinical safety score over general conversational fluency, Hippocratic AI is attempting to solve the 'hallucination' problem that has previously prevented the widespread adoption of LLMs in medicine.
