Atlas Discovery
Atlas Discovery is an AI-native pharmaceutical company that deploys autonomous agents to mine the biomedical corpus and repurpose existing drugs for rare diseases. Its clinical trial prediction system achieves AUROC 0.858 on Phase III-to-approval forecasts, and the company has identified an FDA-approved drug as a candidate for Cerebral Cavernous Malformation, now in preclinical testing.
- Drug Discovery & Therapeutics
- Artificial Intelligence
- Biotechnology
- Healthcare Technology
- AI Agents
- Data & Analytics
Funding
Raised to date
$500KRaised to date based on public sources. This may differ from the amount the company actually raised and is based only on what is publicly available on the internet.
Founders
Founder details are not available yet.
Product
Problem
Traditional drug development is organized around high costs of knowledge work, experimental capital, and regulation, making it economically unviable to pursue treatments for diseases with small patient populations or uncertain commercial returns. The long tail of diseases with tractable biology remains untreated because the cost of thinking about how to treat them has never been justified by market size. Clinical trial forecasting and drug repurposing decisions rely on scattered evidence across papers, patents, and filings, leading to poor success rates and inefficient trial designs.
Solution
Atlas Discovery is an AI-native pharmaceutical company that builds autonomous agents to repurpose existing drugs for new diseases and run clinical trials efficiently. The company shifts the cost of knowledge work to the cost of inference, enabling a differently organized firm driven by agents rather than traditional research teams. Its agent harnesses construct evidence representations tailored to each hypothesis, and its clinical trial prediction models convert assembled evidence into probabilities of clinical success. The company has demonstrated the approach by identifying an FDA-approved drug with over a decade of human safety data as a candidate for Cerebral Cavernous Malformation, now in preclinical testing, and by achieving AUROC 0.858 on Phase III-to-approval predictions, beating leading academic baselines.
Target Audience
Atlas Discovery serves patients with diseases that have tractable biology but lack commercial drug development attention, along with clinical partners and patient advocacy organizations such as Alliance to Cure Cavernous Malformation. The company's tools also serve the broader pharmaceutical industry through its clinical trial prediction and patient response forecasting capabilities.
Features
- Drug repurposing agent harness that builds hypothesis-specific evidence representations rather than relying on static knowledge graphs, recovering dapagliflozin at rank 2 for heart failure where TxGNN placed it 4,388th of 7,957
- RepurposingBench benchmark that distinguishes reasoning from recall, with evaluation sets built from pairs no model recovers under direct probing, achieving weighted recall of 35-41% on rediscovery and under 9/100 on prospective sets
- Clinical trial prediction system that assembles evidence from papers, patents, filings, and prior programs, reaching AUROC 0.858 for Phase III to approval transitions
- Foundational model of patient biology predicting drug response from baseline biopsies at AUROC 0.76, which would have reduced UNIFI trial enrollment by 458 patients at matched statistical power
- ClinicBench evaluation framework drawn from 500,000+ real patient records to assess agent performance on forecasting 30 days of patient care, with best frontier models scoring 51.5 of 100