원문 보기
Maya, quick question. What would you pay for a single day of trading where the market accidentally tells the truth? Ooh, explain. Because in August of last year, for about 6 and a half hours, Wall Street priced a drug company like a software company, and $44 billion showed up out of nowhere. Kathy Wood was on stage at Moonshots Live when the dust settled, and she said the quiet part out loud.
Healthcare is the most underpriced AI theme in the market, and this was just the first blink. All right, I'm in. Let's deconstruct the blink. Part one, abundance hook. Quick heads up before we start. We're about to talk about a stock that moved a lot, a market thesis, and some real clinical results.
Nothing you're about to hear as financial advice. It's a conversation about what artificial intelligence is doing to medicine and why the people who study exponential curves think the market still doesn't get it. All right, picture this. It's August 19th, 2026. I'm your trader on Wall Street. You open your terminal and one of the stocks on your screen, not a meme stock, not some micro cap penny stock, one of the most heavily followed biotech companies on Earth, is up 177%.
In a single day, it nearly tripled before lunch. $44 billion of value appeared in about 6 hours of trading. The company was Madna. And here's the part that makes this a story about artificial intelligence and not just a story about a good drug trial. The reason the stock moved wasn't that Madna stumbled onto something by accident in a lab.
It was that their personalized cancer vaccine worked. A vaccine designed patient by patient by reading a person's tumor, feeding that information into an AIdriven design process and manufacturing a bespoke treatment. Biology done the way software is done. This is the detail Kathy Wood was pointing at on stage at Moonshots Live 2026.
She sat down with Peter Diamandis and Immod Mustaki and circles Nikquille Chandok and she made this claim. Healthcare is the most underpriced AI theme in the market. And she pointed at Madna's record 1-day gain as the first piece of evidence, the market's first blink, the moment Wall Street briefly noticed what was happening.
So, this episode is about that blink. What did the market see for 6 hours on August 19th? Why does Kathy Wood think the pricing is still wrong? And what does underpriced even mean when we're talking about intelligence meeting biology? Two exponential curves colliding inside one human body. Part two, deconstructing the breakthrough.
Let's deconstruct what actually happened because the numbers deserve a slow walk. The treatment is called inaran oagene. Madna developed it with Merc and it's being tested alongside Kitruda, Merk's well-known immunotherapy. The phase 3 trial called Interpath 001 enrolled 1,137 patients with high- risk melanoma who had just had surgery.
The goal, stop the cancer from coming back. And here's what makes it personalized medicine rather than just another drug. When a patient enrolls, doctors sequence their specific tumor. They identify the mutations, the neo antigens, that are unique to that person's cancer. Then Merna manufactures an mRNA sequence tailored to those mutations, encoding up to 34 of them.
The patients own cells read that mRNA and display the neo antigens which trains the immune system to hunt down anything wearing those markers. It's a wanted poster drawn from your own crime scene printed just for you. The AI thread runs through the whole pipeline. Designing which neo antigens to encode, that's a prediction problem.
Manufacturing a bespoke therapy for each of 1,137 patients, each one different, that's a logistics and design problem that collapses without machine learning doing the heavy lifting. This is not a pill made in a factory run. It's software compiled for a patient. And on August 19th, the trial hit. It met its primary endpoint, recurrencefree survival, and a key secondary endpoint around distant metastasis.
It was, by the reporting at the time, the first successful large-scale late-stage trial of an mRNA based cancer therapy. The stock went from roughly $63 to 174.38 at the close. Reuters reported trading volume at about 15 times the 50-day average. The market capitalization jumped by an estimated $44 billion in one session.
Now, optimism check because we always do the honest version first. The topline announcement in August withheld the numerical results. The detailed data came later at a presentation in Madrid and it was genuinely mixed good news. The official findings a 42% reduction in the relative risk of recurrence or death and a 38% reduction in the risk of distant metastasis or death versus Kitruda alone.
Overall survival data was still immature, trending favorable but not proven. Some analysts said the distant metastasis benefit came in below the elevated expectations the August hype had set. The stock sold off afterward, though it later hit a 52- week high of about $254. So, no, this isn't a the trade is proven, go buy the stock story.
It's a the market glimpsed the platform and couldn't unsee it story, right? Because here's what that 177% day was actually repricing. One analyst estimate put Mon Merna's potential 2035 melanoma sales at about $3 billion. The market added $44 billion of value in a day, roughly 15 times that estimate.
The market wasn't pricing a melanoma drug. It was pricing a platform. One successful personalized medicine trial is a proof point. A working platform for manufacturing bespoke therapies is an entirely different company. Wall Street finally looked at Madna and saw software margins sitting inside a drug company.
and the follow-through kept coming. In October, the National Institutes of Health announced plans for a public private cancer initiative fasttrack style starting that December explicitly modeled on the speed of the COVID vaccine effort targeting pancreatic, liver, and colurectal cancers plus pediatric tumors.
Madna joined the NASDAQ 100. And every one of those headlines reads the same way to a thesis like Kathy Woods. The evidence is arriving in waves and the repricing has barely started. It's worth zooming in on what happened between the spike and the follow-through because the sequence tells you how markets actually process exponential news.
After the August pop, analysts rushed to raise targets. RBC Capital went to $130. Bank of America upgraded and went to $170, but plenty of those targets still sat below the new trading price, and some firms called it a do not trade name on the wild swings. Then came the Thursday data leak ahead of the Madrid presentation.
The stock cracked and the sell the news crowd declared the whole thing over. Except it wasn't over because the underlying numbers were strong. The Madrid presentation confirmed the trial hit both end points. Merc and Madna expanded their collaboration with Tempest, the AI powered precision medicine company to support commercialization.
And in October, a New York Times report on the NIH's fasttrack cancer program sent the stock up another 14% in a day on the simple logic that a national push to replicate the CO era vaccine speed would flow straight into the companies with working mRNA platforms. The pattern spike, doubt, leak, recovery, institutionalization.
Every cycle, the price discovery gets a little less frantic and a little more structural. That's the deconstruction. A personalized vaccine designed per patient, manufactured per patient, proven at phase 3 scale, repriced by a market that keeps oscillating between pricing a product and pricing a platform.
Now, why does Kathy Wood think the whole sector is still mispriced? Part three. Yes. And historical context. Markets linearize exponentials. Kathy Wood's shop, Arc Invest, is built on one observation that repeats across every technological revolution. The market prices the present and discounts the future linearly while technology compounds exponentially.
The internet in 1996, computing in 1975. Every single time Wall Street looks at a cost curve falling in half every couple of years and says yes, but this quarter's revenue. And she's not making a single technology bet. ARK's whole framework is five innovation platforms converging at once. Artificial intelligence, robotics, energy storage, blockchain, and multiomic sequencing.
That's the ability to read the full biological instruction set of a cell, not just the genome. Her argument is that when platforms converge, the compounding multiplies. AI doesn't just make drug discovery faster. AI plus cheap sequencing plus robotics run labs changes the whole production function of medicine.
This is also where her bigger economic call connects. She's argued that real GDP growth could run at 7% or more as these platforms hit, roughly triple the long run consensus because productivity compounding is the one force Wall Street's models consistently underestimate. You can argue with the number.
The mechanism is harder to argue with. Every time intelligence gets cheaper, the things intelligence was the bottleneck for get cheaper, too. And then they get abundant. And healthcare has been the last fortress of this pattern because biology was the hardest thing to digitize. The human genome project cost about $3 billion and took 13 years.
Today, a genome can be sequenced for under $1,000 in about a day. That's a cost curve that fell faster than Moore's law. But the market never repriced medicine for it because for 20 years, the digitization of biology was a science story, not a product story. Until the mRNA vaccines, the co shots were the proof that you could go from sequence to injection in months.
The first time biology behaved like a software release. What intaran autogene is is the second act. Not one vaccine for everyone but a different vaccine for every person. When manufacturing becomes personalized and design becomes algorithmic, you stop being a drug company in the old sense. You become a compiler.
So underpriced doesn't mean the stock should be higher. It means the category is being priced with the wrong mental model. Wall Street values biotech on revenue pipelines. How many drugs? How many approvals? What quarter? Kathy Wood's argument is that AIdriven biology needs to be priced on cost curves.
What happens to margins and addressable markets when the design loop keeps getting cheaper and the manufacturing keeps getting more flexible? The Madna spike was the market accidentally using the right model for one afternoon then going back to the old one. Three real world use cases. Let's make this concrete because priced on cost curves is abstract until you can see the cases.
Use case one, the pipeline behind the headline. Intismarin isn't a one drug wonder. The same personalized mRNA platform is being aimed at the cancers where the NIH's fasttrack initiative is heading. Pancreatic, liver, colurectal, pediatric tumors. The Interpath Melanoma result was the platform's proof of concept in the most visible trial.
If the manufacturing works for 1,137 bespoke melanoma vaccines, it works for the next indication and the next. That's why the market priced a platform, not a product. Use case two, the design loop itself. Every cycle of AIdriven drug discovery, predicting protein structures, designing binders, screening virtual compounds, running AI red pathology, compresses timelines and costs.
Development that used to take a decade starts to take years. And it's not one trick. AI agents are now running wet lab experiments, discovering new enzyme systems, designing molecules that human chemists wouldn't sketch. Diseases that were uneconomical to research at old costs become viable at new costs.
That's the part Wall Street can't model in a spreadsheet because it's not a bigger revenue line. It's a bigger universe of addressable problems. The trade isn't one drug. It's the expanding radius of what's curable. And notice what this does to the classic biotech risk model. The old model says most drugs fail, so you price in a low hit rate and demand huge margins on the winners.
The AI model says the failure rate itself falls because you simulate and scream before you ever touch a patient. And the failures you do have cost far less. When the cost of an experiment collapses, the portfolio math of an entire industry changes. That's a structural reperic, not a stock pick. Use case three, the economics of prevention flipping.
When genome sequencing heads toward free and multi-cancer detection tests become a volume play, the cheapest healthcare stops being treatment and starts being early detection. Companies and insurers start paying to keep you healthy because catching disease early costs less than treating it late. That's a structural inversion of a sector that eats roughly a fifth of American GDP from sick care to healthare and it's being built on the same AI cost curves.
The mindset shift. So here's the mindset shift this episode is really about. The old investor question is which drug stock wins the quarter? The new question is what happens when intelligence becomes the cheapest input in medicine? Wall Street still asks the first question. The people building the future asks the second.
And that gap between the two questions, that's the underpriced trade, not a ticker, a noticing. The market briefly noticed on August 19th, priced a platform instead of a product for 6 hours, and then went back to sleep. Kathy Wood's argument is that the evidence wave, personalized vaccines, AI designed molecules, ambient early detection, a fasttrack national cancer initiative will keep forcing that noticing over and over until the pricing catches up with the physics.
One more honest caveat before we close this part because optimism has to survive contact with reality. None of this is guaranteed on any timeline. Trials fail. Overall survival data is still pending. Regulators move at regulator speed. The analyst firms that downgraded after the hype weren't crazy. Valuation discipline is real.
The optimistic case isn't that every AI drug works. It's that the cost curve keeps working even when individual trials don't. Because every failed trial feeds the same learning loop. The platform learns, the category compounds. And there's a second order effect worth naming because it's the part that makes this an abundance story and not just a market story.
When therapies get personalized and manufacturing gets flexible, the economics of rare diseases change. There are thousands of conditions that affect too few people to justify the old billiondoll development pipeline. AId designed, mRNA delivered, platform manufactured therapies could make the economics of the long tale work for the first time.
The most underpriced trade in healthcare might turn out to be the diseases nobody bothered to price at all. That's the yes and in full. The market saw a melanoma stock move. Kathy Wood saw a platform repricing. The deeper read is a category inversion. Medicine stops being an industry that discovers drugs and starts being an industry that compiles them.
And compilations get cheaper every year. Part four, actionable optimism signoff. So what do you actually do with this? Three things. None of them buy a stock. One, learn the vocabulary of personalized medicine. Neo antigen, mRNA platform, recurrencefree survival, cost curve. You don't need a biology degree.
You need 30 minutes and curiosity. The people who understood the software shift in the '90s weren't all programmers. They were the ones who could read the trend line. This is the same game, one level deeper. Two, if you work anywhere near healthare, life sciences, or insurance, push AI literacy inside your organization.
Now, the institutions that reorganize around AI designed medicine in the next 5 years will be the ones that look obvious in hindsight. The ones that wait for the spreadsheet proof will be buying the story at the top of the curve. Three, keep receipts on the cost curves, genome sequencing cost, AI drug discovery timelines, early detection test prices.
Every year, write the numbers down. When someone tells you medicine moves slowly, show them the receipts. Exponentials look slow, write up until they don't. And a bonus fourth for the builders listening. Ask where the platform layer is in your own field. Madna's real business isn't the melanoma vaccine.
It's the manufacturing and design platform underneath it. In every industry AI touches, the durable value migrates to the platform layer. Find the platform, understand its cost curve, and you understand where the abundance is going to land next because that's the real trade Wall Street keeps missing.
It keeps pricing the product. The abundance keeps landing in the platform. and the frame to carry out of this episode. The most underpriced trade Wall Street is missing isn't a stock. It's the assumption that medicine changes the way it always has. On August 19th, 2026, for about 6 1/2 hours, the market priced biology like software and 44 billion appeared.
Kathy Wood's bet is simple. Reality is on her side of that pricing. The curbs only run one direction. Not financial advice, just the future arriving on schedule a little earlier than the analysts expected. That's our show. We'll be back with the next thing the exponentials are doing. Moonshots optimist.
See you next time. End of spoken script. Word count target 2,800 to 3,200. Moonshots. Two words used throughout. Moonshots appears only in this header metadata, never in spoken lines.
Comments 0