Convexia

What do they actually do

Convexia builds an AI‑first sourcing and diligence product that scans public and private data to find overlooked preclinical or early‑stage drug assets, scores them with in‑silico models, and routes the top candidates to human experts for review. They have a public "Playground" demo that shows how a sourcing/diligence query runs end‑to‑end, and they monetize today through pilots and licensing of specific modules (sourcing engine, diligence stack, BD agent) to pharma, biotech VCs, and computational‑biotech teams Convexia site Playground YC launch.

In practice, agents automate global scouting (papers, databases, IP, early disclosures), then a stack of models (they state "50+ custom‑tuned models") evaluates binding, ADME/PK, toxicity, off‑target and mechanistic fit. Internal PhD scientists and external domain experts review outputs; separate agents score commercial fit (market size, reimbursement signals, IP) and operational risk (CMC, CRO feasibility), and the system generates buyer‑ready decks and term‑sheet templates Convexia site.

They do not run a wet lab or in‑house trials today and are not focused on designing new molecules now; instead they partner with CROs when programs advance. Looking ahead, they plan to evolve into a "buy → develop → sell" operator that acquires assets they surface, uses CROs for IND‑enabling work and early trials, and exits those assets to strategic buyers, but this is a stated plan rather than current operations Convexia FAQ YC launch.

Who are their target customer(s)

• Pharma/biotech BD or corporate development teams seeking in‑licensing deals: Internal scouting is slow and fragmented; they need steady, high‑quality dealflow and packaged, buyer‑ready assets to shorten evaluations and move to terms faster Convexia site.
• Biotech VCs evaluating many preclinical opportunities: They need faster, cheaper diligence to sort promising programs from noise and prefer evidence‑backed short‑lists over manual literature reviews Convexia site.
• Computational‑drug/platform teams lacking translational expertise: They want external scientific validation and expert review of model‑flagged candidates to avoid false positives and to package programs for partners Playground.
• Small biotech founders or academic inventors with early assets: They struggle to reach buyers and present clear risk/commercial packages; they need help converting messy data into a sellable diligence packet and draft term sheets Convexia site.
• CRO/CMO program managers and BD at service firms: They need clearer, prioritized sponsor pipelines to bid accurately and manage execution risk; operational scoring and readiness packages reduce ambiguity Convexia FAQ.

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

• First 10: Use warm YC and founder network intros to run high‑touch paid pilots that start from a Playground demo and culminate in 1–2 packaged, buyer‑ready assets via live roundtables with Convexia scientists and the customer’s team Playground YC launch.
• First 50: Productize smaller modules (sourcing engine, diligence stack, BD agent) as standardized pilots and drive referrals through CROs and platform partners; turn early pilots into short case studies to speed demo→pilot conversion Convexia FAQ.
• First 100: Expand channels via CROs, incubators, and consulting partners; add a streamlined path (playground → paid module → packaged deliverable) and an API/white‑label option so platforms embed Convexia scoring. Leverage public proof points from packaged deals and early exits to win larger BD teams.

What is the rough total addressable market

Top-down context:

Near‑term TAM aligns with the AI in drug discovery software category, estimated at about USD 1.9B in 2024, while adjacent markets include large CRO services (roughly USD 56–85B in 2024, scope‑dependent) and biopharma licensing/M&A deal flow (~USD 140B in 2023) that Convexia aims to influence as it scales Grand View Research Grand View Research CRO Fortune Business Insights CRO IQVIA.

Bottom-up calculation:

Anchor on the USD 1.9B AI drug‑discovery market and assume Convexia’s sourcing/diligence modules map to 10–30% of that demand, implying a SAM of roughly USD 190–570M; capture scenarios of 1–5–15% translate to about USD 19M–95M–285M in annual opportunity, before any operator upside Grand View Research.

Assumptions:

• Convexia’s modules correspond to 10–30% of overall AI drug‑discovery tooling (sourcing/diligence/packaging vs. design/screening).
• Market anchors use 2024 third‑party estimates that vary by scope; figures are directional, not precise revenue forecasts Grand View Research.
• Operator upside from CRO‑executed programs and licensing exits is excluded from near‑term TAM and depends on validated results and capital deployment IQVIA.

Who are some of their notable competitors

• Partex: AI‑driven “drug asset manager” that scans large datasets to surface, repackage, and sometimes incubate deprioritized drug assets for in‑/out‑licensing; overlaps on asset scouting and packaging, with a broader asset‑management posture Nature profile.
• Formation Bio: Tech‑native pharma operator that acquires or in‑licenses programs and uses AI tools to prioritize and advance them; similar “buy → develop → sell” logic but focused on operating programs rather than selling modular diligence products.
• BIOPTIC (Optic Inc.): Agent‑based platform automating global scouting, scientific triage, and commercial diligence to prioritize drug opportunities; close overlap on agentic scouting and deal packaging, positioned as a real‑time, language‑agnostic engine agents.bioptic.io.
• Bioneex: AI‑enabled matchmaking/marketplace connecting drug developers with partners and investors; overlaps on helping asset owners get packaged and discovered, with a marketplace‑first model.
• SourceScrub: General deal‑sourcing platform for corp dev/PE teams that crawls public/private sources to surface targets; addresses similar BD pain points but is sector‑agnostic and not a life‑sciences scientific diligence stack.