저자: BradGerstner&CathieWood클립 · 외국인
요약 보기
Cathie Wood는 AI와 멀티오믹스로 신약 개발비가 4년간 약 2.4B달러에서 7억달러로 70% 하락하며 희귀질환 치료제가 경제성을 가진다고 주장. Gerstner는 AI 인프라가 클라우드와 같은 도입곡선을 따르되 자본투입은 10배 크고 3배 빠르다고 보고, 연준 금리인상 사이클이 끝나가 내년 금리인하를 예상. 우주 데이터센터와 스페이스X의 재사용 로켓 우위도 언급.
다듬은 글
I've got three clips for you today from Kathy Wood and Brad Gersonner. Kathy breaks down why drug discovery costs are about to collapse and Brad Gersonner explains why we're at the biggest AI super cycle in decades. And keep watching for the bonus clip at the end on SpaceX. These are some clips I think you will enjoy.
We think this is going to be one of the greatest opportunities for the entrepreneurial spirit in history. uh especially because you can go to chat GBT or Grock and and say you know I think there's a huge unmet need for this that or the other product or service think about it and I would like to start that business boy when we started arc there I would have loved to have a starting point like that and you can work with uh chatbt uh x ai rock uh to build your business.
Uh try it. If you're looking for a job, uh figure out what's that unmet need. Uh I'm going to try and start that business and uh believe me, starting a business very exciting, exhilarating actually, uh but full of risk. Remember, 90% of all startups fail. So just go in soberly that way. make sure you're addressing a real unmet need and then harness AI, harness blockchain technology, harness any of these new technologies.
Now when we get to multiomics normally I start glazing over but uh I think multiomics which is life sciences and really the convergence among sequencing technologies uh artificial intelligence and new technologies like crisper gene editing. What are we doing? We're diagnosing cancer in stage one or before stage one.
Polyps also shed into the blood. And this is just with blood tests. We're collapsing the cost of drug discovery and development. We think it could drop from 2.4 billion for one drug uh to roughly 700 million over the next four years. And that of course includes the failures along the way. Uh, and then most provocatively and and Sheay uh has done an amazing job of documenting uh what a cure is going to be worth.
And I have to give Brett Winton uh incredible credit as he partnered with Sheay uh on this project to uh try and figure out how much is a cure worth. Uh and we think while the cure for HAE, which is a very rare disease, uh would be worth [snorts] $11 million. One and done. You get this treatment and you're done.
Uh that's Intellia's um trial right now and it's in vivo. You don't have to go through any excruciating preconditioning. um it could be worth 11 million. That's how much it would save the system. Uh but they'll probably be able to charge 3 million. And believe me, insurance companies, they will have no problem with that.
Uh so take a look at that section. It's it's fascinating. It's the best one we've ever done. Kathy says AI is collapsing drug discovery costs by 70% over the next four years and that fundamentally changes which treatments become economically viable. So the 2.4 billion number Kathy mentions that's the current cost to get a single drug to market including all the failures along the way.
Most drugs never make it. You test thousands of compounds, run years of trials and maybe one works. That's why drugs are so expensive. The economics only work if you can charge enough to cover all the failures. What AI changes is the failure rate. If you can simulate molecular interactions computationally instead of running physical experiments, you eliminate most of the deadends before you spend money on them.
Kathy's thesis is that this drops the cost to around $700 million. Still a lot, but it's a completely different business model. So, the investor's angle is when drug development costs drop 70%, suddenly treatments for rare diseases becomes profitable. Kathy talks about H AE, this rare genetic condition where a cure would save the health care system $11 million per patient.
At current drug development costs, nobody builds that drug. At $700 million total cost, the math starts working. Insurance companies will pay a couple million for a one-time cure that saves them $1 million in lifetime care. DNA sequencing getting cheap, AI getting powerful, all of this is happening at the same time, which creates new economics of medicine.
early cancer detection through blood tests, gene therapies that actually cure diseases instead of managing them. Biotech's been beaten up for years. But if this thesis plays out, the companies that figure out AIdriven drug discovery first capture an enormous amount of value. And if you want to understand how AI is reshaping industries like biotech, I put together a course called the AI landscape that breaks down all of these things step by step.
Links in the description. If you pay today, you'll get lifetime access and the price goes up as we add new modules. So that's the biotech side. Now let's zoom out to the macro picture. Brad Gersonner was early to cloud when everyone thought it was too expensive and too risky. Now he's calling AI infrastructure the biggest super cycle he's seen in his career.
And his patent recognition has been pretty reliable. By 2008 2009 we had this thing called the iPhone. And we said wow something's really changing here. Um that's no longer going to be search first. That'll be you know something else. And we saw these icons evolving on on page one. And so how do you become an application?
How do you truly leverage? So companies like Uber and Airbnb, companies that were taking advantage of the social graph like Facebook that became applications and were no longer paying the toller, the search engine, in order to gain access. And then I think we were early to the cloud. Um you know people forget but in 2011 2012 the cloud was not viewed favorably and the reason for that is it was more costly it was less performant and it was viewed as less secure than running compute in your own data center.
Yeah, we had a different opinion there. And so our fund, you know, we run, as you know, a long short hedge fund and a long only hedge uh fund. Um, and those are permanent capital. And now we're on our seventh venture fund. But back then when I launched our first venture fund, um, it was almost all dedicated to the cloud.
So early investments in Snowflake and Twilio, and HubSpot uh, really powered the return. We were in the same wave. It took a long time to get the big institutions to actually trust the cloud, but it was clear they were going to and that's that's that's where it is now. We're close to the end of what has been the most parabolic rate hike cycle in the last 50 years in this country that's been hugely destabilizing to financial markets.
It's biggest bond route in 150 years. I think let me be clear on this. I believe the Fed made a huge mistake in the summer of 2021. I think it was knowable then. I think any of the us paying attention were tweeting about it. The cost of a container from China to the US had gone up 20x. Okay. The idea that we didn't have an inflation problem in the summer of 2021 that would warrant just moving off of zero.
Yeah. We had Neil Qashqari, Fed president, um, you know, Minnesota who said at the time there is no chance we raise rates until at least the end of 2023. He was only off by the most parabolic rate hike in the history of the country. Right? So like that to me was nine months that we wasted. We got behind the curve on inflation and once we got behind the curve, there's positive reflexivity to inflation.
So what I mean by that is if your competitor raises their salary, they're going to pay somebody, right? Then you have to raise yours and then they raise theirs, right? That is this uh wage uh price spiral that we talk about. We got behind on that. Now I think we're nearing the front end of that. So we're at 5%.
Could we go to 5.25? Could we go to 5 1/2 on the 10 years? Sure. But think about the convexivity from 0 to five. So that's behind us, right? If you look at the futures curve, it's it'll tell you there's an 84% chance of a rate cut by June. And the Fed itself thinks we'll have two or three rate cuts next year, you know, probably in the back half.
So as investors, you and I shouldn't hyp, you know, spend too much time focused on could it go to 5 and a half or whatever. Like it's not a big move. I look at this and say we're on the front end of a rate cutting cycle. Okay? So think about 2000, the year 2000, May of 2000, the the markets were rolling over, but the Fed again raised rates 50 basis points, its last hike.
By January, 6 months later, they were cutting rates and they would go on to cut rates by almost 500 basis points. So, you're not worried about like some kind of fiscal insanity where they spend so much more money and somehow cause more inflation in the next couple years. You think you think it's pretty naturally just going to go over into a rate cut?
Well, listen, there has been fiscal insanity. There has been. We have to reel it in. um the stimulus package that we did right at the start of COVID, right, when we didn't know whether this was Ebola or whether this was a common like that one is justifiable. The second stimulus package and then the push for the third made no sense to me, right?
Threw a lot of gasoline on the inflation inflation fire. I don't see any appetite in Congress. I don't see any appetite among the Republican contenders for the presidency for any more fiscal stimulus. So, I actually think, you know, call me uh chuck me up to being a little bit more optimistic. I think we're going to see more austerity on the fiscal side out of Washington.
Obviously, the monetary side of the house is already highly restrictive. We're doing quantitative tightening. We've got over 2% real rates in this country, the highest since uh you know uh in over 20 years. And so, I think we're set up for inflation to come down as the Fed expects. We're we're we're nearing the top of tenure, if not at the top.
We're going to see rate cuts next year. And importantly, we're at the beginning of what may be the most important super cycle, accelerant to our economy in a long time. The AI infrastructure buildout is following the exact same adoption curve as cloud, except the capital deployment is 10 times larger and happening three times faster.
So Brad's cloud thesis in 2011 is worth understanding. Back then, enterprise IT departments thought cloud was more expensive, less secure, and slower than running your own servers. All three of those things were actually true at the same time. But Brad bet the economics would flip as the scale improved, and they did.
He made a fortune because he recognized that enterprise adoption was inevitable even when the early product was worse. He's seeing the same pattern with AI infrastructure now. Companies are spending billions of GPU clusters before they even know exactly what they're building with them. That sounds irrational until you realize the same thing happened with cloud.
Enterprises bought AWS capacity they didn't yet need because they knew they needed it eventually and waiting meant falling behind. AIR compute is the same dynamic just compressed into a shorter timeline. But every single GPU being bought is being used now. The numbers are staggering. Microsoft, Google, Amazon, Meta, they're each spending billions and billions a year on capex.
Most of it AI related. Nvidia can't manufacture chips fast enough to meet demand and data center construction is backlogged for years. The investor implication is that we're still probably pretty early. Cloud took a decade to become consensus. AI infrastructure spending is maybe 2 years into what could be a 10-year buildout.
In fact, it may just continue building out forever. The companies supplying picks and shovels, chips, networking, power, cooling have visibility into demand that extends years into the future. Brad's claim is when you see enterprise adoption curves like this, you want exposure before it becomes obvious to everyone else.
Now, for the bonus clip, Kathy also talked about something even more speculative. Data centers in space and why SpaceX's 10-year lead on reusable rockets makes that possible. Before we play the clip, if you enjoyed this video, please consider subscribing. Cheers. And then we get into reusable rockets.
Uh, and it is astonishing. SpaceX 10 years ahead of any competitor. It landed its first reusable rocket in 2015 and now it owns 2thirds of the satellite market and something like 85% of all of the upmass uh put into orbit. Uh and of course now uh we've discovered a new use case and it couldn't happen without reusable rockets uh data centers in space and this is a fairly new introduction to you know our line of thinking and of course inspired by Elon himself.
Uh so take a look at uh at our reusable rocket uh section and be inspired by that whole new world that's being developed. Many people when they're talking about automation and AI and unemployment. Uh one of the things I say is wait a minute there are two new worlds evolving here. Uh outer space a completely new world.
All kinds of jobs are going to be associated with that. Uh and then in the digital realm now thanks to blockchain technology we are seeing immutable property rights in the digital realm for the first time. The the best way to lift people and countries out of poverty is with immutable private property rights.
and that's what we have and it's just starting. Uh, so we're pretty excited about that. One of our clients started with zero audience. Now they're doing $100,000 months thanks to YouTube. And they're not alone. We've helped three businesses hit that level just by growing them a YouTube channel. Want to see how this could work for your business? Book a call with me below.
Comments 0