Blank Bio

What do they actually do

Blank Bio builds a pretrained RNA model focused first on mRNA. It predicts sequence-level properties (like stability or expression proxies) and produces RNA embeddings that partners can use to narrow candidate designs before running wet‑lab experiments. The company distributes this in a partner-driven way (book-a-demo/consulting) and also publishes model code/research for open use [blank.bio, YC].

In practice, partners share sequences or design goals, Blank Bio runs its model to score and embed candidates, and then returns ranked options or design suggestions to cut down the number of lab assays needed. Partners validate top picks in the lab and may fine‑tune models with Blank Bio for specific tasks [blank.bio, YC].

Who are their target customer(s)

• mRNA therapeutics R&D teams at pharma/biotech: Each new sequence requires many slow, expensive experiments to test stability, expression, and safety; they want to pre‑screen candidates computationally to reduce wet‑lab cycles.
• Translational and clinical teams running biomarker and patient‑stratification programs: They need reliable RNA signals to pick the right patients and monitor response, but current RNA assays/analyses are noisy and hard to scale for smarter, smaller trials.
• CROs and in‑vivo delivery partners: In‑vivo testing and formulation screens are costly and time‑consuming; they want higher‑confidence candidates from computational pre‑screening to reduce failed runs.
• Internal computational-biology or ML teams at drug companies: They often lack validated, pretrained RNA models or reusable embeddings to plug into simulations and analytics, forcing them to rebuild heavy data/ML infrastructure in‑house.
• Academic and translational research labs studying RNA biology: Funding and lab capacity are limited; they need open, easy‑to‑run models to prioritize experiments and test hypotheses without large screening campaigns.

How would they acquire their first 10, 50, and 100 customers

• First 10: Run targeted partner pilots via founder networks and YC intros with pharma/biotech R&D and a few academic labs, plus convert 2–4 open‑source users already trying the repo into pilots. Bundle model predictions with a concrete, short validation plan and a clear deliverable.
• First 50: Turn successful pilots into paid POCs, publish 2–3 case studies/benchmarks to support outbound to similar teams, and use CRO/delivery partners as referral channels. Standardize onboarding, outputs, and fixed‑scope pricing to simplify procurement.
• First 100: Launch a self‑serve API/packaged product for computational teams alongside a sales‑led enterprise offering for pharma; hire a small commercial team for targeted account outreach. Formalize partner programs with CROs/delivery platforms and support with peer‑reviewed benchmarks and public repos/tutorials.

What is the rough total addressable market

Top-down context:

Core TAM aligns with drug‑discovery informatics/in‑silico software at about USD ~3.65B in 2024 Grand View Research. Adjacent markets relevant to Blank Bio’s roadmap include companion diagnostics (~USD 9.06B, 2024) and preclinical CRO services (~USD 6.19B, 2024), of which only a software/analytics slice is addressable GVR companion diagnostics, GVR preclinical CRO.

Bottom-up calculation:

Core (today): ~USD 3–4B based on drug‑discovery informatics spend. Near‑term expansion: combining software slices of companion diagnostics and preclinical CROs yields a practical ~USD 10–18B after trimming overlap and non‑software revenue GVR drug discovery informatics, GVR companion diagnostics, GVR preclinical CRO.

Assumptions:

• Only a fraction of companion‑diagnostics and preclinical CRO revenue is software/analytics relevant to RNA AI; most spend is tests, consumables, instruments, or services.
• There is budget overlap across informatics, biomarker programs, and CROs; totals are trimmed to avoid double‑counting.
• Adoption depends on validation with partners; near‑term revenue concentrates in high‑value enterprise deals rather than broad self‑serve.

Who are some of their notable competitors

• Deep Genomics: Builds BigRNA and an RNA‑focused AI platform for RNA biology discovery and design, directly overlapping with pretrained RNA models for mRNA design.
• Moderna: Runs an internal mRNA Design Studio and automated research engine, reducing the need for external sequence‑prediction tools for many use cases.
• BioNTech: Operates multiple mRNA platforms and invests in AI/ML for design and multi‑omics R&D, competing for the same partnerships and validation.
• NVIDIA (CodonFM): Released an open RNA foundation model (CodonFM) for codon‑level optimization and mRNA design; teams can adopt it directly instead of a specialist vendor.
• Aganitha.ai: Offers an end‑to‑end in‑silico mRNA design platform and consulting, competing on partner engagements and model‑plus‑service deliveries.