Funding
Carnegie Mellon Researcher Raises $35M for Clinical Intelligence Work

Key PointsXlue raised $35 million for a health-state model that abstracts patient trajectories…
View more
- Xlue raised $35 million for a health-state model that abstracts patient trajectories instead of tokens.
- Xlue claims its model beat frontier token-based systems in blind evaluations at under 0.1% of the compute cost.
- Xlue is running prospective disease-risk studies and drug-indication simulations to prove real-world value.
Xlue disclosed on November 18, 2026 that a Carnegie Mellon researcher had raised $35 million for clinical intelligence. On the same date, the company published a technical report describing its human health foundation model, designed to understand health, predict how it changes and simulate how treatments can alter its course.
The work is aimed at turning clinical questions into computational problems, allowing physicians and scientists to explore possible outcomes computationally before acting. The model is intended to learn from patients’ changing health, their responses to treatment, and the experiments and trials through which medicines are developed.
Modeling Health States Instead of Tokens
Xlue argues that models trained heavily on medical textbooks and papers can demonstrate knowledge without necessarily being able to predict when patients become ill or which treatment will work. Applied to health data, token-centric systems can also divide patient trajectories into sequences containing uneven information and repeated patterns while entangling different views of the same person.
Xlue’s alternative abstracts tokens into health states—representations of health as it evolves—and treatment actions. Disease progression, treatment response, clinical outcomes and interventions are treated as expressions of one process: health changing over time and actions changing its trajectory. The company describes this as a new pretraining approach for modeling patient trajectories across populations.
Reported Evaluations and Trial Simulations
Xlue reports that its health-state model scales more effectively on health data, generalizes better when modeling patient trajectories across different populations and outperformed frontier token-based models in blind third-party evaluations at less than 0.1% of the compute cost.
The report proposes using the model to simulate clinical trials offline. A patient cohort would be modeled under drug treatment and under control or standard care, with predicted outcomes feeding back into cohort selection, protocol design and possible new indications. Xlue presents the approach as a way to help drug developers design trials with higher success rates, mitigate risks and identify new uses for medicines.
Prospective Studies and Recursive Health Improvement
Xlue says it is conducting prospective disease-risk-prediction studies across multiple health systems to support preventive care. It is also working with pharmaceutical companies to identify new indications for existing drugs using world-model simulations, as part of broader deployment work with hospitals and drug developers.
The longer-term plan is a shared foundation model connecting molecular, cellular, patient, clinical-trial and population-health signals. Xlue assigns it seven clinical tasks: drug discovery, patient identification, care planning, precision medicine, diagnostic workup, indication discovery and clinical-trial optimization.
The company calls its envisioned feedback loop Recursive Health Improvement. Computational models would assist and learn from wet-laboratory experiments, clinical trials and patient care, allowing signals from one phase to inform others. Those projects give the financing its immediate clinical context: they are the real-world settings in which Xlue’s health-state abstraction and trial simulations must demonstrate usefulness beyond the company’s reported evaluations.












