On May 12, 2026, the day FDA Commissioner Marty Makary’s resignation became official, Elliot Turner and John Mihaljevic sat down with Peter Mantas, general partner and chief investment officer of Logos LP and co-creator of Back of the Napkin Bios, for a deep conversation about what comes next for gene therapy, rare disease, and the broader biotech complex. The full episode of This Week in Intelligent Investing rewards a careful listen. What follows is a brief tour of the ideas we found most useful for long-term investors. The complete edited transcript appears below the article.
A regulatory clean-out, not just a personnel change
Peter frames the Makary departure as one node in a larger removal of what he calls “Arnold Ventures plants” inside the FDA. The group includes Makary, Vinay Prasad, Katherine Szarama, and Tracy Beth Høeg. Arnold Ventures, in his telling, has spent more than a decade funding anti-high-cost-drug campaigns, including the Hillary Clinton-era pursuit of Martin Shkreli. Its philosophical capture of the second-term FDA collided badly with a thinly capitalized rare-disease ecosystem that depends on the accelerated approval framework.
Prasad has already exited CBER. Peter argues the political cost of replacing the departed officials with ideological clones is now prohibitive, given quiet pressure from Susie Wiles, Congressman Rutherford, and major Republican donors, including Miriam Adelson, who funds Rick Scott and has long been a rare disease advocate. The current acting commissioner brings a commercial mindset rather than an Arnold Ventures pedigree, which Peter reads as a constructive early signal.
uniQure as the test case
The centerpiece of Peter’s thesis is uniQure’s AMT-130 for Huntington’s disease, an AAV5 gene therapy that delivers an mRNA-silencing tool directly into the striatum. The three-year data, in his words, is “a home run”: a 75% reduction in symptoms and back-to-back negative NfL readings. uniQure had aligned with the Peter Marks FDA for accelerated approval, but the new FDA pulled the alignment twice, including after a Type A meeting following the home-run data, demanding a randomized Phase 3.
Peter argues that a placebo arm is not only scientifically unnecessary in this case but ethically indefensible, because Huntington’s patients who decline during the trial timeline are then disqualified from receiving the therapy at all. Two days after the second rug-pull, the UK government endorsed approval, an asymmetry that compounds political pressure on Washington.
His back-of-the-envelope math on the global TAM, roughly 30,000 US patients, 8,000 in the UK, 40,000 in the EU, plus pre-symptomatic populations at approximately a $3 million price point, supports a share price “well into the triple digits” on a 61-million-share float, before any credit for the platform’s hemophilia and epilepsy follow-ons that ride the same manufacturing rails. AMT-260, the epilepsy program, addresses a market roughly five to ten times larger than Huntington’s. Centene, notoriously stingy, has already drafted coverage guidance for AMT-130, a powerful tell on the value proposition versus standard-of-care spend.
The picks-and-shovels read-across
Peter’s most actionable observation may be the multiple compression the volatile FDA regime has imposed across the biotech tools complex. ClearPoint Neuro, whose brain-delivery platform AMT-130 requires, has been valued as if its biologic customer base is not durable. Repligen’s filtration and Protein A franchise, Cytiva’s bioprocessing exposure, Stevanato’s glass and packaging, Sartorius, West Pharmaceutical, and the end-to-end “shopping malls for pharma” Danaher and Thermo Fisher are all, in his view, mispriced for what happens when a blockbuster like AMT-130 actually clears.
AI: a real tailwind with a regulated ceiling
Peter is bullish on the AI bubble in biotech, “bigger than 2021,” in his words, but realistic about its limits. AI cuts the cost of pre-clinical knowledge dramatically and will jam the early-stage clinical pipeline with candidates. But Eroom’s law, he argues, is a regulatory construct, not a technological one. Until human trials, reimbursement, and label restrictions can be reshaped, AI’s main effect is to grow the queue rather than shorten it.
The candidates that fully break through are those tied to indications where the alternative is unequivocally worse: pediatric rare disease, ALS, Parkinson’s, Huntington’s. In those settings, the NPV math overwhelmingly favors a one-time curative price tag over decades of standard-of-care spending. For lifestyle indications, gene therapy will be a tougher sell. The system can absorb $3 million one-time pricing when it displaces $7 million to $9 million of cumulative care; it cannot absorb the same price tag for an 80-year-old with limited remaining life expectancy.
Why this episode rewards your time
Peter brings a rare combination of corporate-law background, deep scientific-literature work, procedural fluency on FDA process, and a value investor’s discipline on TAM and probability-weighted outcomes. The full conversation is dense with specific names: Sarah Tabrizi and Ed Wild at University College London, Peter Marks’s prior framework, the supply-chain economics of Cytiva and Repligen, as well as the kind of practical handicapping that helps long-term holders separate signal from noise. We encourage members to listen to the episode and to read the edited transcript that follows.
The primary purpose of this podcast is to educate and inform. The views, information, or opinions expressed by hosts or guests are their own. Neither the show nor any of its content should be construed as investment advice or as a recommendation to buy or sell any particular security. Security-specific information shared on this podcast should not be relied upon as a basis for your own investment decisions. Be sure to do your own research. The podcast hosts and participants may have a position in securities mentioned, personally through sub-accounts, and/or through separate funds, and may change their holdings at any time.
Download the full transcript below.
John Mihaljevic: Welcome everybody to a new episode of This Week in Intelligent Investing. We have a great one ahead. We’re going to be talking biotech, and we’re recording this on May 12th — a big day for biotech and specifically for some of the topics and companies our guest, Peter Mantas, has been writing about. Also here on this program is my co-host, Elliot Turner.
Let me welcome Peter Mantas to the show. He is general partner and chief investment officer at Logos LP and co-creator of the Back of the Napkin Bios podcast and Substack, with a unique background spanning corporate law, international technology, and interdisciplinary value investing. Peter brings a wealth of expertise in capital markets, emerging technologies, and in particular life sciences — an area he’s been very deep on for many years, and that I think will be the focus of our discussion today.
Peter, maybe we can start with the timely topics today, and then later we’ll get a bit deeper into how you came into biotech as more of a generalist. Why don’t you fill us in on what the news is that investors should care about?
Peter Mantas: Today’s a pretty big day because Marty Makary is officially gone from the White House administration. Ironically, it’s the day before he’s supposed to testify in front of Congress. He’s been a polarizing figure, along with the entire FDA. This FDA in Trump’s second term is very unusual — a very different FDA than his first term.
Having him removed is quite bullish for the gene therapy companies, cell therapy companies, and gene editing companies — anything in rare disease, expensive drugs that can provide life-changing therapies. It’s good news overall for American science.
Elliot Turner: For those who may not be as familiar with why this is significant: usually when someone like that resigns, the replacement is pretty much the same. What’s so significant about this particular resignation?
Peter: The FDA, for some odd reason, has been infiltrated with the MAHA movement, with a number of members tied to Arnold Ventures, which has a well-known anti-rare-disease, anti-high-cost-drug philosophy. You could argue it’s almost borderline eugenics. Arnold Ventures is the same organization that funded Hillary Clinton to go after Martin Shkreli. Shkreli is nothing like this guy, and I don’t blame them for that.
But Makary, Prasad, Szarama, and Tracy Beth Høeg — I think that’s how you pronounce her last name — are all effectively Arnold Ventures plants. So Prasad leaving is fantastic news for CBER, which heads biologic drugs and gene therapy. Next to go is Makary, the head of the snake. Eventually the rest will be cleaned out.
You have a changing of the guard at the FDA, and an opportunity — not only for Trump to do the right thing, but for any new FDA head, CBER head, or CDER head to almost provide legacy status. This is the chance to approve one of the biggest gene therapies in the world. This is the chance to approve one of the biggest cell therapies in the world. This is the chance to approve some gene editing platforms coming to market soon.
It’s a very interesting time for the FDA, especially for an administration so focused on staying ahead of China and making sure they don’t get out-innovated by the Chinese.
John: What was the Trump 1.0 FDA like? What are the biggest differences here, and what gives you confidence that it won’t be more of the same from whoever’s picked next?
Peter: Trump 1.0 was — I won’t say Reaganesque, but a little bit more pro-industry. My theory is that the MAHA movement sort of took on its own life form. When you go too far right, eventually you get a bit too far left — it kind of comes full circle. Because of the anti-vaccine sort of MAHA movement, you had a window of opportunity for the Arnold Ventures philosophical narrative to get in there and influence.
That blindsided the administration, and I think it blindsided large donors like Miriam Adelson. Miriam Adelson is a very significant rare disease advocate who funds a lot of Rick Scott. That’s almost polar opposite to the Prasad-Makary Arnold Ventures philosophy. So 1.0 is a bit different than 2.0 — just from the perspective of being more pro-industry.
And 2.0 — it’s a post-COVID FDA, a post-COVID Trump term. They may have been scarred from what happened with vaccines back then, and the voter base had a certain view on that. That left the opening to happen. And now that rock is being removed, you have a shot at an FDA that can come back to its senses.
Ironically, the Arnold Ventures philosophy is a Democratic philosophy, but they didn’t infiltrate Peter Marks and they didn’t infiltrate the Biden administration. So it’s a pretty extreme philosophy that, in some way, aligned with the right — and now that’s being removed. So I think it’ll have a more sensible approach going forward.
Another reason you won’t see more of the same: there’s the fear of, what if the replacements are worse? I don’t see that happening. There’s too much political backlash from the electorate, from major donors, from congressmen, from the base, from the amount of advocacy and patient groups and well-funded organizations that want to burn Prasad and Makary. Quite frankly, I think their reputations are finished. They’ll be lucky to get any job outside of academia.
Someone coming in to repeat those mistakes would be a nail in the coffin. It’s just too politically risky, and I don’t think Trump cares that much about it. My personal sense is Trump might have been lied to — yeah, we’re approving a gene therapy a week, which was not the case. So I think it’ll be a little different this time.
Elliot: One of the cynical takes on this — and lumping RFK into the crew — is that the only reason this group has lasted as long as it has is the upcoming midterms. The RFK voter base has been considered a critical piece of the Trump electoral coalition, which is a little different than Trump 1.0 and prior Republican administrations. So there’s a degree to which you can’t chop the head off before you get the tail and take their votes.
Peter: Yeah, I think that’s right.
Elliot: And you’ve had a front-row seat to one of the foremost controversies that seems to have been part of the undoing of Makary. Maybe you want to tell that story — about uniQure specifically.
Peter: uniQure is a gene therapy that uses an AAV5 virus to deliver an mRNA-silencing tool to effectively silence mutant protein production for Huntington’s disease. It’s about the closest thing we’ve come to a cure for Huntington’s. And by cure — it’s such a difficult genetic problem that the standard is, are you getting worse? That’s really the standard. So it’s as close to becoming a game-changing therapy for that group of people as possible.
They’ve released very strong two-year and three-year data, with four-year data coming out this year. They had aligned previously, under the Peter Marks FDA, for an accelerated approval. The Peter Marks FDA had a reasonable approach: look, this is an area we’re not too familiar with. It is a devastating disease, arguably one of the top three most devastating diseases in humanity. Monkeys don’t get this; really just humans. Tell us what needs to happen here — bring the experts in, tell us how we can get this over the line to save people.
That alignment was then rug-pulled under the current FDA. And it wasn’t just rug-pulled pre — you could get a sense it was going to happen before the three-year data — but it happened after the three-year data, which was a home run. 75% reduction in symptoms, back-to-back negative NfL readings, which are biomarkers for neurodegenerative diseases. This thing is working. And theoretically, it’s probably going to get better as time goes on.
So they said, we don’t align with that old version. We’re going to set up a Type A meeting. The Type A meeting happens, and they pulled them again. You had a double rug-pull by the FDA against uniQure, saying they need to do a Phase 3 trial — which for this population group and these patients is unethical.
Why is it unethical? It’s delivered directly into the brain. It requires minimally invasive brain surgery, but that’s not necessarily the unethical part. Part of it is you’re putting people under anesthesia for placebo, but that’s not even the most unethical piece. The unethical piece is this is a disease where time is of the essence. If you deliver someone a placebo in a certain cohort, time goes on, they get worse, and they no longer qualify for the drug. That’s the problem. That’s what’s truly unethical about it.
You’re effectively walking people off the cliff in the name of science, knowing full well they’re going to die — because the moment they pass that threshold, even if the drug gets approved, they cannot take it.
This is the very reason an accelerated approval framework exists: let us approve this, but approve it in a very contained way. You can only dose a few thousand people; you may have to do another trial, but it won’t be a full Phase 3. We’ll restrict the label, we’ll restrict the marketing, and if anything happens, we’ll pull it off the market. That’s the whole reason for AA when it was invented in 1992 — to allow the government to control drugs as they prove out more scientific efficacy, but to allow people who really need it to have access.
Under this administration or this FDA, they’re not even following proper procedure. They’re not even following what the accelerated approval mechanism is meant to do. I don’t think there has been an accelerated approval for any other drug. So what’s odd about this administration is that you’re not only seeing a complete disregard for scientific curiosity in a disease like Huntington’s, which is very difficult and requires specialist expertise, but a complete disregard for the procedural administrative framework that exists under the FDA and under law.
John: It sounds like this is a US issue — in the UK and elsewhere, uniQure has been gaining traction. Has that also been part of the equation in putting pressure on the FDA? It’s not a good look.
Peter: Oddly enough, when they got rug-pulled the second time after the Type A meeting, two days later the UK said, oh by the way, we’ll approve you. It was effectively sponsored, or you could say endorsed, by the UK government. Part of that reason is the UK has a very high Huntington’s disease population — a large concentration of people live there. It’s also the epicenter of HD research. University College London is known as the preeminent hospital for HD research. I think they’re the ones who discovered the gene.
The leading HD researchers, Ed Wild and Sarah Tabrizi, who were involved with the uniQure trial, are out there. uniQure did their trials under NHS guidance and approval — meaning, NHS said, use our hospitals, use our nurses, use what you need to get this done. So the UK is very open and knows what the drug does. Sarah Tabrizi herself was in shock when she saw the three-year data — just in glee that, oh my God, I think we found something here.
And if the UK is open, the EU will eventually be open, and then Australia is very open. It kind of leads to a domino effect. That creates an immense amount of political pressure on the US, because if you’re supposed to be the gold standard of leading the scientific revolution, and you don’t want to be left behind by China — you can’t even catch up with the UK, you can’t even catch up with the EU. So it puts a lot of pressure on the American government, on American policy, to approve a therapy that effectively works. It’s not like they don’t have the tools to follow the UK, the EU, or Australia. It’s just having the right people in place and the right administration to see it through.
Elliot: uniQure shares have been very responsive to news, good and bad — no wonder. How do you think about valuing a company like that? How should we think about what it’s worth?
Peter: It’s similar to a weighted coin flip, or maybe a highly weighted coin flip. You’re taking a probability estimate of every single jurisdiction and its populace and the supply and demand of people who can qualify for the therapy, and you anchor that to a price.
In the US there are 30,000 people who have HD, and maybe another 100,000 who are pre-symptomatic and at some point might meet the threshold to qualify for the therapy. That’s going to cost $3 million. That’s your effective TAM. In the UK there are about 8,000 people with the disease, about 15,000 pre-symptomatic; might cost around $2 million, maybe $2.5 million. The EU is bigger than the US: 40,000 people, maybe 150,000 pre-symptomatic. Australia’s got a large population as well — I think it’s 7,000. Japan has 8,000 who qualify.
If you do these TAM numbers for an effective monopoly at $2 to $3 million, you’re talking about large revenue numbers. Discount it however you want — discount dilution, discount that you may not be able to serve everybody, discount time, go ahead. You get a share price well into the triple digits from here on a share float of 61 million shares. So it can certainly be torquey.
The other piece is that AMT-130 rides the same rails as their hemophilia drug, so manufacturing doesn’t need much of an application there. It’s right in the same rails. They’ve seen that before; Australia’s seen that before. Their epilepsy drug rides the same rails. So it’s more validating of their platform, and platform companies typically get a higher multiple on their valuation just because R&D productivity is far higher. If you’re riding the same manufacturing rails and FDA applications become far easier, you’re cutting time off every single therapy that goes through the channel. Every subsequent therapy becomes more valuable.
AMT-260, their epilepsy drug, has a market I think five to ten times larger than HD. Same exact technology, same exact procedure, same manufacturing rails, with data coming out as well. So if you factor all that together, you have a fairly valuable company. And it wasn’t too long ago, right after the Type A, that the company was trading below cash. They have $8 to $9 of cash. You had a few other platform drugs, like in Fabry disease. It can get pretty violent.
But AMT-130 is the big — I call it the big kahuna — because if you don’t approve this for this biology or this mechanism for this group of people, it becomes hard to approve other things.
John: Maybe talk a little about what you’d look for from here to gain conviction that, yes, this change at the FDA is what you’ve been looking for, and there’s going to be a path forward.
Peter: The current replacement, Demands, is a good start. He’s a hunting buddy, I think, of Trump Jr. He’s a lawyer whose previous experience was helping food companies comply with the FDA — so he has a commercial mindset already. We’ll have to see; I don’t think he’s going to be permanent, so we’ll have to see who they appoint.
It is positive that he’s coming from HHS and not from the internal FDA, so you don’t have that Arnold Ventures influence — I don’t think, anyway — especially with someone coming from a commercial background helping companies get over the line with the FDA. You want to see who they choose to replace him and become FDA head. He may still help in selecting the CBER head, and the CBER decision will be important.
I get the sense that Trump is aware of the issue and needs to clean house. I think it was Congressman Rutherford — I forgot the other guy’s name — they’re very close with Susie Wiles, and they sent a letter saying, hey, this is ridiculous, you need to approve AMT-130. Part of that was through the patient advocates who really pushed for this to get over the line. So I’d be very surprised if the next group of people coming in are as bad as Makary and Prasad. Having them removed is a big benefit to the company.
And not just to uniQure, but to anybody in the rare disease gene therapy space who’s had difficulty aligning with the FDA — mainly because it’s now been made aware by Susie Wiles, who is going through her own health issues. Congressmen are fully aware, MAGA congressmen are fully aware, people like Senator Johnson are fully aware, Ted Cruz is fully aware. So there’s an inclination that we might make it over the line. We’ll be okay.
Elliot: Talk a little more about some of these knock-on effects. I know there’s a tools company you follow that’s directly involved in the uniQure therapy. And how has this impacted the entire biotech landscape — what has it meant for the sector at large, and for the tools companies that service the sector? How much of what has looked like a really bad time in biotech is the consequence of a bubble that popped post-2021, versus the chaos wrought by a very volatile regulatory regime?
Peter: People might disagree, but I think it’s brought on compression across the entire sector, from small companies to large. The most obvious example is ClearPoint Neuro, which delivers the uniQure drug to the brain. They’re a drug delivery platform.
Companies like that are typically valued based on the duration and durability of their biologics business. Similar to bioprocessing: if you’re going to grow 10 to 15% a year, because there’s a stable amount of drugs coming through your pipeline and you’re in that downstream delivery mechanism — as close as you can be to the human body, or in the cleaning of drugs, the filtration of drugs — same thing, you’re tied to the supply chain of manufacturing and delivering therapeutics. You get a certain multiple in that because the thought is, this is a durable part of the market, a durable part of the ecosystem. Aging is secular, therapies will keep coming out, these things should command double-digit enterprise value to sales.
But if you have drugs like uniQure, which work biologically and should have been approved under accelerated approval, then companies like ClearPoint — their entire book of biologic customers gets compressed. You shouldn’t value it as a bioprocessing or biologics business; you should value it as a faster-growing Medtronic. Instead of 12 to 15 times enterprise value to sales, you know, five times sales.
If you get something like uniQure approved — again, a drug that has shown patients getting better, back-to-back negative NfL readings, the biomarker in neurodegeneration — ClearPoint was designed, created, made for uniQure. You’re going right into the striatum of the brain; no pill, no IV, no injection can get there. Not that we know of anyway. Maybe a nanorobot in 30 years, but that has a whole pile of issues.
So the multiple gets compressed for ClearPoint. It also gets compressed for companies like Repligen, which sells a lot of filtration and a lot of Protein A pegs and chromatography columns for gene therapy and cell therapy. If you have a blockbuster, multi-multi-multi-billion-dollar therapy like AMT-130 that biologically works and doesn’t get over the line, then maybe the Repligen end markets aren’t as durable as we thought. Maybe Cytiva’s end market does carry some regulatory risk that impacts the timing of its cash flows, and that creates a problem.
The second piece is that these blockbuster drugs fuel the biotech funding ecosystem — they fuel the VCs, ignite R&D diagnostics, ignite the early-stage discovery upstream part of the business. Eventually that’s what feeds companies like Danaher and Thermo, which are end-to-end. They need everything in biotech working — early-stage diagnostics, R&D, the tools, bioprocessing — everything working.
Drugs like AMT-130 are the fuse that can ignite the winds necessary to ignite the biotech sector even further. That’s what a lot of sector specialists or generalists in healthcare are missing: you need these wins. You need a gold medal winner at the end of the race. It may not be the fastest time, but you need a winner so that people can get confidence in the sector, so that the multiples are accurate for companies that are supposed to be compounders, so biotech funds can continue to fund the drugs they want to focus on to compete against China. All these things get impacted if we don’t have blockbuster drugs getting approved.
And for the population: this disease impacts people in their prime working-age years. You get it at 30, and by 40 you’re in a wheelchair. You’re being removed from the tax base. If you’re an executive at Nike making $300K a year, or at any S&P 500 company, and you get Huntington’s, you can’t work anymore — and your wife has to take care of you. So you go from upper middle class to lower middle class in a generation. It becomes very anchoring to the tax base.
It’s not good for the populace; it’s not good for the health insurance companies. Health insurance companies want people working so they can keep paying more health premiums. So it impacts not only the VC biotech ecosystem and the tools and life science companies, but also the healthcare system and the insurance companies. It impacts everybody. It’s a pretty full-scale impact on healthcare if we don’t get a lot of these drugs across the line.
John: Peter, can we talk about the impact of AI adoption in biotech? I believe you wrote about ClearPoint Neuro being a winner due to its delivery mechanism into the brain — which, as you say, is also used for delivering the uniQure drug. I’m wondering whether that’s analogous to a company like Sartorius, that’s also more in the picks and shovels of biotech. Do you believe those are the biggest beneficiaries as we get more candidates into the pipeline?
Peter: Call them the compounder sleeve — what are the most durable, secular, highest-quality companies that can compound on the backs of focusing on science and getting drugs through? It’s the Sartoriuses, Repligens, ClearPoints, Stevanatos, and Wests of the world. Those are the kinds of companies that in a biotech bull market perform fairly well.
The last ones to move are Danaher and Thermo, because they’re end-to-end. They’re like your shopping mall for pharma. You need research, you need diagnostics, you need bioprocessing, and biotech all on fire at the same time — which usually doesn’t happen unless you have policy and low rates simultaneously.
The first step is getting some of these blockbusters over the line — uniQure, AMT-130, and some other cell therapies focused on cancers or other indications with large TAMs like ALS or Alzheimer’s. These are big problems that companies are spending billions of dollars on to solve, and to throw a hiccup in their way for quite frankly dubious reasons is not good for anyone. It’s not good for the ecosystem, and it’s not good for American health.
So if you get some of these over the line, it certainly helps not just the drug companies and small companies like ClearPoint, but eventually bleeds up to the Repligens, the Twist Biosciences, and the Stevanatos of the world. Do you know how many Repligen filters you’re going to need if AMT-130 gets approved and they’re in year five serving, I don’t know, 2,000, 3,000, 4,000 patients? It’s a lot — a lot of cleaning, a lot of AAV5s. So it’s certainly beneficial for that ecosystem as these drugs get approved over time.
John: Do you think we’re going to see an explosion in drug candidates? How bullish are you on AI actually being effective in discovering drugs and curing diseases that haven’t been cured in the past? Is that realistic, and what kind of timeline do you have in mind?
Peter: If you want my hot take: I think you’re going to see the biggest biotech bubble we’ve ever seen. I think it’ll make 2021 look like child’s play. The big reason is that you need winners in the space, like the AMT-130s of the world becoming poster children. Every bubble has a poster child, from Sarepta to Gilead to Regeneron. They’re all products of bubbles that eventually became platform companies and compounded and reinvested their cash flows to expand their platforms. That’s all it is.
AI is the only sector where it’s a true secular tailwind. If you look at it from discovery, or even supplementing discovery, or supplementing clinical trials — certain things may not be completely replaced, but AI is definitely an accelerator of knowledge. What does AI do? It brings the cost of knowledge to zero. And if you’re bringing the cost of knowledge to zero, the number one beneficiary is areas where the cost of knowledge is an input — which is biotech.
So you’re going to see a lot more drugs go from preclinical to Phase 1. That will jam up the pipeline for early-stage clinical companies. But that will require money, because you have to go through the regulatory hurdles, and money will only flow where it’s respected — and it’s only respected if there are approvals. So AI is sort of a coiled spring, but in order for the fuse to be lit, you need some of these drugs approved. You can have as much AI as you want, but it’s a regulated industry, government-controlled, that creates many monopolies. AI cannot get around that.
So I think AI is certainly a tailwind in the whole space and will cut down time to discovery. But at the end of the day, there’s only so much you can do, because you still have to go through reimbursement, you still have to face insurance companies the same way, you still have to be tested on humans. So I don’t see that changing from AI anytime soon.
Elliot: That’s one of those really interesting points — you said the part about don’t have to be tested on humans, and we go back to the FDA portion of this conversation. You look at what Prasad and Makary had said coming in: there were some really promising programs to get rid of human testing in the earliest clinical stages — going in silico, skipping the animal testing and going in silico, getting to humans faster — and a voucher program that would create accelerated pathways for some of these new therapies. There was reason for enthusiasm. And then, as time went on, you saw chaos compounding and statements that made it very clear the FDA went backward on them, punishing capital that moved on those early promises.
It’s interesting: the leading minds of AI — Dario today, Jensen Huang, Demis Hassabis — have all said that biotech is the most obvious place where AI will have a profound impact. I’ve had this theory that between CRISPR, genomics, and AI, you could finally break Eroom’s law. For those listening, Eroom’s law is literally Moore’s law spelled backwards. It was coined by a Bernstein analyst in 2015, and it’s the notion that for incremental innovation, biotech gets proportionately more expensive to find the next drug or therapy — the opposite of Moore’s law. So do you think we could break Eroom’s law, or is it a construct of the regulatory edifice, and it’s going to be hard to break even with these innovations?
Peter: It’ll be hard to break. You could use AI all you want to help discover any platform drug possible, but if the government’s going to force you into a randomized controlled trial for a patient population that doesn’t need it, or can’t access it, or where it’s unethical to do so, you’re lighting money on fire.
At the end of the day, it’s a highly regulated industry that has guardrails — for good or for ill. The guardrails are necessary. And by the way, I’m not anti-Phase 3 trials, I just want to put that out there. But it is a regulated industry that serves to ensure the safety of a population of humans. Because you’re dealing with safety and deaths, and you can’t afford to make a mistake, those regulations are going to continue. It’s going to cost more and more every year to bring a drug to market. That’s good and bad — it’s part of the moat here. It’s why these companies get mini monopolies over a period of time.
Right now, AI is useful for pre-discovery, early early early-stage, preclinical to preclinical kind of work, where you might not need to test fully on a rat model — maybe you supplement with AI. Maybe you’re doing a clinical trial on 10 patients, but with AI you’re only going to do seven. Something like that is possible. I don’t think you’ll fully replace the clinical trial model with AI models — it’s too dangerous.
For certain diseases there might be more stringent regulations, like: because you brought something very quickly to Phase 1, we want to make sure it’s really good and really strong, so your Phase 2 will be a 300-patient Phase 2, a 250-patient Phase 2. Maybe even though you qualify for AA, we’re still going to ask you to do a randomized controlled trial just because, for this population, there is a standard of care but they’re not really dying; it’s not totally devastating to the population, and because you came to clinic really quickly, we’re still going to ask you to do a small Phase 3. So net-net, I don’t think AI does much other than accelerate the early-stage pre-discovery stuff.
Elliot: Is there enough capacity in the system — from capital at the top all the way down through the companies that run the clinical trials for biotechs — to get more molecules into the clinic, push things through, and achieve what you’d need for that bubble bigger than 2021 to take shape?
Peter: You can get more molecules into the clinic for sure. The question is, how fast can those molecules go through the clinic? That’s the question I don’t know — and it’s really a policy-driven question, a question for the regulators. If you had a Peter Marks-level FDA, it would probably go relatively fast. The previous FDA with Makary and Prasad, yeah, no chance.
We had the fewest biologics approvals in 2018. Do we have more biologics in the queue today than in 2018? Yes. So at the end of the day, as much as AI is a tailwind, the output is subject to government regulation. If our regulators and policymakers don’t address that problem, then AI just creates a queue. That’s my opinion.
Elliot: AI creates a queue. How do you think it has its most profound impact? Is it on genomics — identifying the right targets? Molecule design? Patient population selection? There are so many angles. Is it all of the above? When, where, and how does it start leading to progress?
Peter: It starts with molecular design and genomics, and eventually patient population selection. I think it eventually gets to all of that. Right now, it’s probably most useful for discovering new drugs and molecular design.
Genomics analysis is a tailwind for the gene editing and gene therapy companies. You can test out which AAV is best to deliver XYZ therapy. Will this gene editing platform slice and dice the RNA/DNA needed to make XYZ ABC changes? That’s where it’s useful right now. But those are all preclinical — before being tested on humans.
At the end of the day, you still face the problems any drug company faces: you’ve got to go through Phase 1, Phase 2, you might get an AA, a breakthrough designation, an orphan designation — but again, that’s all government. The only way you break Eroom’s law is changes in regulation, not technological disruption.
Elliot: Let me ask you this. One of the debates I’ve been intrigued by with CRISPR — very recently we had the first Phase 3 data for an in vivo gene therapy, a gene edit — and the debate even before approval very quickly turned to: okay, well, now what? What do the commercial prospects look like for this drug? Will people want to take a genetic medicine when there are combinations of small and large molecules on the market that, though less effective, don’t change their genetic substrate? How do you think about that in approaching the space right now?
Peter: It depends on the indication. If the problem is a very difficult one that has been a sink of money, that doesn’t have a standard of care, that is effectively a death sentence — I think people are more inclined to try the gene editing of the world, or more innovative therapies like cell therapies or gene therapies. So if you have rare diseases, especially pediatric rare diseases, or humanity-changing diseases like Alzheimer’s, ALS, Parkinson’s, Huntington’s — people are more inclined to take those kinds of therapies.
If you’re talking about, well, I want to lose 10 pounds, or I want to lower my blood pressure a little bit, gene therapy or gene editing will be a tougher sell. So the standard of care and the type of indication you’re in are more the factors as to whether you’d be inclined to take those platforms.
That doesn’t mean those platforms can’t solve big problems. If you have a diabetic and there’s a gene editing solution that cures their diabetes — sure, that’s game-changing, and I think people would be interested in it. But for other diseases or modalities, it becomes a harder sell to have those sort of invasive platforms that really change a lot of your body.
Elliot: These gene therapies are fundamentally curative, whereas in typical biotech fashion, with biologics, you tend to be on them for life once you start, in many cases for some of these diseases. But with a gene therapy, you take it once and you no longer have recurring revenue for the duration of the patient’s life. Do you think our healthcare system can handle pricing that’s pricing to value — giving something NPV-positive to the health insurers — or do you think there’s going to be significant pushback at how big the price tags are, how much the biotechs ask for to justify their investment over time, and how small some of these patient populations are?
Peter: It depends on the disease and the patient population. Huntington’s is a good example. You get Huntington’s at a fairly young age — at 30, let’s say — it takes away your entire prime working-age life, and it lasts for 25 years. So for sure, insurers are going to pay upwards of $7 million to $9 million per patient for that patient’s life. That money goes to physical therapists, wheelchair ramps, a nurse in your home 24/7, a hospital bed in the house, maybe some pills to help with motor skills, but those don’t do anything other than — you don’t shake as much.
Whereas uniQure’s gene therapy is $3 million. You pay $3 million one time, and that person goes back to work, starts paying United Health insurance, is able to provide for a family, is a net positive to the tax base — on top of being cheaper than the standard of care. Or do you want to keep paying $7 million over 20 years? That’s a fairly simple calculation.
Where it gets harder would be, you have a $3 million or $5 million Parkinson’s cell therapy where the patient might be 80 years old and has five years left, 10 years left. Do you want to pay $5 million for a million 80-year-old Parkinson’s patients who are retired? They’re not paying that much more in health insurance. Their life expectancy isn’t very long anyway. So that becomes a different calculation.
For rare diseases, again, it depends. If a child who’s six is given a gene therapy that extends their life for 50, 60, 70 years — yeah, that is a net benefit to the health insurer as well. So it depends on the modality, the disease, the severity, and a whole bunch of factors that insurance companies are willing to pay for.
I’ll go back to Huntington’s. Centene has draft guidance already approving AMT-130. They’re ready to go — and Centene is notorious for pushing back on everything. AMT-130 has not been approved yet, and they’re like, we’re ready to cover this, because they did the math and saw this is a net benefit to the patients they cover. It costs upwards of $7 million for 25 years for that patient versus a one-time $3 million drug. That person then goes back on the healthcare system, goes back to work, and starts paying insurance again. So that is a very clear NPV. But for others, you’ll see more pushback, depending again on the modality, the patient population, the age range, and a whole bunch of variables they have to look at.
John: Peter, maybe we can step back. You have a very interesting path in how you came to be so deeply engaged in the biotech space. You started out as a generalist — law degree, I believe. Tell us about how you ended up so focused and in the weeds in biotech.
Peter: Around 2018, in a discussion with my partner, we said, look, we should really know — you’ve got commodities or biotech. In my previous life, I did a rotation in IP and was around defending drug companies in patent infringement cases. I’m not a patent lawyer, but I got a taste of it.
I just started diving super deep into the scientific literature — this is before I transitioned my entire book into that. I got deep into the scientific literature, deep into the legal procedure, the IP patents. I got deep into understanding how the FDA works from a procedural administrative perspective. And it just fit my curiosity and my strengths fairly well. It fits my analytical skills well.
Since 2018, the big focus for me has been the life science biotech sector. AI has also advanced my knowledge — there are thousands of drug companies, and I don’t know every single one. I’ll use AI as my pocket biotech analyst: hey, this stock is down, why, what happened? Teach me up on this particular molecule. The result didn’t hit a response rate, whatever.
I also think AI is going to accelerate generalists starting to look at biotech, because there can be good businesses in biotech. Like I said, there are platform companies — Vertex, Gilead, Regeneron. AI now gives you a biotech analyst in your pocket. That wasn’t the case when I first started.
It’s the only sector I can find that meets the academic requirements with the capital markets mindset, which makes it very interesting. There’s not a lot of attention other than from a select few who can understand how some of these things work. So you have a little bit of an edge. You’re not competing against generalists who are in Mag 7 or compounder bro stocks or quality rock stocks. You take some of those frameworks and apply them, but because it’s really off the beaten path, uncorrelated, asymmetric — and then you tie in the academic piece and the variety of drugs and solutions and problems — it just became natural. I was attracted to it.
John: Tell us a little more about your Substack, Back of the Napkin Bios. What do you focus on there? What’s the publishing cadence, and what can folks get if they sign up?
Peter: There was a point in May of 2025, I believe, where biotech was just in a really bad place. The XBI was maybe 70 bucks, 65 bucks. I started seeing so much biotech at such ridiculous prices. I’m like, this is getting out of hand. So I’m going to write a blog — or, call it a signal-based research publication, whatever you want — where I’m just talking about ideas that I think are interesting.
At the time it was a lot of what I call biotech beta, where you could literally throw a dart at some of the major covered ones — Arrowhead, Ionis, Cytokinetics, and so on — and do really well. Now it’s a little harder, because biotech has run a little more than the rest of the group versus life science tools or other areas of healthcare. Now it’s a bit more alpha versus beta.
It’s also a place for me to house my ideas. I actually often go back to my own publication and ask, what did I write about on this date? Oh yeah, okay. So it allows me to go back and remind myself of the original thesis and keep track of notes I published before. It’s a bit of both — note tracking, helping the community where I can talk about ideas I like, and coming at a time when I thought valuations were silly and thought it could be helpful to people interested in biotech.
John: We’ll certainly keep watching what happens next with uniQure, ClearPoint Neuro, and the other companies you’ve talked and written about. Maybe just lastly, for folks who are on X — what’s your X username?
Peter: It’s Peter underscore Mantas, M-A-N-T-A-S. It’s my name.
John: I follow you, and I know it’s a great real-time source of what you’re thinking and often timely ideas as well. Thank you so much, Peter, for joining us today. This has been a fascinating conversation, and I’m sure as AI ramps up — and maybe we get another administration, who knows — there will be a lot to talk about and get updated on.
I also look forward to having you as a speaker at our upcoming online conference in June, the Wide-Moat Investing Summit hosted by MOI, where we’ll go deep on a specific idea. Thanks also to my co-host Elliot, and to everybody listening. We’ll catch up with you next time. Take care for now.
Featured Events
The Zurich Project 2026 (FULLY BOOKED) (Jun. 2-4, 2026)
Latticework 2026, Chicago, Illinois (Nov. 10-11, 2026)
Ideaweek 2027 (FULLY BOOKED), St. Moritz (Feb. 1-4, 2027)
Enjoying Latticework? Help us make it even more special.
Share Latticework (simply click the above button!)
Introduce us to a thoughtful speaker or podcast guest
Be considered for an interview or idea presentation
Volunteer to host a small group dinner in your city
Become a sponsor of Latticework / MOI Global
Volunteer by reaching out directly to John (john@moiglobal.com).
