Blank Bio raises $7.2M seed to build RNA foundation models for cancer care
What's the deal? Blank BioDealroom has a profile for this one. Try Dealroom →, a San Francisco-based applied AI research lab, has closed a $7.2M seed round and struck a strategic collaboration with sequencing giant Pacific Biosciences (PacBio). The startup is training foundation models on RNA data from tumour samples to improve patient-level predictions in oncology.
The oversubscribed round drew backing from Define Ventures, Leonis CapitalDealroom has a profile for this one. Try Dealroom →, Nova ThresholdDealroom has a profile for this one. Try Dealroom →, Ripple VenturesDealroom has a profile for this one. Try Dealroom →, SignalFire, and Y Combinator, among others. Proceeds will fund model development, new long-read RNA sequencing datasets, and expanded partnerships with pharma and diagnostics companies.
Why now? Bulk RNA sequencing is already widely used across cancer research, drug development, and clinical care — but standard workflows compress the data into simplified gene-level summaries, leaving valuable biological signal on the table. Blank Bio argues that foundation models can extract far more from each patient's tumour transcriptome, turning raw molecular complexity into clinically useful predictions.
Under its PacBio collaboration, Blank Bio will generate long-read RNA sequencing data from up to 100 fresh frozen patient tumour samples across multiple cancer types. The sequencing will take place at Seattle Children's Research Institute. Blank Bio will use the resulting data to train and evaluate its models for biomarker discovery, patient stratification, and clinical trial design.
"Bulk RNA-seq is one of the most clinically accessible and information-rich assays in oncology, but much of its signal is still reduced to simplified gene-level summaries," said Jonathan Hsu, chief executive officer and co-founder of Blank Bio.
What could go wrong? Foundation models in biology are still early-stage, and translating lab-trained AI into validated clinical tools is a long, regulation-heavy road. The company will also need to prove that the additional resolution from long-read sequencing justifies its higher cost compared with standard short-read methods — especially if the end goal is broad clinical deployment.
The signal: Blank Bio's seed round lands as AI-driven drug discovery and diagnostics continue to attract early-stage capital, but the startup's focus on extracting more from an already ubiquitous assay — rather than building a novel wet-lab platform — is a distinctive bet. The strategic tie-up with PacBio, a late-growth-stage sequencing company, suggests incumbent platform players increasingly view computational partners as a route to differentiate their own hardware in a crowded market.
Read more: blank.bio
