Bios Life has emerged from stealth with $25 million in seed funding to build an artificial intelligence-driven cancer surveillance service for survivors and people considered at elevated risk, targeting a US population it estimates at more than 40 million. The Boston-based company was founded by former BioNTech executives Ryan Richardson and Zach Taylor. Richardson, BioNTech’s former chief strategy officer, is chief executive, while Taylor, previously a senior vice-president at the German biotechnology group, serves as chief operating officer. The funding round was led by Redmile, Vsquared Ventures and Kindred Capital.
Bios Life plans to establish a virtual clinic providing personalised cancer risk assessment, screening and survivorship care. Its approach is intended to move monitoring away from fixed intervals towards a system in which an individual’s risk profile can change as genomic, laboratory and clinical information becomes available.
The company is initially focusing on cancer survivors and people with factors that place them at higher risk of developing the disease. About 18.6 million people in the US were living with a history of cancer as of May 2025, while the broader population Bios Life considers potentially suitable for its service exceeds 40 million when higher-risk individuals are included.
A strategic alliance with Tempus AI will provide access to a large pool of oncology information for developing and testing predictive models. The dataset covers more than seven million de-identified patient records and includes genomic information, medical imaging, transcriptomic profiles and clinical text. Bios Life also plans to incorporate hereditary cancer information through Ambry Genetics, which is owned by Tempus.
Central to the company’s technology plans is the Nucleotide Transformer, a genomic foundation model developed by InstaDeep with collaborators including Nvidia and the Technical University of Munich. BioNTech acquired InstaDeep in 2023 as part of its expansion into artificial intelligence for drug discovery and personalised medicine. Richardson chaired InstaDeep’s board following the acquisition.
The Nucleotide Transformer was trained to recognise patterns within DNA sequences and has demonstrated capabilities in research tasks such as identifying regulatory regions, splice sites and genetic variant effects. Bios Life aims to combine such genomic modelling with clinical and longitudinal health information to create continuously changing assessments of an individual’s cancer risk.
That ambition remains distinct from having a clinically proven system capable of predicting cancer or recurrence in individual patients. No publicly disclosed prospective clinical study has yet demonstrated that Bios Life’s proposed model can accurately determine when a particular person will develop a new cancer or experience recurrence, or that care guided by the system improves survival. The underlying genomic technology has been evaluated on research benchmarks, but that does not itself establish clinical effectiveness for continuous cancer surveillance.
The distinction is important as AI increasingly moves from biomedical research into patient-facing decisions. US regulators assess AI-enabled medical devices according to their intended use and risk, with systems providing patient-specific diagnostic or treatment information potentially subject to regulatory oversight and evidence requirements concerning safety and effectiveness.
Cancer surveillance also presents challenges that extend beyond algorithmic accuracy. More intensive screening can identify disease earlier in some circumstances, but additional testing can produce false positives, unnecessary procedures and overdiagnosis. Any system that generates more frequent risk alerts would therefore need to demonstrate that additional investigations improve meaningful outcomes rather than merely increasing clinical activity.
Bios Life argues that existing follow-up arrangements are too fragmented and episodic for many survivors and high-risk patients. Kevin Oeffinger, founding director of the Duke Cancer Institute Center for Onco-Primary Care, has said cancer screening and surveillance have changed comparatively little over the past two to three decades and that many people at elevated risk do not receive screening consistent with established guidelines.
The company intends to work with oncology centres while delivering much of its service virtually. Licensed clinicians would use personalised risk assessments to help determine surveillance plans rather than leaving patients to interpret algorithmic predictions independently. Bios Life has teams in the US and UK and plans to launch its platform during the second half of 2026.
The venture arrives as cancer care companies increasingly combine genomics, medical records and machine learning to predict risk and tailor treatment. Tempus has built one of the largest collections of linked clinical and molecular oncology information, while regulators have authorised a growing number of AI-enabled medical devices for narrowly defined diagnostic and risk-assessment applications.
Topics
Technology