AI-Bio
PrivateA capital-efficient AI-protein-design platform that just earned the field's loudest validation (Lilly $2.25B) while deliberately open-sourcing its crown jewels — the bet is whether "be the OpenAI of proteins" beats "own a drug," and the moat is the data flywheel, not the sequences.
Research
The verdict
A capital-efficient AI-protein-design platform that just earned the field's loudest validation (Lilly $2.25B) while deliberately open-sourcing its crown jewels — the bet is whether "be the OpenAI of proteins" beats "own a drug," and the moat is the data flywheel, not the sequences.
What it actually is. Profluent Bio is an AI-first protein-design company — it builds large generative "protein language models" (the ProGen family) that write novel functional proteins from scratch, the way an LLM writes text. Founded 2022 in Emeryville, CA by Ali Madani (CEO) and Alexander Meeske (co-founder/scientific). The thesis in one line: biology is a language, and if scaling laws hold for it the way they hold for text, whoever trains the biggest, best-aligned protein model can design proteins to order across every therapeutic and industrial format.
How it makes money (business model). Profluent is not a traditional drug company that owns and advances its own clinical assets. It is a platform / B2B licensing + partnership business with an unusual open-source funnel layered on top:
Key products / "modalities." Gene editors (OpenCRISPR-1), site-specific recombinases (the Lilly modality — large-payload DNA insertion), base editors (Rett/MECP2 program), antibodies (OpenAntibodies), plus enzymes, antigens and other biologic formats. The unifying product is really the model + the data beneath all of them.
Customers / suppliers / competitors.
Contract structure / payment terms. The pharma model is the classic biotech milestone-heavy structure: small upfront, R&D funding to do the work, then big numbers that only pay out on preclinical → clinical → regulatory → commercial triggers, plus royalties on net sales. Concentration risk is real and rising — Lilly is now the dominant economic relationship, and the headline $2.25B is optionality, not cash until a program is named.
For a generative-bio platform the "supply chain" is a data-and-compute pipeline → model → designed sequence → wet-lab validation → partner who develops/sells the product. Named stakeholders along the chain:
Chokepoints / single-source dependencies: (1) the alignment-data flywheel is the only durable chokepoint, and it is self-built; (2) delivery (AAV capsid size limits) is a hard physical constraint the Rett program explicitly designs around ("compact editor that fits a single AAV"); (3) partner concentration — Lilly is now the load-bearing customer. Names or it didn't happen: Lilly, Revvity, Corteva, Ensoma, IDT, Rett Syndrome Research Trust are the named chain stakeholders.
The honest moat assessment — this is the crux of the whole name. Profluent deliberately open-sources its best outputs (OpenCRISPR-1, OpenAntibodies designed to be non-patent-infringing on purpose). That makes the sequence explicitly not the moat. So what is?
Verdict on the moat: real but narrow and contested — a data/credibility flywheel, not IP. The open-source strategy is a deliberate bet that distribution + data beats secrecy (the "OpenAI of biology" play), which is brilliant if scaling laws hold and fatal if a better-funded rival's model simply gets bigger.
No segment financials are disclosed — Profluent is private and reports no revenue breakout. n/a — private, not disclosed. What can be qualitatively segmented (by modality / end-market), all ``:
| "Segment" (modality) | What it is | Lead external proof-point | Status |
|---|---|---|---|
| Gene editing — recombinases | Large-payload site-specific DNA insertion | Eli Lilly deal, up to $2.25B | Flagship commercial; preclinical |
| Gene editing — CRISPR/base editors | OpenCRISPR-1; compact AAV-fit base editor | Rett Syndrome Research Trust (MECP2) | Open-source + foundation program |
| Antibodies (OpenAntibodies) | De-novo antibodies vs 20 validated targets | 20 targets, ~$660B historical sales, 7M patients | Pre-release / showcase |
| Enzymes / ag-bio / tools | Industrial + agricultural proteins | Corteva, Revvity, IDT | Commercial relationships |
Trend & cause: the center of gravity is shifting from open-source demonstrations (2024 OpenCRISPR) → monetized pharma partnerships (2026 Lilly) — i.e., from proving the platform works to capturing value from it. That is the right direction, but it concentrates economics in a single partner before a second whale appears.
+private + +clinical overlay — funding/valuation + pipeline replace earnings/EPS)Round-by-round, all ``, unaudited:
| Round | Date | Amount | Total to date | Lead(s) | Notable participants | Implied valuation |
|---|---|---|---|---|---|---|
| Seed | 2023 | $9M | $9M | Insight Partners | Air Street Capital, AIX Ventures, Phoenix VP | n/a — not disclosed |
| Series A | 2024-03 | $35M | $44M | Spark Capital | Insight, Air Street; angels incl. Jeff Dean (Google DeepMind chief scientist), OpenAI/Salesforce/Octant figures | n/a — not disclosed |
| Series B | 2025-11-19 | $106M | $150M | Altimeter Capital + Bezos Expeditions (co-lead) | Spark, Insight, Air Street | ~$1B ("approaching $1B post-money") |
Reading the trajectory: classic steep step-up — $9M → $35M → $106M, with the lead investor upgrading each round (seed-stage Insight → growth-stage Spark → crossover-grade Altimeter + Bezos Expeditions). The crossover tell: Altimeter is exactly the kind of public-markets-adjacent fund whose entry signals IPO-proximity intent (see Lens 11). Total $150M is modest for the field (cf. Xaira's $1B+) — Profluent is the capital-efficient contender, not the capital-heavy one.
Burn / runway signals: not disclosed. With ~41–44 employees and a $106M fresh raise, runway is plausibly multi-year `` — flagged as estimate only; no disclosed burn. The Lilly upfront + R&D funding materially extends runway by offloading program cost onto the partner.
Profluent has no in-house clinical pipeline — it is a platform, so its "pipeline" is partner programs + open releases. Every program ``:
| Program / modality | Partner | Indication / target | Stage | Next value-inflection |
|---|---|---|---|---|
| Site-specific recombinases | Eli Lilly | Undisclosed (genetic medicine) | Preclinical / research | First named indication / target class / IND-enabling disclosure (the catalyst that converts $2.25B optionality → cash) |
| Base editor (compact, AAV-fit) | Rett Syndrome Research Trust | Rett (MECP2 hotspot) | Discovery | Demonstrate in-AAV editor correcting MECP2 |
| OpenCRISPR-1 | Open-source / community | Generalizable gene editing | Released (2024), Nature-published (2025) | Adoption → conversion to paid programs |
| OpenAntibodies | Open-source (planned) | 20 validated drug targets | Pre-release showcase | Public DNA "recipes" release; partner uptake |
| Enzymes / ag proteins | Corteva, Revvity, IDT | Industrial / ag | Commercial | Recurring tools/ag revenue (undisclosed) |
Probability-of-success: Not formally estimable — these are design programs upstream of even IND. The honest statement: the company has zero assets in clinical trials, so there is no rNPV of a drug — the value is entirely in the platform option (Lens 11). The asset table is a list of partnerships, which is the +private reality.
No earnings calls exist. The sentiment signal comes from founder interviews, conference talks, and publication cadence:
Cap-table quality (the syndicate tell): the syndicate is top-tier and improving — Insight (seed) → Spark (A) → Altimeter + Bezos Expeditions (B), with Jeff Dean and OpenAI/Google/Salesforce angels. Altimeter is the crossover-fund signal (IPO-proximity). No public mutual-fund markups/secondary marks are disclosed — n/a — private, not disclosed.
Comps — by category, not P/E (private + platform). Multiples for private peers are not sourced; what is sourced is funding scale and posture:
| Peer | Category | Funding / status | Posture vs Profluent | Source |
|---|---|---|---|---|
| EvolutionaryScale (ESM3, 98B params) | AI protein foundation model | $142M; acquired by CZ Biohub Nov 2025 | Closest model-first peer; now inside a non-profit — exits the commercial race | |
| Isomorphic Labs | AI drug design (DeepMind spinout) | Alphabet-backed (deep pockets) | Best-funded model rival; broader (structure+design) | |
| Xaira Therapeutics | AI biotech (owns assets) | $1B+ launch (2024) | The capital-heavy opposite model — owns drugs | |
| Generate Biomedicines | Generative protein medicines | $425M IPO (early 2026), ticker GENB | Public-market comp; asset-owning | |
| Absci | Generative antibody (Origin-1) | Public (ABSI) | Direct OpenAntibodies competitor | |
| Chai Discovery / Cradle / Basecamp | Private AI-protein tooling | VC-stage | Tooling-layer peers | |
| Recursion / Schrödinger / AbCellera | Public AI-bio platforms | Public | Public sentiment proxies |
EV/Sales, P/E, ROE, div-yield: n/a (Profluent private; no audited financials; private-peer multiples not disclosed). Do not fabricate. The one clean comp statement: Profluent (~$1B, $150M raised) is the lean platform play in a field where Xaira/Isomorphic are the capital-heavy asset plays and EvolutionaryScale just left the field via the CZ Biohub acquisition.
No stock to move; the equivalent mark-moving events (each materially re-rated the private valuation / narrative), all ``:
Pattern: the market (private capital) reacts to (1) top-tier publications and (2) blue-chip partner validation — science credibility and pharma endorsement, not revenue (there's no disclosed revenue to react to). The next catalyst that matters is a named Lilly program / IND-enabling milestone.
Ali Madani — Founder & CEO.
n/a — private, not disclosed). Deep domain + ML credibility; the company's reputational moat is largely his publication record.Alexander Meeske — Co-founder (scientific). CRISPR biologist; PhD Harvard Microbiology, postdoc in Luciano Marraffini's lab at Rockefeller (CRISPR-Cas13 anti-phage immunity), ran a UW Microbiology lab. Provides the wet-lab CRISPR credibility that complements Madani's ML — the editor designs are grounded in real CRISPR biology, not just sequence statistics. (Some sources list him as co-founder/advisor, suggesting his day-to-day role may be advisory while he retains his academic lab — a typical scientific-founder structure.)
Board / investor influence: Altimeter, Bezos Expeditions, Spark, Insight, Air Street now shape governance — a credible, IPO-capable board if the company heads that way.
No audited financials exist (private, no CIK) — so classic forensic income-statement/balance-sheet/cash-flow analysis is not applicable; there is no revenue-recognition, receivables, inventory, SBC-vs-non-GAAP, or goodwill line to interrogate. n/a — private, unaudited. The forensic lens re-points (per +private/+clinical) to deal-quality and scientific-integrity risks:
Regulatory findings (required sub-section).
regulatory/regulatory-findings.md (fetched 2026-06-30) reports total_sec_findings: 0 — Profluent has no CIK, is private, and is not required to file; no EDGAR enforcement search is possible."Profluent Bio" (FTC OR DOJ OR FDA OR consent decree OR settlement OR fine OR penalty) returned no material enforcement actions in the reviewed results — coverage was entirely funding/science/partnership news. (Note: as a pre-clinical platform, Profluent has no FDA-regulated marketed product yet; FDA exposure arrives via partner INDs, e.g. Lilly's.)n/a — no 10-K exists (private).No EPS projection — the company is private and pre-revenue-disclosure; an EPS line would be fabrication. The +private question is what unlocks a tradeable security, and when. No forecast.ts EPS forecast is logged (per --watchlist rules; and there is no EPS to forecast).
IPO-readiness read (no private-watch.json entry exists for Profluent yet — recommend adding one; see Open items):
Base / Bull / Bear path-to-value (qualitative, ``):
Bull case. Profluent is building the foundation-model layer of biology and has the three things that matter: (1) proof it works — the only company to put an AI-designed functional CRISPR in Nature and validate scaling laws at NeurIPS; (2) blue-chip validation — Lilly staked up to $2.25B, the loudest possible endorsement that pharma will pay for AI-designed enzymes; (3) capital efficiency — it reached this on $150M and ~45 people, so every incremental dollar (Bezos/Altimeter) buys more model and more data. The scaling-law thesis is the growth lever: if bigger models reliably make better proteins, Profluent's data flywheel (115B-protein Atlas + partner-feedback loop) compounds, and the open-source funnel keeps filling the top while partnerships monetize the bottom. The upside surprise is a second and third pharma whale turning a one-customer platform into the default protein-design vendor for the industry — a "picks-and-shovels for the genetic-medicine era" outcome.
Bear case (2–3 ways it permanently impairs).
Pre-mortem (18 months out, thesis broke — what happened?): A rival (Isomorphic or an open academic model) matched OpenAntibodies/OpenCRISPR-quality designs for free, Lilly quietly de-prioritized the recombinase programs without naming a target (so the $2.25B never started paying), no second whale signed, and the ~$1B mark proved to be peak-narrative private capital — the next event was a flat/down round or a strategic acqui-hire.
Are multiples too high? Unmeasurable (n/a, private). But ~$1B on zero disclosed revenue is a pure-optionality mark — fine while the catalysts keep landing, fragile the moment they pause.
Contrarian view (what the market refuses to see): The crowd reads the Lilly $2.25B as the bull-confirming headline; the contrarian read is that trading platform exclusivity to a single pharma for validation + capital is exactly what a platform does when it can't monetize broadly on its own — the deal is as much a constraint (Lilly locks up the best recombinase programs) as a trophy. Conversely, the under-appreciated bull is the open-source-as-distribution flywheel: if "OpenCRISPR/OpenAntibodies" become the default free tools researchers reach for, Profluent owns the standard, and standards monetize eventually (the GitHub/Hugging-Face-of-proteins outcome) — a moat the spreadsheet-driven bears can't see because it doesn't show up as revenue yet.
Dismantling the bull case:
A de-risked cash shell ($373M, no debt, ~$207M EV) wrapped around a still-shrinking lab-automation pivot — the balance sheet is the asset, the income statement is the warning; long the optionality only below cash, not the story.
The credible enzymatic-DNA-synthesis survivor — a real fidelity moat (1,005-base record, 50 kb clonal, ~99.9% stepwise yield) now distributed through Danaher/IDT — but it is a sub-$25M-revenue tools shop selling a faster picks-and-shovels commodity into a brutal synbio funding winter; WATCHING as a private until an IPO path or an IDT buyout crystallizes the value.
A fortress-margin vertical-SaaS monopoly trading at a growth-stock funeral price (~20x forward EPS, near 52-wk lows) because the market is pricing a Salesforce-Agentforce CRM war that threatens the contested ~40% (Commercial) while ignoring the defensible, faster-growing ~60% (R&D/Quality); BULLISH at $153 on a 1–3Y view, but the CRM-migration-to-2030 is a real, watchable execution overhang — not a phantom.