An Australian tech entrepreneur used AI to help create the first-ever bespoke cancer vaccine for a dog to treat his beloved pet Rosie

A Sydney entrepreneur used AI tools (ChatGPT, AlphaFold) and UNSW collaboration to design a bespoke mRNA cancer vaccine for his dog, developed in under two months. Rosie received the first injection in December and a booster in February, with most tumors shrinking and noticeably improved quality of life, though some tumors have not responded and a cure is not claimed.

Analysis

AI-driven design workflows are lowering the fixed costs and calendar time for generating candidate nucleic-acid therapeutics; expect candidate-design cycles to compress from industry norms of 6–18 months to sub-2-month iterations for sequence selection and in-silico triage. That will shift the bottleneck downstream to GLP toxicology, regulatory validation and GMP manufacturing capacity — those three stages will determine commercial velocity and who captures margin as bespoke treatments scale.

Competitive winners are likely to be platform owners that control both high-performance compute and model IP, plus scalable CDMOs and LNP/lipid suppliers that can convert designs into regulated doses on demand; traditional mid-sized oncology players that rely on a small number of blockbusters risk margin pressure if payors push bespoke-priced therapies into value-based models. Expect a proliferation of niche CRO/CDMO entrants focused on low-volume, high-complexity mRNA manufacturing — this will bifurcate the supplier market and compress pricing power over 2–5 years.

Key tail risks: regulatory pushback on AI-designed biologics (data provenance, explainability), immunogenicity/adverse-event clusters that trigger clinical holds, and an IP morass around algorithmically-derived sequences that could lead to multi-year litigation. Near-term catalysts include first controlled human proof-of-concept readouts (6–18 months) and clarifying guidance from regulators on algorithmic design and data ownership, either of which could rapidly re-rate incumbents and service providers.

The consensus underestimates scaling friction: clinical validation, batch-release testing, and reimbursement negotiation are multiyear processes that will cap revenue growth even if design tools proliferate quickly. Monitor patent filings, CDMO capex announcements and initial human efficacy reads as high-signal indicators for when adoption translates into durable revenue.

Key Decisions for Investors

• Long GOOGL (12–24 months): buy GOOGL LEAP calls (12–24 month tenor) to capture optionality from platform AI licensing and cloud revenue tailwinds if biopharma workflows standardize on Big Tech models; downside = option premium, upside = 2–3x if adoption accelerates and monetization follows.
• Long Thermo Fisher Scientific (TMO) or Catalent (CTLT) (6–18 months): buy calls or 3–6% notional overweight in cash/equity to play manufacturing and analytics demand from bespoke biologics; reward = mid-teens revenue acceleration if mRNA/CDMO demand grows, risk = 15–25% downside in macro slowdown.
• Pair trade (12 months): long TMO / short CRISPR Therapeutics (CRSP) — expresses preference for service providers over binary R&D-risk biotech. Rationale: capture steady demand for capacity vs high binary clinical/regulatory risk; target asymmetry = limited upside on short if CRISPR clears catalysts, large capital preservation if platform demand wins.
• Event-watch small long (6–18 months): buy CTLT Jan-2027 $40 calls (or equivalent) ahead of expected CDMO capacity announcements and margin-accretive contracts; set stop-loss at 30% to limit downside from macro/clinical shocks.