Retro Biosciences

Phase A — Understand the business

Lens 1 · Company Overview

Retro Biosciences is a Redwood City, CA longevity-biotech founded May 2021 by Joe Betts-LaCroix (CEO), Sheng Ding (UCSF/Gladstone stem-cell chemist), and Matt Buckley (ex-Illumina/Bayer, Stanford genetics PhD). Its stated mission is blunt and singular: add 10 healthy years to the human lifespan. It emerged from stealth with $180M from Sam Altman as sole funder — among the largest individual investments ever into a longevity startup; Altman said he "basically just took all my liquid net worth and put it into these two companies" (Retro + Helion).

Business model. Traditional therapeutics R&D, run as a speed play. Revenue today is ~zero — pre-revenue, early clinical. The intended monetization is the standard biotech ladder: pharma partnerships (licensing/royalties), co-development JVs, and proprietary-dataset value. The differentiator is velocity — Betts-LaCroix targets a first drug to market "by the end of the 2020s" vs. the industry's 10–15 years, with a minimized-bureaucracy operating model (a two-person board of Betts-LaCroix + Altman, weekly win/fail memos).

What it actually is: a three-to-five-program platform, not a single-asset bet. This is the central fact for valuing it. Retro runs parallel shots at aging:

1. Autophagy — lead clinical asset RTR242, a small-molecule oral autophagy/lysosomal restorer, in Phase 1 for Alzheimer's.
2. Cellular / partial reprogramming — Yamanaka-factor-based, including the OpenAI-engineered "RetroSOX/RetroKLF" variants.
3. Plasma-inspired therapeutics — a scalable therapeutic-plasma-exchange ("molecular reset") program. 4–5. Sub-programs: iPSC-derived microglia replacement (Alzheimer's), HSC reprogramming (blood disorders, via the Murdoch Children's Research Institute deal), and tissue reprogramming (osteoarthritis, hearing loss).

Customers / suppliers / contract structure. No commercial customers exist yet. The economically meaningful counterparties today are research/manufacturing partners: OpenAI (AI protein engineering), Calico/Alphabet (a 2023 mouse-longevity collaboration), Multiply Labs ($85M robotic-manufacturing agreement, May 2024), and Murdoch Children's Research Institute ($35M research + commercial licensing, May 2025). There is no take-or-pay or recurring-revenue structure — the "contracts" are R&D collaborations and an in-house cGMP cell-therapy line.

Lens 2 · Supply Chain (→ CDMO / manufacturing, per +clinical overlay)

For a clinical-stage biotech the "supply chain" is the discovery → manufacturing → trial-site chain. Named stakeholders along Retro's chain:

• Upstream / discovery inputs: OpenAI (GPT-4b micro protein-engineering model — designs the enhanced Yamanaka factors); academic IP from UCSF/Gladstone (Sheng Ding's small-molecule stem-cell chemistry) and the Conboy lab at UC Berkeley (plasma-dilution science Retro's program is "inspired by").
• Manufacturing: In-house cGMP cell-therapy manufacturing (a deliberate vertical-integration move — rare for a company this young); Multiply Labs for robotic/individualized-therapy manufacturing ($85M deal); Murdoch Children's Research Institute for iPSC-derived blood-stem-cell production ($35M).
• Clinical: Phase 1 RTR242 is run at a specialized early-phase unit in Adelaide, Australia — a common choice for speed and the Australian R&D tax incentive.
• Regulator: FDA ("multiple successful interactions" claimed); Australian TGA for the Adelaide trial.

Chokepoints / single-source dependencies. The reprogramming program's discovery edge is single-sourced to OpenAI — a relationship that is also a conflict-of-interest node (Altman funds Retro and runs OpenAI). The cell-therapy programs depend on iPSC manufacturing scale and consistency, the classic bottleneck for autologous cell therapy. RTR242's path depends on a single Adelaide CRO site for first-in-human data. None of these is disclosed in audited detail — all ``.

Lens 3 · Competitive Advantages (moats) (→ IP / data / platform moat, per +clinical overlay)

• Talent density. Retro claims "one of the world's largest concentrations of leading aging-biology scientists" at ~60+ staff. Real, but contested — Altos Labs reportedly recruited ~50% of the leading reprogramming scientists with seven-figure salaries in 2023. Retro is talent-rich but not talent-dominant.
• The OpenAI/GPT-4b micro data-and-compute moat. The strongest genuinely-differentiated asset. GPT-4b micro produced RetroSOX/RetroKLF variants delivering a ~50x increase in reprogramming-marker expression and cut iPSC generation from ~3 weeks to ~7 days. >30% of AI-generated RetroSOX variants beat native SOX2; ~50% of RetroKLF variants beat hand-engineered versions. This is a privileged, possibly exclusive, AI-for-protein-design pipeline that peers cannot trivially replicate — the closest thing Retro has to a durable moat.
• Speed / process as moat. A "repeatable process for turning aging-biology advances into medicines" — 15 months from indication selection to first-in-human for RTR242. Process speed is a real edge in a field where most programs die in the lab, but it is not a defensible moat — it's an operating advantage rivals can copy.
• Bargaining power: weak today. Pre-revenue, pre-efficacy, and dependent on continued external capital — Retro needs investors, partners, and the FDA more than they need Retro. Bargaining power only arrives with clinical data.

Verdict on moat: the AI-protein pipeline is the one genuinely hard-to-copy asset; talent and speed are advantages, not moats. The moat is narrow and unproven, not wide.

Lens 4 · Segments

n/a — private, pre-revenue; segments.csv is a header-only scaffold. There is no revenue, EBITDA, or earnings to break out by product or geography. The closest analog to "segments" is the program portfolio (autophagy / reprogramming / plasma + sub-programs), covered in Lens 1 and the pipeline table in Lens 5. Reporting all n/a — not disclosed rather than fabricating a split.

---

Phase B — Measure performance

For a private pre-revenue clinical co, "performance" = funding/valuation trajectory, traction, and the catalyst calendar — not earnings.

Lens 5 · Pipeline by phase (+clinical swap of "Earnings Result")

The asset table is the company. Probabilities of success are industry base-rates ``, not Retro disclosures.

The single near-term value-inflection event: RTR242 Phase 1 data, ~August 2026. It is a randomized, double-blind, placebo-controlled trial in healthy volunteers in Adelaide, reading safety/tolerability + exploratory autophagy biomarkers (LC3, p62, lysosomal flux). As of May 2026, CEO says it is "going super good" with no dose-limiting toxicities.

Critical caveat for reading the August print: this is a healthy-volunteer safety/PK study, not an efficacy study. It can show RTR242 is safe and that biomarkers move; it cannot show it slows Alzheimer's. Treat "positive August data" as de-risking the molecule, not validating the thesis. The field's reality: as of April 2026, no cellular-rejuvenation peer has reported any human efficacy data — the entire field "remains at the hypothesis stage in humans".

Lens 6 · Founder interviews / narrative trend (+private swap of "Earnings Calls")

There are no earnings calls. The sentiment proxy is the founder/funding narrative arc, and it has shifted in a telling way:

• 2023 (stealth-exit): mission-grandiose — "add 10 years," Altman's whole net worth, three moonshot programs.
• Dec 2025: maximally promotional — chasing a $5B valuation on slides claiming "longevity will be the greatest pharma market of all time" and that Retro could "surpass Eli Lilly and Novo Nordisk," with no clinical data in hand.
• May 2026: notably more grounded — the round priced at $1.8B, the language shifted to "repeatable process," "multiple successful FDA interactions," and concrete first-in-human milestones for 2026/2027.

The tone trend is the story: Retro moved from "we will out-pharma pharma" (Dec 2025) to "we have a process and a clinical asset" (May 2026) — after the market refused the $5B ask. That is a healthy maturation, but it is also the market forcing discipline onto the narrative. What they stopped saying: the Eli-Lilly-beating TAM rhetoric. What they started saying: FDA interactions and dated milestones.

Lens 7 · Cap table & secondary marks + mechanism comps (+private & +clinical swaps of "Comps")

Funding & valuation trajectory — all ``, unaudited:

The headline number that matters: the $1.8B mark is a ~64% haircut to the $5B ask of six months earlier. The round closed, which is itself a positive in a brutal 2025-26 biotech funding climate — but it closed well below the founder's ambition.

Syndicate quality — the IPO-proximity tell is weak. Per the +private overlay, a Fidelity / T. Rowe / Coatue-type crossover entry signals IPO proximity. Retro's named institutional backers — 4P Capital, Immortal Dragons, Mana Ventures, Prosto Venture Club — are specialist/longevity-thematic and smaller-name funds, not tier-1 crossover capital. That is a meaningful signal: as of mid-2026, the blue-chip public-market-adjacent money has not anchored this name. IPO is not near.

Mechanism comps (by target/thesis, not P/E) — the relevant peer set:

Comp read: Retro is the second-most-capitalized pure reprogramming play (behind Altos), and arguably the furthest along clinically among the Altman/Bezos/Armstrong cohort (it has a Phase 1 asset; Altos and NewLimit do not disclose one). But Life Biosciences is also in Phase 1, and BioAge is the live cautionary tale of how fast an aging-biotech can lose half its value on a halted trial. EV/Sales, EV/EBIT, P/E: n/a — pre-revenue, private. Do not fabricate a multiple.

Lens 8 · Funding/product-event catalysts (+private swap of "Stock-Price Catalysts")

No public stock, so the catalysts are valuation/narrative-moving events. The pattern of what has actually moved Retro's perceived value:

• 2023 stealth exit + Altman $180M → put the name on the map.
• Aug 2025 OpenAI/GPT-4b 50x reprogramming result → the single biggest credibility catalyst; converted "longevity moonshot" into "AI-bio platform with a real lab result".
• Dec 2025 $5B chase → over-reach; the market pushed back.
• Dec 2025 first RTR242 dose → became a clinical-stage company, the key de-risking milestone.
• May 2026 $1.8B close → validation that capital is available, at a chastened price.

What the "market" reacts to for this name: (1) hard lab/clinical milestones (the OpenAI result, the first dose) and (2) the Altman halo. It does not reward TAM rhetoric — the $5B-on-slides ask was refused. The next catalyst, August 2026 RTR242 data, fits the pattern: a hard data event that will move the mark up or down.

---

Phase C — Judge people & books

Lens 9 · Management

CEO — Joe Betts-LaCroix. This is the strongest single line in the bull case. He is a genuine repeat builder, not a longevity tourist:

• Track record (quantified): co-founded OQO (world's smallest Windows PC, Guinness 2006; ~110 staff, 100+ patents, 10,000+ units; sold to Google) and Vium (AI/automation for in-vivo research; acquired by Recursion Pharma). Scientific pedigree: Harvard/MIT/Caltech, co-discovered protein electron-tunneling factors (~1,500 citations), 80+ patents.
• Aging credibility: founded the Health Extension Foundation (2012) and was a Y Combinator partner (2017, 2.5 yrs) — where he met Altman. The Altman relationship is earned, not transactional.
• Skin in the game: founder; ownership undisclosed but board is just him + Altman.
• Capital-allocation read: built in-house cGMP manufacturing early (capital-intensive, vertical-integration bet) and signed selective partnerships (Multiply $85M, MCRI $35M) rather than spraying capital. Aggressive but coherent.
• Co-founders: Sheng Ding — top-tier stem-cell chemist (Gladstone/UCSF, scientific co-founder of six biotechs) is real scientific weight; Matt Buckley — solid operator (ex-Illumina, Stanford genetics PhD).

Red flags (governance): (1) Two-person board (Betts-LaCroix + Altman) — minimal independent oversight for a now-$1.8B company; this should change post-institutional round. (2) The Altman conflict node — he funds Retro and runs OpenAI, the sole supplier of Retro's AI-protein moat; the GPT-4b collaboration is genuinely valuable but the relationship is not arm's-length. (3) Promotional drift — the $5B / "beat Eli Lilly" slides (Dec 2025) were classic hype-cycle behavior; the May 2026 reset is reassuring but the tendency is on record.

Archetype: founder-operator with a real exit history and deep field roots — the favorable archetype for this stage. The governance is the weak spot, not the operator.

Lens 10 · Forensic Red Flags + Science & exclusivity (+clinical add)

Standard forensic accounting analysis is n/a — no audited financials, no GAAP statements, no 10-K. "Unaudited per public sources." The forensic lens re-points (per the +clinical overlay) to trial-design integrity, going-concern, dilution, and scientific exclusivity:

• Trial-design integrity: RTR242 Phase 1 is healthy volunteers, double-blind, placebo-controlled, at a single Adelaide site, reading exploratory autophagy biomarkers. The integrity risk is over-reading exploratory biomarkers (LC3/p62) as efficacy. The honest framing: a clean safety/PK readout is necessary but nowhere near sufficient.
• Going-concern / runway: $180M (2022) was estimated to fund ops to ~2030; the May 2026 Series A-II extends that. Burn rate is not disclosed `, unverified]. Drug development to approval is estimated at ~$6.1B with ~8% Ph1→approval success — i.e. Retro will need many more rounds.
• Dilution: the $5B→$1.8B reset means existing holders (chiefly Altman) took a markdown; future rounds at this scale imply continued heavy dilution. No share count disclosed.
• Scientific exclusivity / IP: the RetroSOX/RetroKLF enhanced factors and the GPT-4b-micro pipeline are the crown-jewel IP. Patent estate specifics are undisclosed. The mechanism class (autophagy in AD) is not exclusive — hydralazine (Ph3 AD) and nilotinib (Ph3 AD) are autophagy activators already in late-stage AD trials, so RTR242 competes in a validated-but-crowded mechanistic space.

Regulatory findings (required sub-section).

• SEC (EDGAR EFTS — LR + AAER): No findings. Retro has no CIK and is not an SEC filer — no EDGAR enforcement search is possible.
• Non-SEC (web search — FTC/DOJ/FDA/CFPB/consent-decree/settlement/penalty): No material enforcement actions, consent decrees, fines, or penalties against Retro Biosciences surfaced in web search as of 2026-06-24. The only FDA contact disclosed is constructive — "multiple successful interactions with the FDA" on RTR242.
• Item 3 (Legal Proceedings): n/a — private, no 10-K.
• Conclusion: No material regulatory or legal findings — verified via SEC EDGAR EFTS (LR, AAER, nil — no CIK), web search (nil), and noting no 10-K exists, as of 2026-06-24. All findings unaudited per public sources.

---

Phase D — Project & stress-test

Lens 11 · IPO-readiness & path-to-tradeable + rNPV framing (+private & +clinical swaps of "Forward Projection")

No EPS projection — pre-revenue, no P&L. Per the overlays, the questions that matter are runway-to-catalyst and path-to-tradeable.

rNPV framing (lead asset RTR242) — illustrative ``, every input a guess given zero disclosure:

• Alzheimer's is a multi-$10B market, but RTR242 is Phase 1, healthy volunteers, no efficacy signal. Applying an honest Phase-1-CNS PoS (~5–8% to approval ) against even a successful-drug peak, the risk-adjusted NPV is dominated by the discount, not the headline market. Any rNPV here is a `` with error bars wider than the point estimate — I will not manufacture a dollar figure that would imply false precision. rNPV: n/a — not credibly sourceable pre-efficacy.
• The question that actually matters — does cash reach the next inflection? Yes for the immediate one: the May 2026 round funds the company through the August 2026 RTR242 readout and the 2026/2027 first-in-human milestones for the cell programs. The longer-horizon answer (does it reach pivotal data) is no without several more large raises — consistent with the ~$6.1B-to-approval and ~8% success figures.

IPO-readiness: Low / early. Markers that would unlock an S-1 — positive efficacy data in a lead program, a tier-1 crossover syndicate, and a clean governance structure — are absent. The syndicate is specialist/thematic (4P Capital, Immortal Dragons, Mana, Prosto), not crossover; the board is two people; there is no efficacy data. Estimated path-to-tradeable: multi-year (2028+ at the earliest, contingent on the reprogramming programs reaching the clinic and at least one asset showing human efficacy). ``

No forecast.ts create in --watchlist mode (per skill). The scoreable forecast to log later, when conviction is committed, is binary: "RTR242 Phase 1 reads out with no dose-limiting toxicity AND ≥1 statistically credible autophagy-biomarker movement by 2026-09-30."

Lens 12 · Bull vs Bear

Bull case. Retro is the best-executed of the celebrity-backed reprogramming cohort. It is the only one in the Altman/Bezos/Armstrong group with a disclosed Phase 1 asset (RTR242), it built a genuinely differentiated AI-protein-engineering pipeline with OpenAI (50x reprogramming-marker gain, the field's standout AI-bio result), it is run by a founder with two real exits (OQO→Google, Vium→Recursion), and it closed a $1.8B round in a hostile biotech funding market. It is a platform — five shots at aging — so it is not a single-binary bet. If the August data is clean and the reprogramming programs reach the clinic in 2026/27, this is a top-3 longevity franchise with optionality across Alzheimer's, blood disorders, and tissue rejuvenation.

Bear case (permanent-impairment risks).

1. The whole field is pre-efficacy. No reprogramming peer has shown any human efficacy; "hypothesis stage in humans". Retro could execute flawlessly and still find the biology doesn't translate — the BioAge halted-trial precedent shows how fast value evaporates.
2. The $5B→$1.8B markdown is a real signal, not noise — sophisticated capital repriced this name down 64% in six months. The smart money is skeptical of the valuation, and the syndicate that did show up is second-tier.
3. RTR242's mechanism is crowded and the indication is a graveyard. Alzheimer's drug development is "a never-ending string of failures"; autophagy activators (hydralazine, nilotinib) are already in Phase 3 with no approval. Retro is not first in this mechanism.
4. Capital intensity is existential. ~$6.1B-to-approval, ~8% success — Retro must raise repeatedly, diluting heavily, in a funding environment that just marked it down.

Pre-mortem (18 months out, thesis broke): The August 2026 RTR242 data came back "safe but biomarkers ambiguous"; the market read it as a non-event. The reprogramming IND slipped. With no efficacy and no new catalyst, the next round priced flat or down from $1.8B, the second-tier syndicate couldn't lead a larger raise, and Retro entered a slow down-round grind — the classic longevity-biotech "great science, no clinical proof, capital fatigue" trap.

Are the multiples too high? There are no multiples — it's a private mark. But $1.8B for a single Phase-1 healthy-volunteer asset plus preclinical programs is rich on absolute terms, justified only by platform optionality + the Altman/OpenAI halo. The market already said so by refusing $5B.

Contrarian view (what the market refuses to see): The bear consensus fixates on "no efficacy data, hype valuation." What it under-weights is that Retro's real asset is the OpenAI AI-protein pipeline, not RTR242. If GPT-4b-micro-class models keep compounding reprogramming efficiency (50x is a 2025 result), Retro's discovery engine — not any single drug — is the durable value, and it is mispriced as "a longevity drug company" rather than "an AI-native therapeutics platform." That is the asymmetric upside the $1.8B mark doesn't capture.

Lens 13 · Devil's Advocate (short-seller)

Dismantling the bull case:

• What structurally breaks the model: the core scientific bet — that partial reprogramming / autophagy restoration safely and durably reverses aging biology in humans — is unproven and may be wrong. Yamanaka factors are pro-oncogenic (tumor risk is the field's defining hazard); epigenetic "age reversal" headlines "rarely match peer-reviewed statistics". If reprogramming can't be made safe in humans, three of Retro's five programs are impaired at once.
• Concentration risk: value is concentrated in one near-term asset (RTR242) and one privileged relationship (OpenAI/Altman). If RTR242 disappoints and the OpenAI collaboration cools (or the Altman conflict becomes a governance problem), the story loses both its near-term catalyst and its moat narrative simultaneously.
• Most dangerous competitor bulls underestimate: Altos Labs — $3B+, disease-agnostic, and it out-recruited Retro for top reprogramming talent. If Altos reaches the clinic with a better-capitalized program, Retro's "furthest along" edge evaporates. Secondarily Life Biosciences, already in Phase 1.
• Worst capital-allocation / governance moves: the two-person board, the $5B promotional slides with no data, and the non-arm's-length Altman/OpenAI dependency — exactly the related-party + weak-oversight pattern a short looks for.
• What must hold for $1.8B: continued access to large external capital at non-punitive valuations + at least one program producing credible human signal within ~24 months. If growth/progress disappoints by 20–30% (e.g., RTR242 ambiguous + reprogramming IND slips), a down round well below $1.8B is the base case, not a tail.
• Single permanent-impairment scenario: a reprogramming safety event (tumorigenesis) in any company's human trial would tar the entire mechanism class and could strand Retro's reprogramming portfolio regardless of its own execution. Plausibility: moderate — it's the field's most-discussed risk for a reason.

Lens 14 · Management Questions (ordered by information value)

1. RTR242 August readout: what specific, pre-registered biomarker thresholds (LC3, p62, lysosomal flux) would you consider a positive signal vs. a non-event — and are those endpoints powered, or exploratory? (The whole near-term thesis turns on this.)
2. What is your current monthly cash burn and runway post the Series A-II, and at what milestone do you next need to raise?
3. The round priced at $1.8B vs. the ~$5B sought in Dec 2025 — what changed in investor diligence, and what does the second-tier syndicate (vs. a tier-1 crossover) imply for your next raise?
4. Governance: when does an independent board form, and how do you manage the Altman/OpenAI conflict (funder + sole AI-supplier) at arm's length?
5. Is the GPT-4b-micro / OpenAI protein-engineering relationship exclusive to Retro, and for how long? What happens to the reprogramming moat if it isn't?
6. On tumorigenicity — what is your specific safety strategy for partial reprogramming in vivo, and what preclinical evidence retires that risk before first-in-human?
7. Of the five programs, which two are you willing to kill to concentrate capital, and what would trigger that?
8. RTR242 competes with autophagy activators (hydralazine, nilotinib) already in Phase 3 for AD — what is your differentiated mechanistic claim, and where's the evidence?
9. What are the IND timing and gating criteria for the microglia (2026) and HSC (2027) programs — and what's the realistic slippage?
10. How do you convert "multiple successful FDA interactions" into a concrete regulatory path — what did the FDA actually agree to?
11. What is the economic structure of the Multiply Labs ($85M) and MCRI ($35M) deals — milestones, royalties, IP ownership?
12. What is your in-house cGMP capacity and cost, and was vertical integration the right capital allocation at this stage vs. outsourcing?
13. Peak-sales / rNPV: what does your internal model assume for RTR242, and what PoS do you apply?
14. What is the path to revenue before approval — partnership/licensing milestones, and on what timeline?
15. What would make you walk away from the reprogramming thesis entirely?

---