Key points from article :
Biostate AI, a generative AI startup, has launched K-Dense Beta, a multi-agent AI system designed to dramatically accelerate biomedical research, including ageing and longevity studies. Unlike conventional AI tools that focus on narrow tasks, K-Dense manages entire research cycles—designing experiments, analyzing biological data, reviewing literature, executing code, and generating publication-ready reports. Its system of specialized agents cross-checks results to ensure accuracy and traceability, mimicking the workflow of a collaborative team of scientists.
K-Dense’s potential was demonstrated in collaboration with Professor David Sinclair at Harvard Medical School, where it helped develop a transcriptomic ageing clock. The AI analyzed over 600,000 RNA expression profiles, filtered them down to 60,000 high-quality samples, and identified 5,000 key genes from more than 50,000 possibilities. The resulting model predicts biological age and includes measures of confidence for each prediction, providing researchers with valuable insight into reliability across diverse biological samples and conditions.
The AI runs on Google Cloud’s Gemini 2.5 Pro and has outperformed widely used models like GPT-5, GPT-4o, and Claude 3.5 Sonnet in bioinformatics tasks. Its modular design allows integration with standard bioinformatics pipelines and tools such as AlphaFold, broadening its applicability across biomedical research. Biostate claims that K-Dense can compress research timelines from years to weeks without compromising scientific rigor.
Founded by David Zhang and Ashwin Gopinath, Biostate AI raised $12 million in a Series A round and is expanding collaborations with hospitals, academic centers, and biopharma partners globally. K-Dense Beta is currently being tested with select partners, and broader availability is expected later this year, potentially transforming how ageing and longevity research—and biomedical discovery more generally—are conducted.
