Tal Zaks, ex-Moderna CMO, explains biotech investing: science is unpredictable, but mRNA is like 'information software' – change the sequence, get a new drug. Moderna's platform scored 8 wins in a row on vaccines. He's bullish on personalized cancer vaccines (halved relapse risk in Phase II) and Exilio, where he's interim CEO, using mRNA for gene editing. Warning: gene editing raises ethical debates that may slow adoption.
Tal Zaks, former Chief Medical Officer of Moderna, transitioned from a physician-scientist to a biotech investor. In this episode, he shared the key challenges of translating scientific breakthroughs into profitable drugs. Drawing on his experience with the Moderna mRNA platform in the development o
Tal Zaks (former Chief Medical Officer of Moderna, now a Partner at OrbiMed) has transitioned from physician-scientist to biotech investor, sharing the core challenges of translating scientific breakthroughs into profitable drugs. Key judgment: Biotech investment returns are constrained by the unpredictability of biology and pharmacology, but platforms such as mRNA, through their "information as medicine" property, are shifting certain steps from "biological experiments" toward "engineered reproducibility," while public trust and cross-disciplinary collaboration form the foundation of long-term innovation.
Tal Zaks argues that biotech investment returns are determined by three variables: required capital, time horizon, and probability of success, with the most difficult to predict being "whether the biology holds" and "whether the pharmacology works." This is fundamentally different from hard-tech investing—engineers can predict "whether a moon landing is possible" and allocate resources accordingly, but when faced with "can we cure cancer," a doctor can only answer "I don't know, we have to try." He cites OrbiMed's investment philosophy: "People give us money, and a few years later we have to give them back more, otherwise nothing else matters." Investment decisions must be broken down step by step: whether the target is relevant (biology), whether the molecule can intervene (pharmacology), and whether the clinic can confirm it (development). He uses a vaccine project as an example: the early concept seemed reasonable, but it had to go all the way to Phase III for validation, and the capital and time required did not match the risk, so it was ultimately abandoned. Signal validation: If early data (e.g., Phase I) cannot significantly reduce uncertainty, the project is not viable.
Zaks attributes Moderna's success to the combination of "engineer thinking" and "doctor thinking"—engineers have a vision and a process, while doctors are accustomed to experimental validation. Key data point: By early 2020, Moderna had used its mRNA platform to successfully generate neutralizing antibodies in humans against 8 different viruses, achieving "8 wins out of 8"—unprecedented in drug development history. The speed of COVID-19 vaccine development benefited from prior preparation: in September 2019, Zaks and CEO Stefan Bancel visited the NIH, and Fauci described the platform as "the best vaccine platform I have ever seen," leading to the initiation of a demonstration project; in January 2020, Stefan was the first to identify the outbreak, and the company quickly pivoted. Mechanism breakdown: The essence of mRNA is "information medicine"—the physical structure is identical, only the information sequence changes, resulting in extremely low marginal costs. This software-like property allows the platform to iterate rapidly, but it also requires enough "freedom" to explore multiple applications in parallel (vaccines, rare diseases, oncology). Extrapolation: In the future, nucleic acid drugs will expand to therapeutic vaccines, gene editing, etc., but they will need to address delivery (lipid nanoparticles targeting different tissues) and public trust.
Zaks believes that personalized cancer vaccines are his proudest scientific contribution at Moderna, with Phase II data showing a reduction in recurrence risk of approximately 50%. The principle: For early-stage cancer patients (e.g., melanoma), after tumor resection, there is still a risk of recurrence. By analyzing the patient's tumor-specific mutations, a personalized mRNA vaccine is manufactured to activate the immune system to eliminate residual cancer cells. Historical context: Zaks attempted mRNA cancer vaccines (in mice) at the NIH in the late 1990s, but it was not until 20 years later that Moderna's manufacturing capabilities (rapid small-batch production) and sequencing technology matured enough to make clinical application feasible. Key data point: A Phase II randomized controlled trial showed that the recurrence rate in the vaccine group was reduced by approximately 50% compared to the standard-of-care group; Phase III enrollment is ongoing. Signal validation: If Phase III succeeds, it will open a new paradigm in oncology—shifting from "standardized chemotherapy" to "individualized immunotherapy."
Zaks categorizes mRNA, siRNA, and gene editing collectively as "nucleic acid drugs," whose core feature is encoding information into the drug—changing the information changes the drug, making it essentially "software-based biology." Exilio, where he currently serves as interim CEO, is using mRNA delivered in lipid nanoparticles to carry gene-editing tools, enabling a one-time treatment to permanently correct genetic defects. Mechanism breakdown: Traditional gene therapy (e.g., AAV) delivers a protein, whereas nucleic acid drugs deliver information, allowing the body to produce the required protein itself. mRNA is transient (requiring repeated dosing), while gene editing can permanently alter the genome. Risks and extrapolation: This raises new ethical controversies—is it acceptable to intentionally alter human DNA? Zaks notes that mRNA itself does not change DNA (based on first principles), but gene-editing technology does. Public fear of "gene modification" requires ethical dialogue and may affect regulatory acceptance. Uncertainty: The technology is feasible, but social consensus and ethical frameworks could become a slower bottleneck than science.
Zaks takes a "cautiously optimistic" stance on AI's impact in healthcare: it has potential, but for now, it is more of a fragmented application in "thin verticals" than a systemic transformation. Example: AI-based image reading (pathology) has received FDA approval and outperforms humans in accuracy, but fewer than 1% of pathology slides have been digitized—because there is no economic incentive to replace microscopes. Mechanism breakdown: He cites research by Israeli scientist Amos Tanai: using AI to perform "personalized correction" on routine blood tests could eliminate about half of the "normal value variation," thereby detecting abnormalities earlier. The problem, however, is whether there is a validated intervention for early detection. Extrapolation: Bear case: AI penetrates slowly only in individual verticals (e.g., nurse transcription, clinical trial matching) and struggles to integrate. Bull case: System-level integration (e.g., linking electronic medical records with AI diagnostics) drives productivity gains, but this requires redesigning incentives, and current EMR systems have actually reduced productivity. Zaks believes that digital twins and full simulation remain science fiction because human health involves ethical autonomy and cannot be fully reduced to utilitarian calculations.
| 标的 | 嘉宾态度 | 关键数据 |
|---|---|---|
| Moderna | Insider with firsthand experience, highly regards platform capabilities | 8 wins out of 8 before COVID; Phase II personalized cancer vaccine reduced recurrence rate by 50% |
| Exilio | Current interim CEO, bullish on gene editing potential | Uses mRNA+LNP delivery to achieve permanent gene repair, still early stage |
| Teva | Mentioned only as board member, not an investment target | World's largest high-quality generic drug manufacturer, emphasizes long-term social benefits |
1. Tal Zaks: "Biotech investment returns = f(capital, time, probability), and the unpredictability of biology and pharmacology is the biggest variable." — Unlike hard tech, where engineers can plan a path to the moon, doctors can only answer "I don't know, we have to try" when asked to "cure cancer."
2. Tal Zaks: "The mRNA platform is a 'software-ized drug' — changing the sequence of information yields a different drug while the physical structure remains the same, with extremely low marginal cost." — This explains why the platform can achieve 8 wins in 8 attempts, and why COVID vaccine development was unprecedented in speed.
3. Tal Zaks: "In the Phase II trial, the personalized cancer vaccine reduced recurrence risk by about 50% — this was the dream I had when I was doing experiments at the NIH 20 years ago, but it only became possible when manufacturing capacity, sequencing, and understanding of tumor immunology matured." — Scientific breakthroughs require cumulative time, not linear progress.
4. Tal Zaks: "Nucleic acid drugs (mRNA, siRNA, gene editing) are creating a new pharmacology paradigm, but deliberately altering human DNA triggers ethical debates, which may be a bigger bottleneck than the technology itself." — mRNA does not change DNA, but gene editing does; public trust and ethical frameworks will determine how quickly the technology is applied.
5. Tal Zaks: "The bear case for AI in healthcare is slow penetration in thin verticals; the bull case is system integration boosting productivity, but for now EMR systems have actually lowered productivity." — He cites the work of Amos Tanai: AI can cut the "normal variation" in routine blood tests by half, but early detection still requires effective intervention tools.
6. Tal Zaks: "The COVID vaccine is the most thoroughly studied medical intervention in history, with AE reporting rates 10–20 times higher than previous vaccines, but the safety data is clear — the incidence of rare myocarditis is about 1 in 150,000, and it was promptly added to the label." — The public trust problem stems from scientists overstepping boundaries and failing to fully respect ethics and autonomy.
7. Tal Zaks: "The most common reason for investment failure is underestimating the unpredictability of the 'magnitude of clinical benefit' — even if the biology and pharmacology are correct, the benefit is often insufficient to achieve differentiation." — This echoes the "hard rock" analogy: investors always think there is a soft core in the middle, but in reality the entire path is solid rock.
8. Tal Zaks: "The best investors learn more from success than from failure alone. When success occurs, they must ask: Is it luck, talent, or a combination? Pattern recognition behind success is what is sustainable." — He quotes Nelson Mandela: "Either you win or you learn," but notes that success also needs to be deconstructed.