Valar Labs is tackling one of oncology's hardest problems: predicting which treatment will actually work for a given cancer patient. Even with the same diagnosis and stage, patients respond very differently to therapies, and oncologists often lack the tools to know in advance who will benefit. Valar applies AI and computational pathology to routine tissue slides, the same biopsies already collected in standard care, to extract predictive signals invisible to the human eye.
The company's flagship product, Vesta, is positioned as the first AI diagnostic test to predict treatment response in bladder cancer. Rather than only classifying or grading a tumor, Vesta forecasts how a patient is likely to respond to specific treatments, giving oncologists actionable guidance at the moment of decision. Because it works from standard histology images, the approach can scale without requiring expensive new sample collection or specialized lab infrastructure, an important practical advantage for adoption.
Valar's founding team blends machine learning and clinical AI expertise; co-founder Pranav Rajpurkar is a well-known researcher in medical AI. This combination of rigorous ML and clinical grounding underpins the company's emphasis on building tests that are validated for real clinical use rather than research curiosities.
In May 2024, Valar Labs announced a $22 million Series A co-led by DCVC and Andreessen Horowitz (a16z), with participation from Pear VC. The round followed an earlier $4 million seed led by a16z and brought a16z's Vineeta Agarwala and DCVC's James Hardiman onto the board. The funding supports the advancement of Vesta and expansion of Valar's predictive approach into additional cancers.
Led by co-founder and CEO Anirudh Joshi, Valar Labs sits at the intersection of computational pathology, precision oncology, and AI, aiming to make response prediction a routine part of how cancer is treated.