【Meet Kevin】AI正在重塑医疗板块：这3只潜力股不容错过！

浑水摸鱼清源

AI-assisted summary of 【Meet Kevin】AI正在重塑医疗板块：这3只潜力股不容错过！ — original source. Check the original for full context.

Source published: 2026-10-24

Key ideas

莫德纳（Moderna）即将公布与默沙东合作的IV期黑色素瘤个性化癌症疫苗三期临床数据，AI算法已将靶点预测周期从60天大幅压缩至约两周并筛选出32个特异性切片。 (0:00)
抗生素耐药危机严峻，2025年美国耐药菌死亡人数达3.5万人，初创企业Radical Numerics正利用生成式AI精确改造噬菌体结合位点以实现靶向杀菌。 (7:00)
Tempus AI（TEM）高毛利的AI数据与应用业务实现28%的年增长，但最新季度净利润主要由约1.02亿美元的股票公允价值变动收益支撑，预计全面盈利需等到2028年。 (11:12)
医药巨头礼来（Eli Lilly）市值突破1.22万亿美元且PEG约为1.9，凭借充沛现金流和并购整合机制化解单药试验失败风险，被定位为AI医疗领域的“标普500”。 (12:34)

Detailed notes

## Background 该视频由BilibiliUP主“浑水摸鱼清源”于2026年10月24日发布，搬运并翻译了财经博主Meet Kevin的分析内容。视频主要围绕人工智能在医疗健康板块的颠覆性应用展开，深度剖析了莫德纳（Moderna）、Tempus AI和礼来（Eli Lilly）等三家极具代表性公司的业务亮点与投资前景。

## TL;DR Meet Kevin在视频中探讨了AI如何通过加速基因靶点预测重塑癌症疫苗研发，并重点介绍了莫德纳即将在周末公布的第四期黑色素瘤三期临床关键数据。同时，他分析了AI在噬菌体疗法应对抗生素耐药危机中的前沿应用，以及Tempus AI在软件数据层面的高速增长与账面收益隐忧。最后，Kevin将市值超万亿美元的制药巨头礼来定位为医疗AI领域的“标普500”，建议投资者在博取单药高弹性的同时，通过巨头配置有效对冲单一临床试验失败的二元风险。

## Atmosphere & style 视频整体呈现出快节奏、极具煽动性且干货满满的投资直播风格。Kevin借助白板手绘通俗易懂地拆解复杂的生物医药技术，语言直白生动、充满激情，同时穿插着严谨的财务审视与个人仓位披露。在谈及周末即将揭晓的二元催化剂和致命耐药危机时，语气中带着紧迫与刺激感，随后又冷静切换回理性配置策略，给人一种在惊险过山车中寻求稳健锚点的专业互动体验。

## Bulletin - [00:00](https://www.bilibili.com/video/BV1cHni64ENA?t=0s) 莫德纳（Moderna）即将公布与默沙东合作的IV期黑色素瘤个性化癌症疫苗三期临床数据，AI算法已将靶点预测周期从60天大幅压缩至约两周并筛选出32个特异性切片。 - [07:00](https://www.bilibili.com/video/BV1cHni64ENA?t=420s) 抗生素耐药危机严峻，2025年美国耐药菌死亡人数达3.5万人，初创企业Radical Numerics正利用生成式AI精确改造噬菌体结合位点以实现靶向杀菌。 - [11:12](https://www.bilibili.com/video/BV1cHni64ENA?t=672s) Tempus AI（TEM）高毛利的AI数据与应用业务实现28%的年增长，但最新季度净利润主要由约1.02亿美元的股票公允价值变动收益支撑，预计全面盈利需等到2028年。 - [12:34](https://www.bilibili.com/video/BV1cHni64ENA?t=754s) 医药巨头礼来（Eli Lilly）市值突破1.22万亿美元且PEG约为1.9，凭借充沛现金流和并购整合机制化解单药试验失败风险，被定位为AI医疗领域的“标普500”。

## Key takeaways - [00:00](https://www.bilibili.com/video/BV1cHni64ENA?t=0s) AI在医疗领域的结合正在颠覆传统研发模式，从基因测序到噬菌体工程均蕴含超越传统芯片和纯科技股的投资潜力。 - [00:00](https://www.bilibili.com/video/BV1cHni64ENA?t=0s) 莫德纳联合默沙东开发针对IV期黑色素瘤的个性化mRNA癌症疫苗，通过在12个月内注射约7剂训练免疫系统清除转移性癌细胞。 - [05:00](https://www.bilibili.com/video/BV1cHni64ENA?t=300s) 传统人工切片测序耗时约60天且仅能寻找极少靶点，AI技术将耗时缩短至两周内并可同时锁定32个特异性切片以抵御癌细胞变异逃逸。 - [05:00](https://www.bilibili.com/video/BV1cHni64ENA?t=300s) 莫德纳高度依赖单一关键疗法突破，面临极高的单品与单试验失败风险（Single-trial risk）。 - [07:00](https://www.bilibili.com/video/BV1cHni64ENA?t=420s) 全球抗菌素耐药性（AMR）危机日益严重，2025年美国因耐药菌感染致死人数（约3.5万人）已逼近全年车祸死亡人数（约3.66万人）。 - [10:00](https://www.bilibili.com/video/BV1cHni64ENA?t=600s) 传统广谱抗生素杀灭有益菌且面临抗药性，而AI算法能通过计算建模噬菌体的表面受体结合点实现精准杀灭病原体。 - [11:12](https://www.bilibili.com/video/BV1cHni64ENA?t=672s) Tempus AI核心的高毛利AI数据与应用业务年增速达28%，明显快于传统诊断检测业务20%的增长速度。 - [11:12](https://www.bilibili.com/video/BV1cHni64ENA?t=672s) Tempus AI最新财报利润受约1.02亿美元股票投资收益影响，主营业务目前尚未实现可持续运营盈利，市场共识盈利预期在2028年左右。 - [12:34](https://www.bilibili.com/video/BV1cHni64ENA?t=754s) 礼来（Eli Lilly）作为万亿市值制药巨头，通过大规模收购整合顶尖AI技术初创团队，有效规避了早期生物技术公司的二元试验波动风险。 - [12:34](https://www.bilibili.com/video/BV1cHni64ENA?t=754s) 礼来当前PEG估值约为1.9，对于追求低回撤的稳健型投资者而言是分享AI医疗红利的核心资产。 - [15:00](https://www.bilibili.com/video/BV1cHni64ENA?t=900s) 美国银行将莫德纳目标价上调至220美元，但其三期数据读出呈现极高的非对称性与二元催化特征。

## Detailed walkthrough ### [00:00](https://www.bilibili.com/video/BV1cHni64ENA?t=0s) AI in Healthcare and Moderna’s Phase III Catalyst

The application of artificial intelligence in healthcare represents an emerging frontier that many investors have overlooked, even as traditional tech and semiconductor trades dominate the conversation. Ahead of a scheduled presentation on Saturday, October 24th—timed deliberately outside market trading hours—Moderna faces a critical near-term catalyst involving Phase III clinical trial readouts. This update focuses on Moderna’s personalized cancer vaccine development in partnership with Merck’s Keytruda, specifically targeting stage IV melanoma to halt or prevent metastatic spread across the body. Meet Kevin notes that of the three stocks discussed (Moderna, Tempus AI, and Eli Lilly), he holds direct exposure to two of them, with portfolio notifications shared via his transparency service.

```mermaid flowchart TD accTitle: AI-Powered Cancer Vaccine Workflow accDescr: Step-by-step process showing biopsy, AI genome slicing, carrier injection, and immune targeting. subgraph Genomic_Analysis["Genomic Analysis"] A["Patient Tumor Biopsy"] --> B["Genomic Sequencing"] B --> C["AI Target Prediction<br/>(Identifies 32 unique slices)"] end subgraph Immune_Response["Immune Activation"] C --> D["mRNA Injection<br/>(~7 doses / 12 months)"] D --> E["Carrier Proteins in Lymph Nodes"] E --> F["Immune System Destruction of Cancer"] end ```

To explain the mechanism of action without getting bogged down in protein folding arcana, Kevin sketches the entire flow on a whiteboard: - A surgical biopsy extracts tissue from a solid tumor, followed by genomic sequencing of its nucleotide chains (the classic A, C, T, and G base pairs). - The objective is finding distinct slices of that cancer's genome that can be safely targeted without harming healthy tissue. - Because malignant cells constantly mutate and adapt, targeting just one or two markers is inadequate; Moderna designs the therapy to hit 32 unique genomic slices simultaneously so the immune system still detects the cancer even if mutations strip away some targets. - These target sequences are delivered via injection roughly seven times across a 12-month treatment cycle, training carrier proteins in the patient's lymphatic system to display the markers and cue immune cells to hunt down and eliminate metastatic cancer cells throughout the body.

---

[05:00](https://www.bilibili.com/video/BV1cHni64ENA?t=300s) AI Acceleration, Valuation Range, and Single-Trial Risk

The transformative impact of artificial intelligence lies in dramatically shortening the discovery timeline while increasing target density. In previous iterations, manual laboratory work required approximately 60 days merely to isolate two viable genomic slices from a biopsy. Machine-learning algorithms have cut that turnaround down to approximately two weeks while predicting 32 robust slices that reliably eliminate malignant cells without damaging vital organs.

Kevin notes his investment approach to Moderna, highlighting that rather than chasing speculative price run-ups prior to data drops, he waited for initial hype to settle and accumulated his position in the $135 to $145 per share range. While this therapeutic platform holds enormous potential to expand beyond melanoma to other solid tumors (such as lung and breast cancers) or eventually blood cancers like leukemia, Kevin stresses the inherent downside:

> “It’s very much, this stock is very much riding high on one treatment. I mean, they do others. But it’s just an example of how AI, it can enhance these healthcare plays. However, you have a lot of single trial risk here.”

Beyond single-stock biotech plays, broader opportunities exist through diagnostic data companies like Tempus AI, mega-cap pharmaceutical names like Eli Lilly, and private biotech startups tackling alternative biological mechanisms.

---

[07:00](https://www.bilibili.com/video/BV1cHni64ENA?t=420s) Antimicrobial Resistance and AI-Engineered Phage Therapy

AI is also being deployed to solve the escalating global crisis of antimicrobial resistance (AMR). A recent feature covered by *The New York Times* highlighted research from the startup Radical Numerics, which uses generative models to engineer bacteriophages—viruses that hunt and consume bacteria. Rather than relying on broad-spectrum antibiotics, which indiscriminately wipe out beneficial bacteria across the human microbiome, engineered phages can deliver highly targeted strikes against specific bacterial strains.

```mermaid flowchart LR accTitle: Phage Therapy Targeting Mechanism accDescr: Flow diagram showing how AI models optimize phage attachment structures to eliminate resistant bacteria. A["Resistant Pathogenic Bacteria"] --- B["AI-Optimized Phage Legs"] B --> C["Precision Surface Binding"] C --> D["Bacterial Lysis without Damaging Microbiome"] ```

AI computational tools optimize the binding structures on the phages' legs, allowing them to latch onto specific outer membranes of resistant bacteria and neutralize them. This therapeutic alternative has grown urgent as traditional antibiotic efficacy plummets.

- In the United States, annual deaths directly linked to antibiotic-resistant bacterial infections reached approximately 35,000 in 2025. - By comparison, U.S. motor vehicle traffic fatalities totaled roughly 36,640 over the same period, illustrating that deaths from drug-resistant superbugs now nearly match the annual death toll of car accidents on American roads.

[10:00](https://www.bilibili.com/video/BV1cHni64ENA?t=600s) Antimicrobial Resistance and AI-Driven Phage Engineering

Kevin highlights a critical and often overlooked healthcare crisis: antimicrobial-resistant bacterial infections. It is staggering to realize that patients are routinely dying from bacterial infections simply because conventional antibiotics no longer work against mutated, drug-resistant strains. Traditionally, attempting to match or discover an effective treatment manually takes months of intensive lab work, by which point the patient has often already succumbed. Kevin explains that artificial intelligence is directly transforming this paradigm through biotech startups like Radical Numerics. These teams utilize AI to rapidly screen, identify, and deploy bacteriophages—naturally occurring viruses that specifically hunt and destroy bacteria—transforming them into targeted precision weapons against deadly pathogens.

Beyond mere identification, AI enables bioengineers to actively design and modify phages. Kevin explains that scientists can computationally manipulate the structural attachment points—which he colloquially describes as the “little feet” of the viruses—to bind precisely to target bacteria. By taking biological engineering from slow, trial-and-error manual benchwork into predictive, algorithmic modeling, AI can identify life-saving interventions within days or hours rather than months. As Kevin bluntly emphasizes the life-or-death nature of this acceleration: “If it takes months to do the work manually, you're dead by then.”

- **The Clinical Threat:** Patients are dying from bacterial infections that standard antibiotics cannot treat due to advanced bacterial resistance. - **Phage Deployment:** AI allows researchers to predict which specific phage will save a patient's life and generate actionable solutions before the infection turns fatal. - **Precision Bioengineering:** Algorithms can model and engineer the physical receptor attachment points of phages to optimize their bacterial binding efficiency.

---

[11:12](https://www.bilibili.com/video/BV1cHni64ENA?t=672s) Software and Data Plays: Evaluating Tempus AI

Turning his attention toward pure-play healthcare software and data platforms, Kevin analyzes Tempus AI (ticker: TEM). While Tempus has generated considerable buzz as an AI-powered diagnostic and clinical data leader, Kevin urges investors to look closely under the hood of their financial statements. Tempus reported a positive net income in its latest quarter, but Kevin cautions that this headline profit does not reflect core operational turnaround; rather, it was driven by roughly $102 million in mark-to-market stock gains on investment holdings. Kevin notes that while diversifying corporate cash into market investments is understandable, paper portfolio gains obscure the true health of the underlying operations, and consensus estimates do not expect Tempus to reach sustainable operating profitability until 2028.

Despite these profitability headwinds, Kevin acknowledges that Tempus's operational trends show genuine momentum where it counts. The company's highest-margin segment—its AI data and applications division—is also its fastest-growing unit, compounding at 28% year-over-year compared to roughly 20% growth in its legacy diagnostic testing business. This favorable margin mix shift offers long-term promise for TEM, yet Kevin remains sidelined for now, arguing that it is still too early in the cycle to crown Tempus as the definitive long-term winner in clinical data infrastructure. Reflecting on their recent earnings report, Kevin explains: “You have to be clear that when they report a positive net income in this last quarter, it's really because they took about $102 million of stock gains.”

- **Profitability Timeline:** Consensus expectations do not project operational profitability for Tempus AI until roughly 2028. - **Accounting Nuance:** The reported quarterly net income was primarily bolstered by approximately $102 million in paper gains (marked to market) on stock holdings rather than organic operating earnings. - **Segment Divergence:** The AI data and applications business is expanding at 28% with expanding margins, outpacing the legacy segment's 20% growth rate.

---

[12:34](https://www.bilibili.com/video/BV1cHni64ENA?t=754s) Eli Lilly as the Diversified "S&P 500" of Healthcare AI

For investors who find speculative biotech startups and single-study binary catalysts too volatile, Kevin presents Eli Lilly as an alternative cornerstone holding. Trading at a PEG ratio of approximately 1.9—sitting right under Kevin's personal threshold of 2.0 for high-quality growth platforms—and sporting a market capitalization above $1.22 trillion, Eli Lilly functions essentially as the diversified blue-chip index of medical innovation. While a high-beta stock like Moderna could easily swing between $140 and $300 depending on a single weekend trial readout, Eli Lilly completely hedges individual clinical trial failure through its massive balance sheet, industry reputation, and aggressive acquisition engine.

```mermaid flowchart TD accTitle: Healthcare AI Investment Approaches accDescr: A comparison showing high-volatility single-trial biotech bets versus Eli Lilly's diversified corporate aggregator model.

subgraph BinaryRisk ["High-Volatility Biotech Play (Moderna)"] A["Single Clinical Trial Readout"] A -->|"Positive Trial Data"| B["Dramatic Upside ($300+)"] A -->|"Negative Trial Data"| C["Severe Collapse ($140 Floor)"] end

subgraph AggregatorModel ["Diversified Aggregator Play (Eli Lilly)"] D["$1.22T+ Balance Sheet & Free Cash Flow"] D --> E["Acquisition Engine"] E --> F["Absorb AI Software & Tech Startups"] E --> G["Diversified Multi-Vertical Pipeline"] F & G --> H["Low-Drawdown Compound Growth (PEG ~1.9)"] end ```

Kevin describes Eli Lilly as an “acquisition machine” that systematically sweeps up promising technology startups, specialized AI software, and clinical-stage competitors the moment they gain traction. Rather than forcing investors to guess which individual AI lab or software platform will succeed, Eli Lilly lets market winners emerge and then uses its immense capital to buy them out. During this discussion, Kevin also plugs his ReinvestOne platform, highlighting the Alpha membership, daily morning Alpha reports, real-time portfolio alerts, and an upcoming financial research terminal designed to offer Bloomberg- and Refinitiv-caliber data alongside embedded Homes AI, research modules, and agentic AI blueprints. Framing Eli Lilly's unique role in a portfolio, Kevin argues: “Eli Lilly is already an over $1 trillion company but I call them the S&P 500 of artificial intelligence, because this meeting on Saturday could go horrible for Moderna.”

- **Valuation Metric:** Eli Lilly trades at an attractive PEG ratio of approximately 1.9. - **Capitalization Power:** With a market valuation exceeding $1.22 trillion, Lilly possesses virtually unmatched capital to acquire emerging biotech startups and AI tools. - **Platform Announcement:** Kevin previews the upcoming Reinvest terminal, positioning it as an accessible, unified alternative to legacy financial terminals featuring integrated research, video courses, and AI workflows.

---

[15:00](https://www.bilibili.com/video/BV1cHni64ENA?t=900s) Moderna's Binary Weekend Catalyst and Asymmetric Risk

Kevin brings the discussion back to Moderna, which stands at an inflection point heading into its weekend Phase 3 cancer therapy data presentation. Following the peak of the pandemic, Moderna's revenue trajectory collapsed into what Kevin characterizes as merely milking legacy COVID-19 mRNA vaccine demand, leaving the company without a dynamic growth driver. That stagnation could reverse abruptly if the Phase 3 individualized neoantigen therapy data—developed alongside Merck for Stage IV melanoma—validates mRNA and AI-driven neoantigen sequencing as a viable oncology platform. Bank of America has already raised its price target on Moderna to $220, representing substantial upside from its current market cap range of $70 billion to $93 billion.

However, Kevin emphasizes that this setup is the definition of high-stakes asymmetric volatility. If Saturday's clinical readout delivers compelling efficacy, Moderna could recapture the speculative market enthusiasm and retail virality that defined its pandemic ascent. Conversely, if the trial results disappoint or fail to clear efficacy hurdles, the stock risks plunging right back to its baseline lows. Kevin concludes the livestream by engaging with viewer comments about market swings and advertising his tools, reiterating that while high-risk clinical events are not personalized financial advice, the intersection of AI with mRNA sequencing represents one of the most exciting frontiers in modern finance. Capturing the stock's roller-coaster historical trajectory, Kevin remarks: “When you sort of zoom out on the stock chart, it shows you the virality that this company used to have that doesn't have anymore.”

- **Wall Street Target:** Bank of America has boosted its Moderna price target to $220 ahead of major oncology readouts. - **Valuation Context:** Moderna currently trades at an enterprise valuation between roughly $70 billion and $93 billion. - **Binary Catalyst Profile:** The stock's post-COVID lull makes the Phase 3 cancer therapy data a pivotal make-or-break catalyst that could either re-ignite massive momentum or trigger a sharp breakdown.

## Mentioned Meet Kevin, Moderna, Merck, Keytruda, Tempus AI, Eli Lilly, Radical Numerics, Bank of America, ReinvestOne, Bloomberg, Refinitiv, Homes AI

## Video description 【Meet Kevin】 AI正在重塑医疗板块：这3只潜力股不容错过！ The NEXT Frontier： THESE 3 Stocks Boom

掘金AI医疗：不可忽视的下一个超级风口大家好，我是Meet Kevin。今天我们来聊聊人工智能在医疗板块的深度应用，以及那些正在悄然迎来重大突破的创新公司。很多朋友把目光全锁死在传统的科技巨头和芯片上，但实际上，AI正在医疗健康领域掀起一场颠覆性的革命。从基因测序到噬菌体疗法，科技与生物医药的结合不仅在拯救生命，更在孕育巨大的投资机会。

莫德纳的癌症疫苗决战与AI加速作为我个人重点配置标的之一，莫德纳（Moderna）即将在周六迎来极其关键的三期临床数据更新。这次发布会重点展示其与默沙东合作的针对第四期黑色素瘤的癌症疫苗（免疫疗法）数据。过去，人工从癌症细胞中测序并寻找32个特异性切片靶点需要长达60天的时间；而现在借助AI技术，这一过程被缩短至两周内完成，不仅大大降低了成本，还显著提高了预测准确性。目前莫德纳市值在700亿至930亿美元之间，美国银行给出的目标价已上调至220美元。我在近期股价经历回调、市场情绪降温时，选择在135至145美元区间进行了建仓。

抗菌素耐药危机与AI在噬菌体疗法中的应用除了癌症治疗，AI在应对全球抗生素耐药危机中同样大显身手。数据显示，美国2025年的耐药菌感染直接导致的死亡人数高达35,000人，几乎与全年36,640起机动车交通事故死亡人数持平。传统抗生素不仅效率低下且不分敌我地杀死有益菌，而通过AI技术精确设计和改造噬菌体（能够吞噬有害细菌的病毒），可以实现对病原体的精准打击。初创企业如Radical Numeric等正利用AI优化噬菌体的结合位点，这标志着精准医疗正在迈向全新的发展阶段。

软件与数据服务商的多元化考量在软件和服务层，像Tempus AI（股票代码：TEM）这样的公司也备受关注。其AI数据与应用业务保持着28%的高速增长，且毛利率表现优于传统老业务。不过需要理性看到的是，TEM目前尚未实现真正意义上的持续盈利，最近财报显示的净利润增长很大程度上源于约1.02亿美元的股票投资收益（标记公允价值），而非纯粹的主营业务利润，预计公司全面盈利可能要等到2028年。因此，对于投资者而言，单赛道创新公司的单品和单试验风险（Single-trial risk）依然较高。

行业巨头的低风险配置选择如果你觉得像莫德纳这类高波动、高风险的生物科技公司不适合自己的风险承受能力，那么像礼来（Eli Lilly）这样的制药巨头则是更稳健的选择。礼来目前市值超过1.22万亿美元，堪称医疗人工智能领域的“标普500”。凭借雄厚的资本实力与强大的并购整合能力，礼来通过不断收购业内顶尖的初创企业和技术团队，有效分散了单一药物研发失败的风险。其PEG估值指标目前维持在1.9左右，对于希望分享医疗创新红利但又追求低回撤的投资者来说，是一个极具吸引力的优质底层资产。

可执行操作指引理性评估风险承受能力：对于莫德纳等处于重大临床数据发布节点的高波动标的，切忌盲目追高，应等待情绪宣泄和回踩后再寻找建仓机会。构建攻守兼备的医疗组合：将高弹性的创新生物科技股（如莫德纳）与抗风险能力强的行业巨头（如礼来）进行均衡配置，降低单一试验失败带来的组合震荡。密切跟踪行业关键催化剂：重点关注周末及财报季发布的临床试验进展与FDA审批动态，根据最新数据及时调整仓位策略。

简洁时间轴 [00:00] 人工智能在医疗领域的突破与本周关键催化剂 [01:00] 莫德纳癌症疫苗三期数据与AI赋能基因测序 [07:00] 抗菌素耐药危机与AI在噬菌体疗法中的应用 [11:00] 软件与数据服务公司Tempest AI的业务表现 [12:30] 医药巨头礼来作为AI医疗领域的标普500 [15:10] 莫德纳美银目标价与当前高波动投资风险

[Source](https://www.bilibili.com/video/BV1cHni64ENA)

中文简报

Meet Kevin：AI重塑医疗行业，关注3只潜力股 Meet Kevin在视频中探讨了AI如何通过加速基因靶点预测来重塑癌症疫苗研发，并重点介绍了莫德纳即将在周末公布的IV期黑色素瘤三期临床关键数据。同时，他分析了AI在噬菌体疗法应对抗生素耐药危机中的前沿应用，以及Tempus AI在软件数据层面的高速增长与账面收益隐忧。最后，Kevin将市值超万亿美元的制药巨头礼来定位为医疗AI领域的“标普500”，建议投资者在博取单药高弹性的同时，通过配置巨头有效对冲单一临床试验失败的二元风险。莫德纳（Moderna）即将公布与默沙东合作的IV期黑色素瘤个性化癌症疫苗三期临床数据，AI算法已将靶点预测周期从60天大幅压缩至约两周并筛选出32个特异性切片。 (0:00) 抗生素耐药危机严峻，2025年美国耐药菌死亡人数达3.5万人，初创企业Radical Numerics正利用生成式AI精确改造噬菌体结合位点以实现靶向杀菌。 (7:00) Tempus AI（TEM）高毛利的AI数据与应用业务实现28%的年增长，但最新季度净利润主要由约1.02亿美元的股票公允价值变动收益支撑，预计全面盈利需等到2028年。 (11:12) 医药巨头礼来（Eli Lilly）市值突破1.22万亿美元且PEG约为1.9，凭借充沛现金流和并购整合机制化解单药试验失败风险，被定位为AI医疗领域的“标普500”。 (12:34)

Reading summary

Investor Meet Kevin argues that artificial intelligence is transforming healthcare by radically compressing discovery timelines across oncology, infectious diseases, and diagnostics. While single-asset biotechs face high-stakes binary catalysts, broader exposure can be calibrated through clinical data platforms and mega-cap pharmaceutical aggregators. AI-Driven Personalized Oncology and Moderna's Binary Risk (0:00) Moderna, in partnership with Merck, is deploying machine-learning algorithms to produce individualized mRNA cancer vaccines targeting stage IV melanoma alongside Keytruda. Conventional manual laboratory procedures required approximately 60 days merely to extract two viable genomic markers from a solid tumor biopsy. AI workflows have accelerated this turnaround to roughly two weeks while identifying 32 unique genomic slices simultaneously, ensuring the immune system maintains targeting efficacy even if malignant cells mutate. The therapeutic regimen delivers these target sequences via approximately seven injections across a 12-month cycle to direct lymphatic carrier proteins to eliminate metastatic cells. Kevin cautions that Moderna carries acute single-trial risk. Following the erosion of its pandemic COVID-19 vaccine revenues, the company is reliant on its upcoming Phase 3 neoantigen clinical readout scheduled outside market hours on October 24th. Although Bank of America set a $220 price target against Moderna's valuation of $70 billion to $93 billion, Kevin notes that the stock—which he accumulated between $135 and $145—could experience asymmetric swings between $140 and $300 based on the readout's outcome. Engineering Bacteriophages to Combat Antimicrobial Resistance (7:00) AI applications are expanding into antimicrobial resistance (AMR), where traditional antibiotics fail against evolved superbugs. In the United States, annual deaths from antibiotic-resistant bacterial infections reached roughly 35,000 in 2025, rivaling the approximately 36,640 domestic traffic fatalities recorded over the same timeframe. Identifying effective therapies manually often requires months of laboratory benchwork, frequently outpacing patient survival timelines. To bypass these delays, startups like Radical Numerics employ generative AI models to engineer bacteriophages—viruses that selectively hunt and consume targeted bacteria while sparing beneficial human microbiota. Computational design tools optimize the binding receptors on phage attachment structures within days or hours, enabling engineered viruses to latch onto resistant bacterial membranes and trigger lysis. Evaluating Tempus AI: Operational Momentum versus Paper Gains (11:12) In the software and clinical diagnostics sector, Tempus AI (TEM) has attracted market focus, yet Kevin advises close examination of its financial reports. While Tempus posted positive net income in its latest quarter, the profit stemmed from roughly $102 million in paper, mark-to-market investment gains rather than core operations. Wall Street consensus does not anticipate sustainable operating profitability until approximately 2028. Despite the accounting distortion, Tempus exhibits solid underlying business trends. Its highest-margin division, AI data and applications, is compounding at 28% year-over-year, outpacing the 20% growth rate of its legacy diagnostic testing segment. Kevin remains on the sidelines, however, reasoning that it is premature to determine whether Tempus will emerge as the definitive market leader in healthcare data infrastructure. Eli Lilly as a Capitalized AI Aggregator Hedge (12:34) For investors seeking insulation from clinical-trial volatility, Kevin highlights Eli Lilly as a diversified foundation, characterizing it as the medical sector's 'S&P 500' for AI innovation. Backed by a market capitalization exceeding $1.22 trillion and trading at a PEG ratio of roughly 1.9—below Kevin's personal threshold of 2.0 for high-quality compounders—Eli Lilly effectively diversifies single-study clinical risks through its extensive pipeline and capital reserves. Rather than attempting to forecast individual startup successes, Eli Lilly functions as an acquisition vehicle, deploying its substantial balance sheet to acquire validated AI software platforms, specialized laboratories, and clinical competitors once they gain market traction. This aggregator model captures AI-driven medical upside while mitigating the severe drawdowns associated with binary clinical data releases. While AI dramatically enhances biological target prediction and engineering speed, clinical applications remain tethered to binary regulatory milestones and accounting nuances. Investors must differentiate between volatile, single-catalyst biotech bets and cash-generative industry aggregators capable of absorbing external AI breakthroughs.

精炼阅读摘要

Meet Kevin指出，人工智能正在深度重塑医疗健康领域，加速了个性化癌症疫苗研发与抗微生物耐药性治疗突破；与此同时，医疗AI相关标的在资本市场上呈现出从高波动单向临床催化剂到稳健聚合型巨头的显著分化。 Moderna与AI癌症疫苗的二元催化 (0:00) Moderna正迎来关键的业务催化节点，其与默克公司（Merck）旗下Keytruda联合开发的个性化癌症疫苗即将公布针对IV期黑色素瘤的III期临床试验读数。在传统实验室流程中，人工提取并分离出2个可用的基因切片需耗费约60天；而通过引入机器学习算法，这一靶点预测周期被缩短至两周左右，且能同时锁定32个独特的基因切片，大幅降低了癌细胞突变逃逸导致的失效风险。整个治疗流程依托患者活检进行基因组测序，并在12个月内通过约7次mRNA注射训练淋巴系统，引导免疫细胞精准清除全身转移性恶性肿瘤。单一试验风险与极端估值波动 (5:00) 尽管该个性化平台理论上可拓展至肺癌、乳腺癌及白血病等实体瘤与血液肿瘤，美国银行（Bank of America）亦将其目标价上调至220美元，但Kevin强调Moderna目前面临极高的“单一试验风险”。随着疫情后COVID-19疫苗需求的自然消退，公司估值回落至700亿至930亿美元区间，业绩极度依赖该肿瘤管线的兑现。Kevin透露其建仓成本位于135至145美元区间，但他警示称该试验具有极强的二元不对称性：若临床数据超预期，股价或将重新点燃市场狂热并冲击300美元以上；但若疗效不达标，股价将承受大幅回撤至基本面低点的压力。 AI噬菌体工程破解耐药超级细菌 (7:00) 除肿瘤疫苗外，AI正被部署用于应对全球抗微生物耐药性（AMR）危机。传统广谱抗生素在消灭病原体的同时会破坏人体有益微生物组，且细菌突变已令常规药物疗效严重受损。数据显示，2025年美国因抗生素耐药感染死亡的人数达约3.5万人，与同期全美机动车交通事故致死人数（约36,640人）相近。针对传统人工筛选噬菌体耗时数月、极易延误患者救治的痛点，初创公司Radical Numerics利用生成式AI模型对噬菌体实施精准工程化改造。 Tempus AI的财务真实底色与隐忧 (11:12) 在纯软件与临床数据平台方面，Tempus AI（代码：TEM）尽管备受瞩目，但Kevin提醒投资者注意其财报中的会计细节。Tempus在最新财报中录得净利润转正，但该盈利主因是其投资持仓确认了约1.02亿美元的按市值计价账面收益，并未反映底层核心业务的实际扭亏。华尔街共识预期认为该公司要到2028年左右才能实现可持续的运营盈利。不过其业务结构正迎来积极转变，高毛利的AI数据与应用部门同比增长28%，明显快于传统诊断检测业务20%的增速，但Kevin目前仍选择在场外继续观望。礼来：医疗AI领域的多元化对冲基石 (12:34) 对于寻求规避单一生物科技临床黑天鹅事件的投资者，Kevin将市值超过1.22万亿美元的礼来（Eli Lilly）定义为医疗AI赛道的“标普500指数”。礼来当前的市盈增长比（PEG）约为1.9，处于2.0以内的优质成长股估值区间。相较于Moderna因单一临床结果可能引发的剧烈剧震，礼来凭借庞大的资产负债表与充沛现金流构建了极强的安全垫。Kevin称其为系统化的“并购机器”，即允许初创实验室承担早期试错成本，并在AI软件和创新管线验证成功后迅速进行并购整合，为投资者提供低回撤的长期复合增长敞口。 AI技术正在大幅压缩靶点发现与生物工程设计周期，但医疗AI领域的投资逻辑分化显著：单一管线生物科技股面临极高的临床二元归零风险，纯数据软件标的受制于较长的运营盈利周期，而兼具资本厚度与并购整合能力的医药巨头则成为对冲研发不确定性的核心通道。

Original transcript

[0:00] The next sector for artificial intelligence might be hiding in plain sight. And we should talk about some of the breakthroughs that are going on, as well as Something happening this Saturday that could be another catalyst for a particular stock that's been doing pretty well. What we're talking about in this is everything to do with how AI is affecting already and going to affect AI. healthcare stocks. some of the stocks we'll touch on here are going to be tempest ai Moderna And Eli Lilly. Okay. Now, I want to be clear, out of these three, I've got exposure to two of them. As always, we always send our portfolio transparency notices to everybody at reinvestone.com. In that membership, you pay once, you get lifetime access. You could read all about it at reinvestone.com. But for now, let's focus on artificial intelligence. in medicine. So the first thing to know, is that Moderna has a meeting coming up this Saturday. It is going to be a presentation. They purposefully do it when the market is closed. It's going to be a presentation on October 24th. It's the nearest catalyst, and it is an update on phase three data for their cancer therapy, which is the partnership between Moderna, Moderna and Merck's key true to cancer therapy or immunotherapy drug in fighting stage for melanoma and minimizing the risk of spread throughout your body metastasis now how does this work well how it works is and I'm gonna oversimplify because it gets really boring otherwise But basically, how does AI play into this world? Okay. That's really the part that we want to talk about with Moderna. And then I'm going to talk about Tempest AI. And then I'm going to talk about Eli Lilly to show you how AI is touching on some of these, as well as some startups working on artificial intelligence, which somewhat spawned some of us talking about this this morning in our live stream. But take a look at this. This is a whiteboard. Okay, so what does Moderna do? Moderna basically says, hey, let's go ahead and look at somebody's cancer. So we're going to take a hard cell of cancer and we're going to surgically remove a portion of it. We're going to take a biopsy. So we're gonna take a biopsy of you know a cancer cell and then what we're gonna do is we're gonna genomically see whoa Jeez we're gonna genomically sequence that which is you know very long chain of genetic data Think about the T's and the A's and the C's remember back in bio days, right? What we're gonna try to do is we're gonna try to find a of that cancer cell, where are their slices of that genome that we can target for elimination, which obviously means we do not want to target parts of that genome that if we targeted would kill you. Duh, that's a problem. We want to target parts of that genome that are unique to to that cancer cell. The problem is cancer cells mutate all the time. They're always changing. So this underlying genome changes a lot. So what we're going to do is we're going to target 32 different slices And if any single one of those slices survives, we can still target that cancer cell. What we're going to do is we're going to introduce these little slices via injection into you probably seven times over a 12-month cycle and we're going to train your body we won't get into the exact science of it because it gets it gets a little deep with with protein folding and prediction what ends up happening in your lymph nodes basically little carriers, okay, little carriers are going to get trained to, On these 32 slices. And they're going to take these little 32 slices and they're going to attach them. to their outside they're gonna bring it on over to your lymphatic system And they're going to say... you know, See all these little spikes we got here? Kill those. Kill these little proteins we're carrying. Identify and kill those. And then your lymphatic system is going to say, we don't like that. We are going to, how can I make an angry face? There we go. Angry face. We're going to go after those little spikies. Okay, but these spikies are really hard to identify. And this is where AI gets in. Because these guys can go throughout your whole body and go identify where these are and go track down and kill that cancer cell from metastasizing through the rest of your body, right? Okay. And the whole idea is that even if that cancer cell is mutating or evolving, As long as we have one of those slices left over, we could still identify it. That's the idea. The problem is, it [5:00] to take to do this identification process, it used to take somewhere around 60 days and you would get two of them, maybe two slices. Okay, that has been transformed with AI. AI now comes in and says we can do a lot of this in about two weeks and we can get 32 predicted slices that we actually think will A, kill the cancer, and B, not kill you. Both of those things are kind of important, right? So AI is speeding up the time to doing this. And this meeting coming up on the 24th, on Saturday, is expected to provide a lot more information on how successful this Phase III cancer readout is going and how the therapies are going. Obviously, you know, it could also be bad. But given that it's a presentation on Saturday, A lot of people are optimistic that they're gonna put on a pretty good show And I personally am exposed to Moderna stock. I'm very excited about it We we didn't buy it before This this news on the phase three we actually waited for some of the initial momentum to die And we bought it right around here around the 135 to 145 area And so I very excited about Moderna because of how AI is more rapidly enabling more patients to be able to utilize this, reducing the cost of this. It's the long-term goal. And ultimately, the hope is that these sort of therapies could then be expanded to other cancers as well, not just melanoma or other sort of hard tissue cancers, but broadly others. Because this same treatment could already work in lung cancers or where you have actual lumps, maybe even breast cancers. But it'd be nice to see this spread to even other types of cancers, like blood cancers, leukemia style, right? Anywho, Moderna is a special play in this space. It's very much, this stock is very much riding high on one treatment. I mean, they do others. But it's just an example of how AI, it can enhance these healthcare plays. However, you have a lot of single trial risk here. And that where other plays like Tempest startups maybe that you can get your hands on investing in Eli Lilly get very interesting because it not just a company like Moderna that working this kind of stuff The New York Times actually had a great piece that linked out to research that this new company is doing called, well, the startup is doing, called Radical Numerics. And this was the piece right here. If you or your child spent years growing up on social media and experienced serious mental health struggles, you should hear this. Recently, a jury ordered Meta and YouTube to pay millions for their role in designing and promoting addictive platforms. They knew their apps could contribute to anxiety, depression, and body image issues, but did not share those risks while prioritizing profits. The recent historic verdict underscores the consequences of those decisions. Now, Morgan & Morgan is stepping up to hold these platforms accountable for the harm they cost. Click the link in my bio to take a short quiz, and you may be entitled to a potential recovery of over $1,000. It's quite easy to do. Go to the link in the description, fill out the form for yourself or a loved one. It quick easy and you can recover a potential recovery of over This was the piece right here They basically make this argument that they using AI to help them identify similarly how I showed you those little spikies, except now instead of spikies, they're going to try to use AI to genetically engineer or match phages, which are the enemies of bacteria. And these phages are going to have their little legs modified, enhanced by AI, so that they can properly attach to bad bacteria somebody is infected with and kill that bad bacteria. Now, you might think, why not just use an antibiotic? Well, a couple of reasons. One, antibiotics kill a lot of bacteria, including good bacteria. but in addition to that we're actually seeing deaths now for antibiotic resistant deaths in the United States roughly match traffic fatalities in 2025 we had about 36,640 motor vehicle deaths in the United States compared to 35,000 direct links to antibiotics [10:00] Antimicrobial resistant infections. which is crazy. to think about, that people are literally dying because of bacterial infections that we can't treat because those bacteria are resistant to treatment. AI, through this company, Radical Numerics, could potentially be used To weaponize... these little phages, which are like viruses that go kill bacteria, And we can not only predict which phase would save your life, and do it rapidly before you're dead, because that's the other thing. If it takes months to do the work manually, you're dead by then. So AI speeds it up. It's just another example of how AI is affecting healthcare. Or you could engineer the phages to do exactly what you want by manipulating their little attachment points so to speak little feed of these viruses if you will i'm trying to again i'm Now, That's cool. for that particular lab. But how about other plays like How about the software play? If you look at a software kind of play, there's Tempest. Tempest isn't expected to be profitable until 2028. They did report a profit, but that was mostly because they had a ton of stock gains on their income statement over here. So less ideal because... You know, it tells you less about the underlying business. I don't have a problem with a company making money or investing in stocks. I think it's a good use of diversifying some of the cash that they have and having exposure to other aspects of the market while still doing what it is that they're doing. You know, you have to be clear that when they report a positive net income in this last quarter, it's really because they took about $102 million of stock gains. good for them. And that doesn't necessarily mean they sold, it just means they have marked to market those gains. Here's a company that is growing their AI side of the business. at the fastest portion of the entire business. They're growing the AI side, data and applications at 28%. The legacy side of the business is growing at about 20%. Their margins are actually better on the AI side, and that's the side that's growing faster. So there's a lot of hope for a company like Tempest AI, ticker TEM. However, I don't invest in them. I still think it's... potentially a little early to say that they're gonna be, you know, the winners of data and apps. And so I'm gonna show you a different play. This next play, has about a 1.9 peg ratio, which is kind of a magic number where I like being under two for some of these service providers. And it's this company right here. It's Eli Lilly. Eli Lilly is already an over $1 trillion company but I call them the S&P 500 of artificial intelligence, because this meeting on Saturday could go horrible for Moderna. And then it's $140 stock again, and you've lost all your gains. It could also go great and it's a $300 stock. That's your high volatility play. you're less risky less YOLO we play in my opinion is a play like Eli Lilly obviously nothing in this you know, video is personalized financial advice. My goal is just to provide perspective and data and insight. It's the same thing that we do over at reinvestone. I encourage you to join us. You go reinvestone.com. Join us here. You could be a part of the alpha membership. You get all the trade portfolio alerts. You get our alpha report every single morning. You get our stock terminal coming out later this year. I cannot wait for this to come out. So this will be coming out later this year. This will be very, very useful. It's basically... we're trying to create everything that Kevin would use in like a Refinitiv or a Bloomberg terminal in one beautiful package for you to get in the reinvest terminal for a fraction of the price. You get the Homes AI built into it, the Reinvest Research built into it, all the nine courses, course member live streams, as well as our agentic AI business and consumer blueprint, which is really exciting. A lot of cool things coming for this. So really good package you get at reinvestone.com right now. But anyway, you go back to Eli Lilly. I see Eli Lilly as somewhat of this S&P 500 player. of healthcare. And the reason I say that is rather than picking a bunch of individual plays in healthcare, you can kind of look at Eli Lilly as anybody who starts succeeding and starts getting big. Eli Lilly just comes in and goes, you know what? We're going to buy your software. We're going to buy your startup. They do really good at acquisitions. They're an acquisition machine. I mean, they are just like literally... You know, suck up. All the competition that they could get their hands on and they've got the capital do it. They've got the reputation to do it. They've got the profitability to do it. They've got the market cap to do it. I mean, at at. over $1 trillion market cap. Yeah, $1.22 trillion market cap. This is a It's a pretty fat company. And rightfully so. So this, I think, is the lowest risk play. And I still think their valuation is reasonable, despite the fact that they've done very well already. [15:00] So that's sort of my take on how health care is affecting AI and some of the things that I'll be looking at this weekend, especially with that presentation from Moderna. So I'm pretty excited about that. Price target from Bank of America now up at $220 for Moderna. And mind you, Moderna is what a 70, $93 billion market cap company right now. So, after covid and that's really what's been sustaining their income after covid Their growth prospects really plummeted and it turned into this business that was just milking the COVID mRNA vaccine. And they haven't really had a growth vertical trend. until this phase three cancer vax, people are calling it. It's really a therapy. But I'm personally very, very excited about it. When you sort of zoom out on the stock chart, it shows you the virality that this company used to have that doesn't have anymore. But there's a potential it could come back. Unless, of course, they have a bad report on Saturday. Then it just goes right back down to the floor. So, not personalized financial advice. And join us over at... reinvestone.com. And I know there's a comment here, don't leave after this, go refill your mug and put us watching commercials. We don't want that's very nice of you to say, but I have about a million and a half things I got to get done. So I have one very strong message for you, especially for you. Oh, man, your boy Kevin just lost lots of money. Oh, that's okay. I am with a lot. Why not advertise these things that you told us here? I feel like nobody else knows about this. We'll try a little advertising and see how it goes. Congratulations, man. You have done so much. People love you. People look up to you. Kevin Paffroth there, financial analyst. And YouTuber, Meet Kevin, always great to get your take.