【The Motley Fool】Moderna CEO：mRNA将颠覆癌症治疗，股价曾暴跌95%我却坚定加仓！

浑水摸鱼清源

AI-assisted summary of 【The Motley Fool】Moderna CEO：mRNA将颠覆癌症治疗，股价曾暴跌95%我却坚定加仓！ — original source. Check the original for full context.

Source published: 2026-10-25

Key ideas

Moderna在黑色素瘤二期临床中与Keytruda协同治疗，随访5年实现80%患者无复发生存，在皮肤癌领域已达临床治愈标准。 (1:47)
作为跨疾病mRNA平台公司，Moderna的7项三期试验累计成功率达约85%，远高于传统药企约10%的研发成功率。 (5:00)
二期数据显示患者间99%基因与90%肿瘤突变均不相同，证实癌症个性化疫苗为刚需，公司正通过算法与小型反应器推进量产。 (10:00)
Moderna利用AI将三期临床数据整理及提交FDA周期从行业平均6个月压缩至3个月，终极目标是缩短至1天以极大提升资产净现值。 (14:44)
面对股价自497美元暴跌95%至23美元，CEO Stéphane Bancel在29美元全额自费行权增持，强调决策只关注未来10年现金回报。 (17:12)

Detailed notes

## Background 本内容由B站UP主“浑水摸鱼清源”于2026年10月25日发布，转译并解读了The Motley Fool主办的FoolFest活动中对Moderna首席执行官Stéphane Bancel的深度访谈。视频简介指出，许多人对Moderna的认知仍局限于新冠疫苗，而Bancel在访谈中展现了mRNA作为信息分子在癌症治疗、罕见病、AI赋能研发及生物安全防御等领域的宏大版图，并回应了公司股价暴跌95%背景下的长期投资逻辑。

## TL;DR 访谈中，Moderna首席执行官Stéphane Bancel阐述了mRNA作为“人体软件代码”的平台型药企本质，指出其通过调整四个碱基编码不同蛋白的特性将临床研发成功率提升至约85%。在肿瘤领域，Moderna的个性化癌症疫苗与Keytruda协同攻坚黑色素瘤，二期临床实现了术后5年80%的无复发生存率，且正将该技术横向扩展至肺癌等其他癌种与罕见遗传病。Bancel详细对比了Moderna与BioNTech在化学修饰、递送系统及算法壁垒上的本质差异，并展示了AI如何将三期临床数据提交周期压缩至数月甚至未来的单日水准。面对2021至2025年间股价自高点暴跌95%的市场动荡，Bancel表示自己选择在低位坚决行权增持，强调其管理决策完全聚焦于未来10年的现金回报与生命科学底层价值的长期复合增长。

## Atmosphere & style 整场对话在轻松风趣的技术科普开场与严谨厚重的商业战略思辨之间自如切换，主持人以幽默形象的提问破除晦涩的生化门槛，而Bancel则展现出一名兼具工程师理性与创业者坚韧的企业领袖特质。谈及二期临床80%无复发生存与拯救生命的个性化疗法时，对话充满对现代医学奇迹的由衷敬畏；探讨AI生物武器威胁时语调急转直下、凝重警惕；而在直面股价崩跌95%与华尔街短期质疑时，Bancel平静笃定的回答透露出纯粹的长期主义底气，整体氛围既有前沿科技的澎湃张力，又兼具穿越资本周期的从容与清醒。

## Bulletin - [01:47](https://www.bilibili.com/video/BV1XpHw6DEz5?t=107s) Moderna在黑色素瘤二期临床中与Keytruda协同治疗，随访5年实现80%患者无复发生存，在皮肤癌领域已达临床治愈标准。 - [05:00](https://www.bilibili.com/video/BV1XpHw6DEz5?t=300s) 作为跨疾病mRNA平台公司，Moderna的7项三期试验累计成功率达约85%，远高于传统药企约10%的研发成功率。 - [10:00](https://www.bilibili.com/video/BV1XpHw6DEz5?t=600s) 二期数据显示患者间99%基因与90%肿瘤突变均不相同，证实癌症个性化疫苗为刚需，公司正通过算法与小型反应器推进量产。 - [14:44](https://www.bilibili.com/video/BV1XpHw6DEz5?t=884s) Moderna利用AI将三期临床数据整理及提交FDA周期从行业平均6个月压缩至3个月，终极目标是缩短至1天以极大提升资产净现值。 - [17:12](https://www.bilibili.com/video/BV1XpHw6DEz5?t=1032s) 面对股价自497美元暴跌95%至23美元，CEO Stéphane Bancel在29美元全额自费行权增持，强调决策只关注未来10年现金回报。

## Key takeaways - [00:25](https://www.bilibili.com/video/BV1XpHw6DEz5?t=25s) mRNA本质是由四种碱基构成的信息分子，仅需调整排列即可编码人体任何蛋白质，底层工艺通用性打破了传统药物研发模式。 - [01:47](https://www.bilibili.com/video/BV1XpHw6DEz5?t=107s) 癌症个性化疫苗负责为免疫系统提供癌细胞新突变特征以生成新T细胞，与负责给旧T细胞供能的Keytruda形成协同作战。 - [05:00](https://www.bilibili.com/video/BV1XpHw6DEz5?t=300s) 传统生物制药基于孤立分子试错导致成功率仅约10%，而Moderna作为模块化工程平台实现了约85%的三期研发成功率。 - [06:40](https://www.bilibili.com/video/BV1XpHw6DEz5?t=400s) 生物医药资产价值取决于销售峰值、上市周期倒数与成功率的乘积，mRNA平台的复用性彻底改变了传统药企的现金流贴现模型。 - [10:00](https://www.bilibili.com/video/BV1XpHw6DEz5?t=600s) 针对癌症高度异质性（99%基因和90%突变各异），Moderna将水相生物反应器小型化，搭建起专属个性化疫苗的敏捷量产网络。 - [10:00](https://www.bilibili.com/video/BV1XpHw6DEz5?t=600s) Moderna针对林奇综合征（约100万美国患者无药可医）布局预防性疫苗，个性化底层制造能力正在向更多罕见病扩展。 - [11:47](https://www.bilibili.com/video/BV1XpHw6DEz5?t=707s) Moderna与BioNTech路线存在本质差异，两者在尿苷修饰、脂质载体、注射途径及突变选择算法等至少五个关键维度完全不同。 - [13:14](https://www.bilibili.com/video/BV1XpHw6DEz5?t=794s) 公司采用多层堆叠的“千层饼”式专利矩阵，且突变筛选算法作为核心商业机密高度保密，构筑起仿制药无法绕过的监管壁垒。 - [14:44](https://www.bilibili.com/video/BV1XpHw6DEz5?t=884s) AI被用于单管多mRNA复方设计以干预复杂多基因疾病，并助力三期临床数据提交FDA的耗时大幅压缩，极大提升研发净现值。 - [20:00](https://www.bilibili.com/video/BV1XpHw6DEz5?t=1200s) Moderna构建了AI提出科学假设、设计mRNA/脂质、动物实验验证并实时反馈模型重训的自动化研发飞轮。 - [20:00](https://www.bilibili.com/video/BV1XpHw6DEz5?t=1200s) AI辅助的功能增益呼吸道病原体被CEO视作远比核武器廉价且危险的生物安全威胁，亟需建立全球快速防御网络。 - [21:05](https://www.bilibili.com/video/BV1XpHw6DEz5?t=1265s) 在流感获批框架下，Moderna可在10分钟内完成新病毒序列mRNA设计，5至6个月内完成获批并具备交付10至20亿剂的产能储备。 - [23:01](https://www.bilibili.com/video/BV1XpHw6DEz5?t=1381s) 管理层完全摒弃取悦华尔街的季度EPS导向，所有资本配置与管线决策均严格锚定15年以来的10年期现金回报模型。

## Detailed walkthrough ### [00:00](https://www.bilibili.com/video/BV1XpHw6DEz5?t=0s) mRNA 技术本质与人体认知的冰山一角

访谈伊始，主持人开玩笑称虽然只有三个小时，但要直奔主题，并提出希望将 Moderna 的底层技术解释得通俗易懂，让一个讨厌科学的高中十年级学生也能听明白。Moderna 首席执行官 Stéphane Bancel 笑着回应道：“Come on, you cannot hate science in school.” 随后他解释了信使核糖核酸（mRNA）的基本原理：就像所有数字软件都是由 0 和 1 构成的代码一样，mRNA 是一种由四个化学字母（碱基）构成的信息分子。这四个字母的组合能够编码人体内的任何蛋白质。这种机制并不是 Moderna 发明的，而是整个地球生命运转的基础规律，广泛存在于植物、动物和人类细胞之中。

Moderna 在新冠疫情前整整探索了十年，所攻克的真正核心难关是在生物反应器中合成安全可用的 mRNA 分子，以此向身体细胞下发指令，按需生成特定的蛋白质。当主持人询问现代医学究竟对人体的运作机制理解了多少时，Bancel 坦言人类目前的认知可能只有 5%、10% 到 15%，因为“我们不知道我们不知道什么”，现有理解依然非常有限，但这也正是他感到极其振奋的原因。在未来的五到十年内，医学技术即将为人类带来爆发式的飞跃与突破。

- **mRNA 信息属性**：由 4 种碱基字母组合而成，负责向细胞传递合成特定蛋白质的代码指令。 - **Moderna 的技术攻关**：在疫情前历时 10 年，解决了在反应器内安全合成高保真 mRNA 药物分子的工程难题。 - **医学认知现状**：人类对人体底层生理机制的了解仅在 5% 至 15% 左右，存在巨大的探索与发展空间。

---

[01:47](https://www.bilibili.com/video/BV1XpHw6DEz5?t=107s) 癌症疫苗与 Keytruda 协同攻坚黑色素瘤

主持人将个性化癌症疫苗与默沙东的免疫检查点抑制剂 Keytruda（帕博利珠单抗）的协同机制比喻为一张“通缉令”：疫苗负责为机体标出癌细胞的模样，让免疫系统顺藤摸瓜去搜捕，而 Keytruda 则是给免疫系统发放开火权，允许其痛击罪犯。Bancel 对这一比喻进行了专业补充与修正：当患者长期患癌、肿瘤不断增大时，体内的免疫细胞会消耗殆尽陷入疲态，Keytruda 的核心作用是“重新给现存的 T 细胞注入能量”；而 Moderna 的癌症疫苗则是向免疫系统展示癌细胞新突变产生的特异性特征签名，促使机体产生全新的 T 细胞。

患者接种两到三针疫苗后，体内的免疫反应就会被彻底激活。从黑色素瘤的临床二期随访数据来看，在术后 5 年的时间节点上，有 80% 的患者保持无复发生存状态。Bancel 指出，在皮肤癌领域，5 年未复发在医学上通常已经被视为临床治愈。这一联合疗法展示了以 mRNA 为载体的个性化肿瘤免疫治疗具有巨大的临床转化潜力。

```mermaid flowchart TD accTitle: 癌症疫苗与Keytruda协同作用机制 accDescr: 展示mRNA疫苗与Keytruda如何分别通过生成全新T细胞和为旧T细胞供能协同杀灭肿瘤

subgraph ModAction ["Moderna 个性化癌症疫苗"] A["提取并测序肿瘤突变特征"] --> B["mRNA 编码特异性新抗原"] B --> C["接种 2 至 3 针疫苗"] C --> D["刺激免疫系统生成全新的特异性 T 细胞"] end

subgraph KeytAction ["Keytruda (帕博利珠单抗)"] E["识别疲劳衰竭的既有免疫细胞"] --> F["解除肿瘤微环境免疫抑制"] F --> G["为现存的耗竭 T 细胞提供能量补充"] end

subgraph JointResult ["双管齐下清除肿瘤"] D --> H["协同打击肿瘤细胞"] G --> H H --> I["临床二期黑色素瘤数据: 5年无病生存率达 80%"] end ```

---

[03:00](https://www.bilibili.com/video/BV1XpHw6DEz5?t=180s) 单一资产 vs. 平台型生物技术：底层研发范式的代际革命

针对外界对 Moderna 最大的误解——将其看作单纯依靠特定单品药物的传统药企，Bancel 从生物制药的发展史切入，深入阐明了什么是真正的“平台型公司”。回顾过去 150 年的制药史，行业从阿司匹林、抗生素、立普妥（Lipitor）等化学小分子药物起步，但这些都是孤立的“单发项目（onesies）”，每做一款新药都要重新摸索化学结构、安全性和有效性，研发几乎全是从零开始，因此绝大多数管线都以失败告终。到了 70 和 80 年代，以重组重叠技术、胰岛素、生长激素以及后来的 Keytruda 为代表的重组蛋白抗体生物药兴起，但开发每一种分子仍然需要经历独立且复杂的摸索过程。

Bancel 强调，正是由于 mRNA 是一种遵循通用规则的信息分子，只要调整四个碱基的排列顺序，就可以编码不同的蛋白质。他在 2011 年加入公司之初就敏锐地意识到：“We knew from the beginning that it was going to be zero drug or a lot of drugs.” 凭借工程师的专业训练背景，他从第一天起就将 Moderna 架构为一家平台公司，大规模布局 IT 信息系统和自动化机器人产线，让成功经验可以低边际成本快速复用。现在公司不仅有 5 款疫苗获批上市，癌症疫苗也已落地，年底还将揭晓罕见肝脏遗传病的临床数据。在平台范式下，只要攻克了向某个特定器官递送的技术，后续研发就变成了高效的“复制粘贴”。

- **化学药与重组蛋白时代**：结构各不相同，研发流程彼此独立，绝大部分候选分子在从零试错的过程中折戟。 - **信息分子的底层逻辑**：化学骨架与递送系统相同，仅需改变模板上的碱基顺序即可靶向全新蛋白质。 - **工程化平台架构**：通过引入工业自动化、信息技术与标准化流程，使技术跨适应症复制如同“复制与粘贴”。

---

[05:00](https://www.bilibili.com/video/BV1XpHw6DEz5?t=300s) 85% 临床成功率的平台红利与三大疾病扩张领域

在谈到这种平台范式的产出效果时，Bancel 给出了极具反差的对比数据：传统制药行业从临床一期到三期成功的整体成功率仅有 10% 左右；而 Moderna 截至目前共完成了 7 项三期临床试验（包括 5 款获批疫苗、1 项癌症三期阳性结果，以及巨细胞病毒 CMV 疫苗的一项未达终点试验），其平台研发成功率高达约 85%。Bancel 谦虚地表示：“And so it's not because we're smarter. I don't think we are. It's again because we have a platform that reduces the risk and increases the probability of drugs to go the whole way to approval.”

谈及下一步向更多治疗领域的纵向迁移，Bancel 详细列出了 Moderna 的战略路线图： 1. **实体瘤全面拓展**：目前正在肺癌、肾癌、膀胱癌等不同癌种中推进 9 项临床试验。因为 T 细胞在血液循环中如同巡逻的安保警察，根本不在乎肿瘤是生长在肺部还是皮肤表面，平台机制在不同肿瘤间具有通用性。 2. **罕见遗传病突破**：重点靶向肝脏罕见代谢疾病，关键临床三期数据预计在年底前揭盲公布。 3. **自身免疫性疾病攻坚**：借助在传染病疫苗和癌症治疗中对人体免疫网络建立的系统认知，向复杂的自免领域进军。

---

[06:40](https://www.bilibili.com/video/BV1XpHw6DEz5?t=400s) 重构生物制药估值模型与极端交付速度

针对二级市场投资者的困惑，主持人提到如果把 Moderna 套用传统药企 10% 的临床成功率模型，其估值逻辑就会被严重误读。Bancel 顺势剖析了决定生物医药资产价值的核心公式： $$\text{资产价值} \approx \text{销售峰值} \times \text{产品上市周期倒数} \times \text{临床成功率（Probability of Success）}$$ 传统药企每做一款新药都要重建供应链并独自承担临床失败风险，而平台型体系不仅大幅拉升了研发成功率，更显著缩短了开发周期，改变了现金流贴现模型。

以新冠疫苗为例，Moderna 从获得病毒基因序列到疫苗最终获批仅耗时 11 个月，跑完了行业常规需要耗费至少 4 年的临床一、二、三期全程。Bancel 客观地回顾道，这一神速一方面得益于监管体系的支持（FDA 评审员提供全天候不间断快速周转）以及政府对多家企业的资金托底，使企业能够不计商业财务风险地并行推进各项工作；另一方面，更关键的内因在于其平台技术无需重新发明工艺。Bancel 回忆起自己在礼来制药（Eli Lilly）的工作经历，指出传统药企通常在三期临床一启动就必须冒着数亿美元的报废风险开工建厂，否则试验成功后根本无法按期供货；而 Moderna 依靠底层统一的 mRNA 生产工艺，彻底摆脱了这种被动局面。

- **估值核心三要素**：销售峰值、上市开发速度、临床成功率三者的乘积决定了药物资产的真实价值。 - **11 个月新冠获批奇迹**：并行推进临床试验、监管机构 24/7 评审支持与底层工艺免重新设计共同促成了这一历史速度。 - **规避传统厂房投资风险**：通用生产模块消除传统药企“三期未明、厂房先行”的巨大资本沉没隐患与产能错配风险。

---

[08:30](https://www.bilibili.com/video/BV1XpHw6DEz5?t=510s) 从零起步到十亿剂产能：定制化药物制造的飞跃

面对来自团队内部分析师 Rachel Warren 关于实体供应链扩张与个性化疫苗量产能力的提问，Bancel 讲述了团队在制造领域的工程突破。大规模制造适用于所有人的通用型疫苗，与为每个癌症患者“量身定制”完全独一无二的个性化 mRNA 药物，在生产组织与质控逻辑上有着天壤之别。为此，Moderna 投资兴建了专门用于生产个性化肿瘤疗法的专业化定制工厂。

值得骄傲的是，操盘这套高精尖个性化产线建设的，正是当年在新冠疫情中创造奇迹的原班生产团队。Bancel 感慨地回忆道，2020 年初疫情刚刚暴发时，Moderna 本质上没有任何大规模商业化生产能力，甚至在此之前连一次三期临床试验都没有主导过。然而，正是这支队伍以超常规的工程执行力，在短短一年之内实现了从零到 10 亿剂产出的极限突破。如今，正是这支战功赫赫的团队，正在将这种敏捷、模块化的高强度制造经验，复刻到更为复杂的全球个性化癌症治疗供应网络之中。

[10:00](https://www.bilibili.com/video/BV1XpHw6DEz5?t=600s) Miniaturized mRNA Manufacturing and the Necessity of Individualized Cancer Care

Stéphane Bancel explains that scaling individualized cancer vaccines required Moderna to completely rethink biomanufacturing by shrinking reactor sizes down to a miniature scale. Because mRNA production takes place in water using enzymes—mirroring how human cells act as tiny biological reactors taking DNA instructions to synthesize proteins like insulin in the pancreas—the company faces no fundamental scientific limits on compaction. This gives Moderna a major advantage over living cell therapies such as CAR-T, where fragile human cells must be delicately handled, cultured, and protected from damage during processing.

```mermaid flowchart LR accTitle: mRNA vs CAR-T Manufacturing accDescr: Comparison of manufacturing processes between cell therapy and enzymatic mRNA synthesis. subgraph CART["CAR-T Cell Therapy"] C1["Human cells"] --> C2["Delicate handling & culture"] C2 --> C3["High fragility & scale risk"] end subgraph MRNA["Moderna mRNA Platform"] M1["Aqueous enzymatic reaction"] --> M2["Miniature bioreactors"] M2 --> M3["Scalable individualized output"] end ```

When asked whether broad clinical data might eventually uncover enough common cancer characteristics to allow for standard off-the-shelf treatments rather than extreme personalization, Bancel acknowledges that the team originally hoped that would be the case. However, the phase 2 clinical data proved just how uniquely mutated human tumors really are across patients:

> “Out of half people, 99% of the genes were different. 90% mutations were different.”

Because personalization has proven essential and effective, Moderna is actively leveraging that individualized infrastructure for other indications, including Lynch syndrome. Lynch syndrome is a hereditary genetic condition affecting roughly 1 in 300 people—amounting to approximately one million Americans—which can be easily detected via sequencing but currently lacks therapeutic options. Moderna is designing a targeted vaccine program intended to protect this entire patient population.

- **Reactor compaction:** Production occurs purely in aqueous enzyme solutions without the fragility bottlenecks of living cellular therapies. - **Genetic divergence:** Phase 2 data revealed that 99% of genes and 90% of tumor mutations diverged across patient cohorts, validating personalized neoantigen designs. - **Lynch syndrome target:** Affects roughly 1 in 300 individuals (~1 million people in the United States) who currently have no preventative or therapeutic treatment.

---

[11:47](https://www.bilibili.com/video/BV1XpHw6DEz5?t=707s) Moderna vs. BioNTech: Divergent Strategies, Chemistry, and Delivery

When the interviewer playfully offers Bancel an out on comparing Moderna to its German peer BioNTech, Bancel embraces the question, emphasizing that the two companies share a foundation in mRNA but diverge fundamentally in corporate philosophy and biochemical execution. BioNTech was conceived from day one as an oncology-first enterprise pursuing cancer targets across multiple drug modalities, including small molecules, biologics, and CAR-T cell therapies. Moderna, by contrast, set out from its founding to build a horizontal mRNA platform spanning diverse therapeutic areas. Even where both companies pursue mRNA cancer vaccines, Bancel stresses that their products cannot be equated or correlated because they differ across at least five critical dimensions:

> “So it is impossible in a positive or negative way to correlate those two products. You're going to have to wait for the data.”

Bancel compares equating the two therapies to claiming two automobiles are identical simply because they are both cars: their internal engineering, delivery biology, and clinical profiles are fundamentally distinct.

---

[13:14](https://www.bilibili.com/video/BV1XpHw6DEz5?t=794s) The "Lasagna" Patent Architecture and the Uncopyable Algorithm

Addressing how Moderna protects its intellectual property compared to traditional biotech companies that patent a single standalone molecule, Bancel uses the analogy of a multi-layered lasagna. Rather than patenting one isolated compound, Moderna continuously stacks intellectual property across every tier of its platform:

> “Because we're a platform, think about it almost like a lasagna. So we build layers and layers and layers of patent on mRNA. And then when you do a drug, you just take a slice of that lasagna.”

When Moderna brings a new candidate to the clinic—whether for COVID-19 or individualized oncology—it slices through stacked layers of mRNA sequence modifications, proprietary lipid delivery chemistries, and specialized manufacturing methods.

Beyond patent filings, Bancel argues that individualized cancer therapies carry an insurmountable regulatory moat rooted in trade secrets. A generic or biosimilar competitor cannot simply copy an individualized therapy and petition the FDA for an abbreviated pathway:

> “How can you go to the FDA and say it's the same drug and waive the clinical trial cost? I think this will not happen.”

Because the computational algorithm that evaluates a patient's tumor sequencing data to select the optimal mutations is a closely held internal secret—known only to a tiny handful of people inside Moderna—no third party can duplicate the drug product or prove bioequivalence without funding their own comprehensive clinical trials from scratch.

---

[14:44](https://www.bilibili.com/video/BV1XpHw6DEz5?t=884s) Accelerating Complex Biology and Smashing Clinical Timelines with AI

Looking forward to the integration of advanced artificial intelligence and AGI agents, Bancel highlights two core breakthroughs: mastering complex multigenic biology and compressing operational timelines. Because human disease states rarely hinge on a single biological target, Moderna is deploying AI models to design therapies combining multiple mRNA sequences into a single vial—mimicking the natural cellular environment where thousands of distinct mRNAs operate in harmony.

On the operational side, Bancel points out that clinical phases encompass hundreds of individual administrative workflows. Historically, consolidating and cleaning Phase 3 trial data for an FDA submission took the pharmaceutical industry an average of six months. Moderna has compressed that milestone to roughly three months, but Bancel envisions a far more radical AI-driven transformation:

> “We are starting more and more to get to three months at Moderna, and I want that to become one day.”

Reaching a single-day turnaround between final Phase 3 data and regulatory submission will require a multi-year effort, but shaving off months of data processing translates directly to earlier commercialization, expanded patent runway, and a dramatic boost to the net present value (NPV) of pipeline assets.

- **Multivalent formulations:** Using algorithmic design to combine up to four distinct mRNAs within a single vial to modulate complex biological pathways. - **Data assembly compression:** Moving from the pharmaceutical industry standard of 6 months down to Moderna's current 3 months, with an ultimate target of 1 day. - **NPV impact:** Accelerating market entry by a full quarter yields immediate revenue advantages and extends effective commercial exclusivity.

---

[17:12](https://www.bilibili.com/video/BV1XpHw6DEz5?t=1032s) Weathering the 95% Stock Crash and Playing the Platform Long Game

Reflecting on the turbulent period between 2021 and 2025, where Moderna's stock price collapsed from an all-time peak of $497 down to $23—a 95% drawdown—Bancel acknowledges that the collapse was difficult, though not unexpected. Management understood that the extraordinary revenue volumes driven by emergency COVID-19 vaccination would naturally taper off, prompting Wall Street to aggressively reprice the stock. However, COVID-19 served as proof of concept, demonstrating that the mRNA platform operated safely and effectively at scale. Rather than panicking, leadership kept investing in scientific expansion:

> “COVID got us out of jail because we knew the technology was working safely. So then he was just sticking to the science, taking a long view, investing, because when the thing was going to go back, you are going to go back strongly as it's starting to happen.”

Bancel illustrates his personal conviction through his capital allocation: when his 10-year executive stock options were set to expire, he exercised them fully in cash at roughly $29 per share, paying his taxes out of pocket to retain 100% of the underlying stock. He notes that platform enterprises naturally concentrate the vast majority of their enterprise value in the later stages of maturity.

When the interviewer mentions an article he published floating a potential acquisition by Merck at $215 per share—right as Moderna shares rebounded to $213—Bancel brushes aside buyout speculation. Supported by a board of directors that reviews and updates detailed five-year plans annually, Bancel believes the market is only beginning to understand the platform's breadth. Specifically, he points to autoimmune diseases as a massive frontier that is completely absent from Wall Street analyst models, yet represents a colossal commercial opportunity where Moderna's mRNA capabilities could outperform conventional therapeutic approaches.

[20:00](https://www.bilibili.com/video/BV1XpHw6DEz5?t=1200s) Building the AI Flywheel in R&D and the Threat of AI-Enabled Bioweapons

Stéphane Bancel explains that Moderna’s technological toolkit is continuously expanding, extending deep into antibody development and discovery. Reflecting on their Science Day held that past June, Bancel highlights how Moderna is building an autonomous "flywheel" supercharged by artificial intelligence. In this closed-loop engine, AI formulates scientific hypotheses, generates novel mRNA sequences, designs new lipid nanoparticles (LNPs), evaluates their efficacy in animal testing, feeds the resulting data right back into the computational system, and immediately loops back to run subsequent iterations at unprecedented speed.

```mermaid flowchart LR accTitle: Moderna AI Scientific Flywheel accDescr: Closed loop showing hypothesis generation, mRNA and lipid design, animal testing, and data feedback. subgraph Computation["In Silico Discovery"] A["AI Scientific Hypothesis"] --> B["Design Novel mRNA & Lipids"] end subgraph InVivo["Wet Lab Testing"] B --> C["Animal Model Testing"] end subgraph FeedbackLoop["System Optimization"] C --> D["Data Feedback to Engine"] D --> A end ```

The interview takes a serious, darker turn as the host turns to an issue Bancel has previously highlighted as the single gravest societal risk posed by artificial intelligence: the creation of a cheap, gain-of-function respiratory bioweapon engineered by a malicious actor. When asked about the true scale of that threat and what can be done to mitigate it, Bancel is unsparing in his assessment: “So as many of you have read gain research aided by AI is much cheaper to do than build a nuclear weapon Much much cheaper.” Synthesizing a high-consequence pathogen no longer requires the astronomical budgets, specialized materials, or state-level infrastructure of an atomic program, which makes biological hazards far more asymmetric and readily weaponized.

- **The Closed-Loop AI Flywheel**: Integrates in silico hypothesis formation, mRNA encoding, specialized lipid formulation, in vivo animal data collection, and algorithmic model re-training. - **Biosecurity Cost Disparity**: AI-assisted gain-of-function research on respiratory viruses demands orders of magnitude less capital and equipment than uranium enrichment or nuclear weapon assembly.

---

[21:05](https://www.bilibili.com/video/BV1XpHw6DEz5?t=1265s) Rapid Pandemic Response: Global Manufacturing and 10-Minute Vaccine Design

Moderna is proactively constructing defense-in-depth infrastructure across manufacturing and clinical development so the world can respond swiftly to an intentional or natural biological catastrophe. Bancel notes that only an airborne respiratory virus can produce a catastrophic global pandemic. In contrast, viruses with other transmission dynamics spread far more slowly; an outbreak like Ebola is horrific, but due to its biological nature, it cannot spread into a true airborne pandemic. Moderna already has an Ebola vaccine candidate advancing through clinical trials, which they believe will be highly effective and expect to test in Africa very soon. Concurrently, Moderna has advanced candidates targeting around 10 distinct respiratory viruses into the clinic to build pre-existing regulatory and technological templates.

Because Moderna now holds FDA approval for a seasonal flu product, the timeline for countering a weaponized flu variant would be radically compressed compared to the early months of COVID-19. Once an engineered viral sequence is identified, Moderna can design a custom mRNA vaccine candidate on a computer in roughly 10 minutes. Under that approved regulatory umbrella, bringing a modified formulation to authorization would require only a single clinical bridging study rather than an exhaustive multi-phase trial. Applying that pipeline to an outbreak emerging early in the year, a finished vaccine could secure FDA clearance by May or June. Furthermore, Moderna's global production capacity has evolved immensely since 2020:

```mermaid flowchart TD accTitle: Rapid Response Workflow for Weaponized Pathogen accDescr: Process showing sequence identification down to billions of doses within months. Seq["Identify Sequence (T=0)"] -->|"10 Minutes In Silico"| Design["Design mRNA Construct"] Design --> Synthesize["Formulate mRNA in Lipids"] Synthesize --> SingleStudy["Single Bridging Study (Prior Approval Platform)"] SingleStudy --> Approval["FDA Approval (Month 5-6 / May-June)"] Approval --> ScaleUp["1-2 Billion Doses Deployed Globally"] ```

- **Global Manufacturing Footprint**: Moderna has scaled well beyond the United States, operating operational and developing plants in Canada, the United Kingdom, and Australia, alongside additional public-private national partnerships. - **10-Minute Sequence-to-Design**: Using digital sequence files directly, vaccine candidates can be computationally constructed in minutes without culturing physical pathogen samples. - **Accelerated Approval Timelines**: Leveraging a baseline approved respiratory vaccine framework allows single-study bridging, slashing regulatory approval times down to roughly 5 to 6 months. - **Mass Scale Output**: Unlike December 2020, when ramping bottlenecks meant only 20 million COVID-19 doses were ready at FDA approval, Moderna's expanded network is positioned to deliver 1 to 2 billion doses by mid-year during an emergent crisis.

---

[23:01](https://www.bilibili.com/video/BV1XpHw6DEz5?t=1381s) CEO Philosophy on Stock Price, 10-Year Cash-on-Cash Return, and Value Creation

Closing the interview with a lighthearted inquiry, the host asks Bancel whether he checks the Moderna stock price on a daily or weekly basis, or if he shields his eyes from market volatility altogether. Bancel admits candidly that he monitors the share price because investors consistently bring it up and it would look bizarre for a public company CEO not to know where the stock trades. However, he emphasizes with total conviction that market quotes never dictate his leadership or allocation of capital.

Bancel outlines how the leadership team actually evaluates their portfolio: “I always think about 10 years cash on cash return That's what I've done since day one of the company 15 years ago.” Pointing directly to his calendar for that very afternoon, Bancel notes that he and his executive team are spending a two-hour lunch minutely dissecting their entire R&D pipeline—debating which programs generate enduring long-term value and how to rigorously insulate the business from downside clinical risk.

He bluntly rejects the culture of meeting Wall Street expectations quarter by quarter: “I'm not looking at the EPS of next quarter. I never have. and most probably never will because what happens next quarter is things we've decided and worked on two or three years ago.” Rather than selling, Bancel reiterates that he has actively bought shares into the downturn because the company's valuation had become deeply disconnected from fundamental reality. Bancel remains anchored in the conviction that an industrialized mRNA platform will ultimately transform human health, and that doing so will systematically yield massive rewards for patient shareholders.

- **15-Year Capital Allocation Anchor**: Strategic decision-making is evaluated strictly against a 10-year cash-on-cash investment horizon rather than quarterly performance benchmarks. - **Active Insider Buying**: Viewing the multi-year market sell-off as an extreme mispricing of the underlying technology platform, Bancel has been accumulating shares rather than distributing them. - **Platform Compounding**: Executing systematically on high-probability R&D assets positions the underlying platform to simultaneously revolutionize patient care and unlock sustained shareholder value.

## Notable quotes > [03:00](https://www.bilibili.com/video/BV1XpHw6DEz5?t=180s) We knew from the beginning that it was going to be zero drug or a lot of drugs. — Stéphane Bancel

> [23:01](https://www.bilibili.com/video/BV1XpHw6DEz5?t=1381s) I'm not looking at the EPS of next quarter. I never have. and most probably never will because what happens next quarter is things we've decided and worked on two or three years ago. — Stéphane Bancel

> [13:14](https://www.bilibili.com/video/BV1XpHw6DEz5?t=794s) Because we're a platform, think about it almost like a lasagna. So we build layers and layers and layers of patent on mRNA. And then when you do a drug, you just take a slice of that lasagna. — Stéphane Bancel

> [23:01](https://www.bilibili.com/video/BV1XpHw6DEz5?t=1381s) I always think about 10 years cash on cash return That's what I've done since day one of the company 15 years ago. — Stéphane Bancel

> [05:00](https://www.bilibili.com/video/BV1XpHw6DEz5?t=300s) It's again because we have a platform that reduces the risk and increases the probability of drugs to go the whole way to approval. — Stéphane Bancel

## Mentioned Moderna, Stéphane Bancel, Keytruda, Merck, BioNTech, FDA, Rachel Warren, Eli Lilly, The Motley Fool, FoolFest, CAR-T, 黑色素瘤, 林奇综合征, 埃博拉病毒, 巨细胞病毒

## Video description 【The Motley Fool】 Moderna CEO：mRNA将颠覆癌症治疗，股价曾暴跌95%我却坚定加仓！ Moderna CEO Reveals His Massive Plans for Cancer, AI & Beyond!!

许多人对Moderna的印象还停留在新冠疫苗。这期FoolFest访谈里，CEO Bancel给出了更大的蓝图：mRNA本质是信息分子，只用四种碱基就能编码人体内任何蛋白质。公司花了十年才掌握安全、高效量产mRNA的方法。目前人类对自身运行机制的理解只有5%到15%，未来5到10年医学可能迎来爆发式进展。

癌症是Moderna的主战场。 mRNA药物与Keytruda协同作战：后者为T细胞供能，前者教会免疫系统识别癌细胞的新突变，产生全新T细胞攻击肿瘤。黑色素瘤二期数据显示，五年后80%的患者病情受控，在皮肤癌领域已接近治愈。目前还有9项覆盖肺癌、肾癌、膀胱癌的临床研究在推进。

平台型公司的底气。传统生物制药每款新药成功率只有10%左右，而mRNA只需改变碱基排列就能编码不同蛋白。Moderna不是一家疫苗公司，而是一个平台：一旦技术在某个领域成功，后续药物开发将大大加速。已有五款疫苗获批，平台成功率高达85%；新冠疫苗从拿到病毒序列到获批只用了11个月，行业通常需要四年。疫情初期团队就把年产能做到了10亿剂，mRNA生产在水中进行，没有科学上的限制，可以高度小型化和规模化，公司还专门建厂生产个性化治疗药物。

下一个增长曲线。管线重点是罕见病，一项三期临床预计年底揭盲，再往后是自身免疫疾病，这部分市场潜力巨大，当前股价尚未体现其价值。个性化癌症疫苗二期覆盖150名患者，其中99%的基因和90%的突变各不相同，个性化是刚需；针对林奇综合征的疫苗将覆盖美国约100万目前无药可医的患者。

和BioNTech不是一条路。对方专注癌症、多条技术路线并行；Moderna做跨疾病平台，在mRNA修饰、脂质递送、注射方式和突变选择算法上都有差异，个性化癌症疫苗的突变选择算法更是高度保密，加上"千层饼"式多层专利，构成几乎无法复制的知识产权壁垒。

AI是下一个引擎。它能帮Moderna设计含多个mRNA的复杂药物，还能把三期临床数据整理并提交FDA的时间从行业平均6个月压缩到3个月，长期目标是一天，直接提升药物上市速度和产品净现值。AI还在帮公司构建"提出假设—开发新mRNA和脂质—动物实验反馈"的飞轮，不断加速创新。

暴跌95%后，CEO自己在买进。 2021到2025年Moderna股价从497美元跌到23美元，最大回撤95%。Bancel坦言自己曾在29美元附近增持，认为被严重低估；平台型公司的大部分价值会在后期体现，董事会每年更新五年计划，专注长期价值创造。他每天看股价，但决策只看未来十年的现金回报。他始终相信mRNA平台会给人类健康带来巨大影响，最终也会为投资者创造巨大价值。

AI的另一面：生物安全。最大的风险是被滥用，制造廉价的"功能增益"呼吸道病毒武器。Moderna已针对约10种呼吸道病毒开发疫苗，一款流感产品已获批，埃博拉疫苗在推进，并在全球建了更多工厂以备紧急扩产。一旦出现新病原，从拿到病毒序列到设计新疫苗只要10分钟，获批时间可缩到5-6个月，产能可达10-20亿剂。

三条操作指引： 1. 跟踪Moderna癌症、罕见病管线的临床进展，尤其是年底揭盲的三期数据，这是长期价值的核心变量。 2. 用平台型公司85%成功率的逻辑给它估值，别套用传统药企10%的框架。 3. 关注AI在药物研发中的落地速度，提交FDA周期从6个月到3个月再到1天，可能显著改写竞争格局。

时间轴： [0:25] mRNA：可编码蛋白质的信息分子 [1:47] 癌症疫苗与Keytruda协同 [2:30] 黑色素瘤二期：五年80%受控 [5:02] 平台研发成功率85% [6:16] 下一步：罕见病与自身免疫 [14:44] AI压缩临床试验周期 [17:12] 股价跌95%与长期主义

[Source](https://www.bilibili.com/video/BV1XpHw6DEz5)

中文简报

Moderna首席执行官：mRNA将颠覆癌症治疗，股价暴跌95%后坚定增持访谈中，Moderna首席执行官Stéphane Bancel阐述了mRNA作为“人体软件代码”的平台型药企本质，指出其通过调整四个碱基编码不同蛋白的特性将临床研发成功率提升至约85%。在肿瘤领域，Moderna的个性化癌症疫苗与Keytruda协同攻坚黑色素瘤，二期临床实现了术后5年80%的无复发生存率，且正将该技术横向扩展至肺癌等其他癌种与罕见遗传病。Bancel详细对比了Moderna与BioNTech在化学修饰、递送系统及算法壁垒上的本质差异，并展示了AI如何将三期临床数据提交周期压缩至数月甚至未来的单日水准。面对2021至2025年间股价自高点暴跌95%的市场动荡，Bancel表示自己选择在低位坚决行权增持，强调其管理决策完全聚焦于未来10年的现金回报与生命科学底层价值的长期复合增长。 Moderna在黑色素瘤二期临床中与Keytruda协同治疗，随访5年实现80%患者无复发生存，在皮肤癌领域已达临床治愈标准。 (1:47) 作为跨疾病mRNA平台公司，Moderna的7项三期试验累计成功率达约85%，远高于传统药企约10%的研发成功率。 (5:00) 二期数据显示患者间99%基因与90%肿瘤突变均不相同，证实癌症个性化疫苗为刚需，公司正通过算法与小型反应器推进量产。 (10:00) Moderna利用AI将三期临床数据整理及提交FDA周期从行业平均6个月压缩至3个月，终极目标是缩短至1天以极大提升资产净现值。 (14:44) 面对股价自497美元暴跌95%至23美元，首席执行官Stéphane Bancel在29美元全额自费行权增持，强调决策只关注未来10年现金回报。 (17:12)

Reading summary

Moderna首席执行官Stéphane Bancel认为，基于mRNA底层信息属性构建的工业化研发与制造平台正在根本性改变生物制药模式，其高成功率与拓展速度不仅将重塑癌症等复杂疾病治疗，也将长期兑现商业价值。 mRNA平台范式与高成功率优势 (3:00) Bancel指出，过去150年的化学药与重组蛋白开发多为彼此孤立的单发项目，每次从零试错导致行业临床整体成功率仅约10%。而mRNA属于遵循通用编码规则的信息分子，只需调整4个化学碱基的排列顺序即可指导细胞生成对应蛋白质。Moderna从创立初期便按工程化平台架构布局IT系统与自动化产线，一旦攻克特定器官递送技术，研发便能以极低边际成本快速复用。截至目前，Moderna完成的7项三期临床中包含了5款获批疫苗与1项癌症三期阳性结果，平台临床成功率达约85%。公司正将该能力拓展至肺癌、肾癌等实体瘤、罕见遗传代谢病及自身免疫疾病领域。个性化癌症疫苗协同机制与制造护城河 (1:47) 在肿瘤免疫领域，Moderna的个性化癌症疫苗通过测序获取肿瘤特异突变特征并编码新抗原，接种2至3针后刺激生成全新的特异性T细胞；与为衰竭T细胞重新充能的Keytruda联合使用，在二期黑色素瘤随访中实现了5年80%的无复发生存率。针对通用型癌症疗法的设想，二期临床数据显示不同患者间99%的基因与90%的突变各不相同，证实了个性化定制的必要性。制造上，Moderna通过微缩酶反应器规避了类似CAR-T等活细胞疗法脆弱且难以放大的瓶颈。Bancel强调，Moderna采用“千层饼”式的专利组合覆盖序列、递送与制造，叠加针对患者测序筛选突变的内部核心算法，构成了仿制药企难以凭借生物等效性跨越的监管与商业护城河。 AI全流程提速与生化防御应急架构 (14:44) Moderna正在将人工智能融入研发飞轮，实现从AI提出假设、设计mRNA与脂质纳米颗粒（LNP）、动物实验测试到数据自动回流训练的闭环，并尝试在单支药瓶中组合最多4种mRNA以调控多基因复杂疾病。在运营端，公司已将三期数据整理提交FDA的周期从行业通行的6个月缩短至3个月，终极目标是压缩至1天，以显著提升研发资产的净现值。面对Bancel认为成本远低于核武器的AI辅助呼吸道生物武器风险，Moderna凭借已获批的季节性流感疫苗监管通道建立起防线：针对新测序病原体可在10分钟内完成电脑设计，经单项桥接试验在5至6个月内获批，并依托跨国工厂网络在年中交付10亿至20亿剂疫苗。应对95%股价回撤与长期资本逻辑 (17:12) 针对2021至2025年间公司股价自497美元峰值暴跌95%至23美元的过程，Bancel表示管理层早有预期，因疫情紧急疫苗采购消退必然引发华尔街重新定价，但新冠实际验证了mRNA技术的安全性与规模化能力。Bancel拒绝迎合华尔街季度每股收益预期，坚持以15年来一贯的10年现金回报周期进行资产组合配置，并在自身期权到期时以约29美元全额现金行权并自费缴税保留全部股票。他否认了被默沙东收购的市场传闻，认为分析师模型完全忽视了自身免疫性疾病等潜在管线，并坚信工业化mRNA平台的长期复利效应终将展现。 Moderna的商业逻辑依赖于将其极高的三期临床成功率与工程化流水线持续变现，但其实体瘤管线的大规模拓展、自身免疫性疾病的转化效果以及应对突发疫情的极端响应速度，仍有待后续更多未盲临床数据与现实监管周期的全面检验。

Source quotes

[3:00] Stéphane Bancel: We knew from the beginning that it was going to be zero drug or a lot of drugs. [23:01] Stéphane Bancel: I'm not looking at the EPS of next quarter. I never have. and most probably never will because what happens next quarter is things we've decided and worked on two or three years ago. [13:14] Stéphane Bancel: Because we're a platform, think about it almost like a lasagna. So we build layers and layers and layers of patent on mRNA. And then when you do a drug, you just take a slice of that lasagna. [23:01] Stéphane Bancel: I always think about 10 years cash on cash return That's what I've done since day one of the company 15 years ago. [5:00] Stéphane Bancel: It's again because we have a platform that reduces the risk and increases the probability of drugs to go the whole way to approval.

Original transcript

[0:00] Hey, how are you doing? Quite well, quite well, Stefan. I'm so excited about this time that we have together. Because we only have three hours, I'm just going to get right into the questions. So let's start with explaining the foundational technology of Moderna so that a 10th grade student in high school who hates Moderna. science can understand it. Oh, come on, you cannot hate science in school. So very simply, as many of you have learned during the pandemic, mRNA is an information molecule. It's made of four letters. Like all softwares are made of zero and one mRNA is made of four letters that can code any protein in your body. And it's not something we invented. It's how life works. It's in plants. It's in animals. It's in humans. You have thousands of animals in every one of yourselves. And what we at Moderna... figure out and because you know, 10 years to do so before a pandemic. is how to make a safe mRNA molecule in a reactor, That will be the instruction. of any portier remote your body to make on demand. And for that student, one more question. At this point, how much, Stefan, do we understand about how the human body works. Oh, maybe 5%, 10%, 15%. I don't know because we don't know what we don't know, right? But it's tiny. Which is what I'm so excited about. what medicine is going to bring to people. in the next five, ten years is going to be amazing. So, Ines Merrin tells the body, "This is what your cancer looks like. Go find anything carrying these markings." It's like a wanted sign. And Keytruda then gives the immune system the freedom to attack the criminal when it finds it. Is that an accurate expression of how Ines Maron and Keytruda work together? So the way I will describe it is, the cathedral basically gives energy back to the T cells, this is a part of your immune system, that attack your cancer when you have cancer over time, when the tumor grows, your system gets... tired, and ketidra basically is boost of energy to the T-cell that already exists. What in Tismaran the modern mRNA drug does is we basically teach your immune system how to look for new signature, new mutation in your cancer. So it's basically new T-cells. that are brand new that you make after we give you a product it takes two to three dollars of a vaccine. to get your image system going, And then, as you know, the results are quite impressive. From the Phase 2 data, at Fabia's heart, 80% of people who have disease 3, which in skin cancer, is considered a cure. arrival to Moderna, I think in 2011. And I want to hear your explanation because I think this is the most misunderstood thing about Moderna. is your description of how the company has created a platform and how you would distinguish that from the single asset biotech companies. So Um, What is it that makes Moderna's solutions a platform. Yes, it's because of mRNA being an information molecule. If you think about the last 150 years where the world has had medicine, We started with small molecules, chemicals like aspirin and then later on, you know, antibiotics and Lupito and other products. But those are onesies. Every time it's a different chemical structure, you have to go reinvent everything. That's the efficacy, test of safety. Most things fail because you are relearning everything from literally scratch. When the biotech industry happened in the 70s and 80s, with Recombine and Technologie. Insulin growth hormone. We'll talk about KT Dry. KT Dry is a biotech product from its chemical standpoint, but again, Those are products that you have to relearn how to make every time. Because mRNA is an information molecule. because I'm only use the same four letters, and you just put them in different order. to code for a different protein, We knew from the beginning that it was going to be zero drug or a lot of drugs. He made no science piece. scientific sense that it would be one drug or two drug company. And so we built from the beginning the company. And I'm an engineer by turning... It would be the company to be a platform company, meaning how do you scale IT? How do you scale robotics? So that when you have something working once, you can get it again, which is why now we have five vaccines approved already. and the cancer vaccine is a new application. By the end of the year, we should have data in rare genetic disease of the liver, yet another application. And each of those applications, getting the first product done is really hard because you have to invent how to make the technology work in that organ. But then the second, the third, the fourth is like a copy and paste. So... It's fair to say that With each approval, Each new medicine makes the next one in that vertical Easier to create with a higher approval rate you would expect or... [5:00] That's the dream of the system you've built. Correct. And if you look at the data, so far we have five products approved. We have the cancer data, never positive phase three. And then we had CMV, which was a negative phase three. So that's seven phase three done. If you look at the industry average starting from phase one to phase three positive, it's around 10%. And if you look at those numbers in Moderna, we're around 85%. And so it's not because we're smarter. I don't think we are. It's again because we have a platform that reduces the risk and increases the priority of drugs to go the whole way to approval. And how likely do you think it is that Moderna will translate this success into other therapeutical verticals and which ones are at the top of the list for you. Sure. So in cancer, first, we're going to go very broad. We already have nine clinical studies ongoing in lung cancer, kidney cancer, bladder cancer. and more to come. Basically, in cancer we want to go and try for different tumor types, using the same technology because T cells do not care. if your tumor is in your lung or your skin, It's a tumor that T cell just goes around your body and your blood all the time. It's like little police checking for security incidents. The next big one is rare disease. And this we should know before the end of the year. We have a phase three study ongoing right now. And then the next thing is going to be autoimmune disease. If you think about Moderna between what we did in infectious disease vaccine, And then now in cancer, we're basically starting to really understand very well the immune system. So this is a seeking affirmation and clarification for all of us. So when an investor looks at the approval rates let's say, as you mentioned, a phase one drug, 10% approval rate, and tracks that to Moderna. and the solutions you've created thus far. One would have to conclude that Moderna stock is incredibly overvalued if it had the same approval rate as an average drug through the FDA process. But because you've built a system, I see a number of people misunderstanding the platform system that Moderna has constructed here. You're expecting to see much higher rates within each vertical as you've proven out the first drug in that category. Correct. If you think about the pharma business or the biotech business, three things really matter. What will you pick status of a drug? How long to launch? times the probability of success. Those three things define most of the value of an asset. And so by being able for a platform to increase the probability of success, but also accelerate development. The other piece that the platform gives us is you don't have every time to reinvent how to make a drug. It is the same thing as the overdrug. You just change again the order of the letters on the template that used to make the mRNA molecule from. So, the speed, we have not spoken about the speed. I mean, look, you all know that COVID, we got a vaccine in 11 months out. while doing the phase one, phase two, phase three, where usually is at least a four-year process was a record before. The 11 months was spectacular because the FDA helped us a lot. to have a reviewer else. just spending 24/7 on our files. and not having a long clock cycle, which happens in a regular way of doing business. The government helped us a lot by funding six companies, and those companies taking a lot of business risk by doing things in parallel to not take safety risk, but to go faster. So I was very unique and exceptional. But the piece that's still true is that the place at which we can go, because we don't have to invent the technology, we don't have to be in new factory. I used to work at Eli Lilly, an amazing company. At Lilly, like in most pharma companies, what happens is you have to start building your factory when you start your phase three. And the reason is if you don't build a factory while the phase 3 is ongoing in the clinic, If it's positive, you cannot sell product. But in our case, because we use the same technology for everything, we don't have that financial risk. And we don't have a risk of getting the forecast wrong. Rachel Warren in our Hidden Gems team wants to know about your ability to scale physical manufacturing. We did talk a little bit about that in the pre-interview call, but there are a number of questions about what... your ability to scale an individualized vaccine to meet the needs of so many patients around the world. So maybe some of the manufacturing facility investments you've made. Sure. So we build a dedicated factory to make individualized treatments, very different where you make the same product for everybody, of course. But it was done by the same team that did COVID. And the team that did COVID was remarkable. Because in 2020, when the pandemic started, we had no manufacturing capacity. We never have even run a phase three. And so the team was able to get a billion dollars output in 2020. And so that same team, we put them [10:00] in view as cancer vaccine. We had to shrink everything down. So you have very, very small reactor size. And what we have done is invest a lot in technology to be able to shrink everything down. But because mRNA is made in water, and lacking your cells. If you think about it, your cells are little reactors of a money It takes the information from your DNA. makes an mRNA molecule, let's say for insulin in your pancreas, and then make an insulin protein. And so that's all happening in water with enzymes. So we have no limited scientific limitation on how we can compact things. And so that's why we can scale out. the individualized treatment. It's very different than other technologies like CAR-T that use human cells And you have to manage human cells where you have to be very careful not to damage them. We don't have that problem. Is it possible that you will find numerous common characteristics in cancer across patients such that there are actually some standardization rather than the level of individualized treatment that you might be assuming now. So initially we thought so. But the data we've had so far in the phase two study, I don't have yet publicly available the data for the phase three study. Out of half people, 99% of the genes were different. 90% mutations were different. But what we're trying to do with this technology, because we know it works now, is how can we use that in other applications? So there's a disease... that some people might have heard of because somebody in your family or friends might have it, is called Lynch disease. It's one in 300 people that have lynched disease. So it's around one million Americans. Mm-hmm. You can diagnose it with simple sequencing test. There's no treatment available to there. But what we are basically doing is developing a vaccine that will cover all the people that have linching. This next question is unfair. And I give you I give you the freedom to simply say I'd like to pass on that one. Bring the next question. Compare and contrast Moderna with BioNTech. It's just very different companies. So the similarity is mRNA as a technology, but biotech has always been a cancer driven company. You can go back and read the S1. They are about cancer across technology, small molecule, large molecule, cell therapy, like CAR-T and mRNA. We have said from the beginning we're going to be a company across disease areas. So very different strategy. Of course, there's an overlap. in cancer mRNA. Both companies are doing it. But it's impossible to correlate those products because We don't modify the urodine. It's one of the four letters on it. We use lipid. They don't, they use a lipoplex, it's different chemical. We inject it intramuscular. We know it drains into your lymph node, your immune system. They inject it IV. I have no idea where it goes. You need to ask them. Our algorithm is different than the algorithm to pick from the thousands of mutations the 34 in our case and the 20 in that case. So as you can see, at least on five dimensions, The products are very different. So it's like if you had two cars and you say those are the same things because they are cars. No, they have very different features, right? Different cars. And so it is impossible in a positive or negative way to correlate those two products. You're going to have to wait for the data. To know if their product works, it's impossible to know on the onset because the technology is so different. Thank you for that. I want to have you shed some light on patents at Moderna because the average biotech company is patenting a drug in a single molecule, but you're patenting layers of a system. So can you talk about the difference between the two? Sure. Because we're a platform, think about it almost like a lasagna. So we build layers and layers and layers of patent on mRNA. And then when you do a drug, you just take a slice of that lasagna. Let's say for COVID. COVID for the mRNA, COVID for the lipid, COVID for the manufacturing process, you take a slice. And so we keep adding new layer to lasagna by filing new patents as we discover new technology. And then we just apply it across biology for the drug that we are making and taking to the clinic. And some drugs, by the way, like in Tismaran, the cancer drug we've been speaking about that is individualized, I do not know how somebody copies that drug. And the reason is simple. How can you go to FDA and say, I'm not going to do the clinical studies. It's the same drug in the fight. How can you prove it? Because the algorithm to pick the mutation is not known to the public. It's not fine. Even I don't know it. We're very few people in the company who know what's in the algorithm. And so how can you go to the FDA and say it's the same drug and waive the clinical trial cost? I think this will not happen. I'd like to hear a little bit about your thoughts of – the application of artificial intelligence and AGI at Moderna going forward rather than what you've accomplished so far in your use of the tools right now. the most advanced frontier models in the hands of sophisticated [15:00] AI coders are delivering agents that have the equivalent of a human IQ of 165 when the average IQ for a human is 100, so 65% higher. So what will be the two or three major breakthroughs you think at Moderna, for applications of AI. going forward. As you started our discussion, biology is very complicated. Most diseases don't need one thing, they need two or three or four things. We have drugs now with four mRNA in the vial. You can use AI to go after very complex biology, and decide what are all the things you need to code in the product and you can combine mRNA in the values like you have thousands of mRNA inside each of your cells. So it can help with peak cells through innovative medicine. It can help with time by compressing timeline, especially clinical trials. Because if you think about phase one, phase two, phase three, It is hundreds of business processes. And as you start to use AI and agent to compact those business processes, you're going to reduce time to go, file a product to be FDA. It takes around six months on average between the phase three positive study... and the filing of the file to the FDA. Six months of data gathering and cleanup, right? We are starting more and more to get to three months at Moderna, and I want that to become one day. Because again, you're just cleaning of data, putting everything in a file that the FDA wants to receive it when you send the file electronically. But it's just data processing. So are we going to get there overnight? No. It might be a two to three year journey to get there. But getting from three months, which are already best in class compared to six months industry average to one day, that's three months more of sales. Think about the impact it has on your MPV to launch your product three months earlier. And then you can go at all the big business processes from when you start a phase one, that will create a lot of value by just compressing time of clinical development. You said you believed that when you joined Moderna, that it would either be a bankrupt company or one with a flood of new medicines, that you were building a system going from zero to one as a company. So what was it like between 2021 and 2025, years later, years after you joined, so 10 years after you joined, to see your stock go from $497 a share down to 23? with a 95% decline in the stock, and how was the interaction with your board, etc. So it was clearly not fun, but we kind of knew this was going to happen because it was clear that as we are building for the volume of COVID, the volume was going to come down. And as the sales come down, the market reprice the stock. But because we always believed it was going to be a platform and remember what I said, zero drug or a lot. COVID was number one. COVID got us out of jail because we knew the technology was working safely. So then he was just sticking to the science, taking a long view, investing, because when the thing was going to go back, you are going to go back strongly as it's starting to happen. which is why last year I had, it's all public information, of course, I had options expiring. They were 10 years old. I always keep my options to the end. They were 10 years old. I was not going to exercise them. So at 29 bucks, I bought all of them cash. I paid my taxes. I bought 100% of the stock. In August, I don't know if... Grantvesting.com. I just use, I sold some shares just to pay the taxes, but I held everything else because I wasn't selling at 50. And same thing. This is a platform. And as we all know. platform company value, most of our values in the last couple of years, right? And so being a founding CEO in the company and playing the long game, uh, the market will reprise the company as people start to realize we have a platform that has not been built before. It's always hard when you have an end of one company, for people to understand how to price it. Yes. I know we talked on the on the pre-call about we bought in Hidden Gems at the Motley Fool at twenty four dollars a share. So about a dollar or two above the above the loan. We've only been adding. And I can tell you after our prior conversation, after this conversation, I anticipate holding and adding to my Moderna investment. But I did write an article recently when the stock was at one hundred and forty dollars. that perhaps it would be a good time for Merck to acquire Moderna for $215 a share, but your stock is up another 8% today to $213 a share. So I'm assuming you think that was a pretty misguided article that I wrote. Look, Merck will decide what they have to decide. We know what we have. The great thing about Moderna is... I have a board who understands our platform deeply. understand our science deeply. We do five-year plans at least once a year. We update them regularly as we have new data on the platform. the platform is going to keep on giving. Look, autoimmune disease, I don't think anybody has that price into the stock. Analyst report on the street and that's going to be a gigantic opportunity and we might be able to do autoimmune disease like nobody else can do using. [20:00] antibodies. And we're still expanding. We talked at our Science Day back in June that now we are building a flywheel with AI to generate a scientific hypothesis, make new mRNA, make new lipid, test them in animal, feedback the system and go. Sorry to end on a dark point here. But first of all, Stefan, thank you so much for this time. Thank you for the conversation we had earlier. Congratulations on everything that you all are achieving at Moderna. And it's one of my favorite companies in the public markets in the world. And I enjoy not just profiting from investing in it, but learning about the world and about your work. You've cited the single biggest risk you see from AI as the risk of a cheap gain of function respiratory virus being created by a bad agent. Can you talk a little bit about how substantial you think that risk is and what should be done to prevent or reduce the risk of that substantially? Sure So as many of you have read gain research aided by AI is much cheaper to do than build a nuclear weapon Much much cheaper And so what we are doing at Moderna to if something really bad were to happen, to be ready to help is a few things. First, on the manufacturing front, we have built more factories. Now we have factories not only in the U.S., but also in Canada, in the U.K., in Australia. We're trying to do a few more with some government long-term partnership across the world on the manufacturing. And then on the clinic, we're trying to get ahead. So to be a pandemic, you would have to have a virus that is a respiratory virus. As you know, other virus spread much slower. So you can have outbreaks like Ebola. Ebola, by the nature of the virus, cannot become a pandemic. We have in a clinic right now, an Ebola vaccine that we think will work great and hopefully you can test it in Africa very soon. And so for all the around 10 respiratory virus we are taking those into the clinic even for flu we already have a flu product approved So if there was a flu gain of function you know bioweapon we should be able to very quickly on the computer in 10 minutes from the sequence to make a new vaccine and to get very quickly going. And because we have an approved flu product by the FDA now, we will need only one study. So if you use the 2020 timelines from COVID, we could have a vaccine approved by May, June, with something starting early in the year. So very different. And it's only 20 million doses to ship, which is what happened at the end of 2020 when the vaccine was approved by the FDA, because we were ramping up manufacturing and we had made very little volume by then, we should be able to have maybe one or two billion doses of product available mid-year. So we look very, very different in terms of impact. So we are trying to do a lot of things like this to just be ready in case something bad happens, either from nature or from a bad actor. So, Stéphane, I mean, my last question is only lightly humorous. Do you follow the price of your stock on a daily, weekly basis or do you shield your eyes from it entirely? So I look at it because investors talk to me and they know the stock price and I will look a bit weird as CEO not to know the stock price But it not what drives my decisions I always think about 10 years cash on cash return That's what I've done since day one of the company 15 years ago. And that's what my team and I do. Today at lunch, we're going to spend two hours looking at the portfolio of all the assets in R&D and debating intensely how do we create value, how do we manage downside risk. And that's what we do. I'm not looking at the EPS of next quarter. I never have. and most probably never will because what happens next quarter is things we've decided and worked on two or three years ago. And that's how this business works. So I keep my eyes on the long term. Again, I'm a long-term shareholder company. I'm not selling stock. I'm just buying stock actually on the other way because it was so mispriced. And I want to create a huge impact on humanity when making medicine. And as a consequence, if we do this well, this platform will create a lot of drugs and will create a lot of value for investors. Stefan Bansel, the CEO of Moderna, thank you so much for this time and all these insights, and we wish you the very best. Thanks, Tom. Bye.