Techie uses ChatGPT and AlphaFold to build DIY mRNA cancer vaccine, saves dog

Australian tech entrepreneur Paul Conyngham used ChatGPT and DeepMind’s AlphaFold to design a bespoke mRNA vaccine for his dog after chemotherapy failed; genomic sequencing cost AUD 3,000 and he spent roughly three months preparing a 100-page ethics submission. Rosie received a first injection in December followed by boosters and the tumor has shrunk significantly, though Conyngham cautions this is not a proven cure and instead appears to have extended the dog’s survival and quality of life. He is now developing a second vaccine targeting a remaining tumor.

Analysis

This episode is an operational proof-point that low-cost AI + accessible sequencing can compress the timeline from variant discovery to candidate mRNA design from months to weeks for single-case therapeutics. That compression disproportionately benefits platform owners who control validated model-to-lab workflows and compliance stacks (cloud providers, CDMOs, large sequencers), because liability and GMP/ethical gating will force most commercialization through accredited partners within 6–24 months. Expect demand to bifurcate: bespoke, high-margin personalized therapeutics (premium to platform providers) and a larger long-tail of regulated, recurring diagnostic/sequencing services.

Key regulatory and biosecurity levers will determine whether this becomes a decentralised maker movement or a re-centralised industrial market. In the near-term (0–12 months) expect noise—guidance, export controls on oligos/enzymes, and insurance product emergence—that can intermittently shock small suppliers; over 12–36 months, formal accreditation and reimbursement pathways will lock-in winners with existing QA/QC, manufacturing scale, and IP portfolios. A single adverse regulatory or liability ruling could remove a large fraction of nascent entrants within a quarter.

The market consensus that “AI democratises drugmaking” understates the capture mechanics: revenue accrues to integrators who bundle compute, validated models, and GMP execution, not to hobbyist toolmakers. That creates asymmetric opportunities to own the plumbing (sequencing, sample logistics, regulated CDMOs, cloud infra) and to avoid single-incident biotech small-caps that lack compliance moats. Position sizing should reflect binary regulatory tail risk—small-cap exposure is high-volatility, while platform/CDMO exposure is slower but more defensible.

Key Decisions for Investors

• Buy Thermo Fisher (TMO) shares, 6–18 month horizon — rationale: beneficiary of higher recurring sequencing and regulated reagent/CDMO spend. Target +15–25% if adoption of custom biologics accelerates; downside -12–18% if capex slows. Hedge with 6–12 month protective puts at ~20% OTM if available.
• Buy Alphabet (GOOGL) 12–18 month call exposure (prefer LEAPS or 2:1 call spread) — rationale: cloud + model deployment fees and enterprise integrations into life-sciences workflows. Risk/reward ~3:1 if healthcare AI monetization narratives re-rate; key risk is tighter antitrust/regulatory scrutiny in 6–12 months.
• Pair trade: Long Moderna (MRNA) shares, short Twist Bioscience (TWST) shares, 6–12 month horizon — rationale: long mRNA platform execution vs short DNA/oligo pure-plays vulnerable to export controls and compliance-driven demand shifts. Target asymmetric return where MRNA captures platform premium (+25%) while TWST de-rates (-30%) under regulatory tightening; position size capped to 1–2% NAV each leg.
• Event hedge / catalyst trade: buy a small position in Thermo Fisher or Lonza-equivalent CDMO (regional ADRs if preferred) ahead of expected regulatory guidance on gene synthesis controls (watch 3–9 month window). If guidance tightens, expect 10–20% snap move higher in compliant CDMOs and larger sequencers.