Hello. It's great to see all of you. You know, once a year I get the benefit of thanking all of you for the incredible work that you do. As you know, we pioneered a new way of doing computing called accelerated computing. And we've been working on this for now 33 years. The big idea is that if you could co-design, meaning if you understood the problems you're trying to compute for, and you developed the entire stack, meaning the algorithms, the computers, and the processors within it, that if you co-designed, you solved the problem from top down, you solved the problem from bottom up and inside out, that you could accelerate these applications which are incredibly computing intensive by a level that Moore's law would only dream of. Moore's law runs at about 10 times every 5 years, 100 times every 10 years. In the last 10 years, we've accelerated AI about a million times. A million times versus a hundred times. A million times compounded over 20 years gives us an opportunity to maybe address some of the most incredible, most impactful challenges of humanity. That's you. And so we've been thinking about accelerated computing and this co-design idea for a very long time.
Well, after some 30 years or so, each different field of computing domains became within reach. One of the first ones was computer graphics. And somebody told me about 10 years ago that it would take 30 years for us to solve this incredibly hard problem called ray tracing for computer graphics. Literally following every photon of light as it bounces around the room, eventually reaching your eyes. And using ray tracing, you could solve for really interesting computer graphics problems and create photorealistic images like this. Well, that was 10 years ago that it would take 30 years. It was the computer graphics problem that was always the future. And literally today, we ray trace every single pixel in a computer graphics program in our video games. And not only do we ray trace it incredibly beautifully, we do it at several hundred frames a second, as large of a resolution as you like. And we've taught an AI how to, we've taught our computer how to simulate with first principal simulations very few pixels. And then we use AI to predict the rest of it. We teach an AI what an image should look like and it generates the image for us. It's temporal consistent. It's spatially consistent. It's incredibly beautiful. It looks even better than we compute.
Well, each one of these different domains are becoming within reach. About 10 years ago, we started talking about self-driving cars. Today, I think we are at the ChatGPT moment of self-driving cars. There's no question in my mind that within the next 10 years, you know, almost every car that gets manufactured will have, of course, either robo taxi capability or if you like to drive yourself, you still can, but it has the ability to be a robo taxi. That future is just around the corner. Each one of these domains are within reach. Now, of course, we now have the ChatGPT moment. The last year we've seen incredible progress in generative AI in large language models. These ideas made it possible for us to take an AI that we thought was magical but hallucinated a great deal. It was essentially cute and very interesting to something that is incredibly useful. This last year we saw, last couple years, breakthroughs in multimodality AIs. Not only do you understand the word cat and the letter c-a-t, you could look at an image and hear the word, hear the sound meow and recognize all of those words, all of those information is aligned to the same place in your brain. And we can now do that with all kinds of different modalities which is very important to you. We made huge breakthroughs in reasoning capabilities to be able to decompose through composition. Break down a problem we've never seen before into routine things and through that composition solve problems step by step reasoning through hard problems. The ability to do multimodality reasoning and to be grounded by different sources of information is going to be a revolution even for your industry. It is clearly a revolution in ours.
The idea that we would use multimodality sensors, very sensible. But the idea that a car, a self-driving car stack that can think because it can reason through and experience a situation it's never experienced before. Reason through it step by step and be able to solve that problem, navigate through it. Incredible capabilities. It expanded the operating domain of self-driving cars tremendously. It's now the case that we are very certain we don't have to test every single car across every single scenario in order for that car to do well just like us. We can now have a car that can think and be basically intelligent. These techniques has made, and of course agentic capabilities, the ability for agents to reason, use tools, work with other agents completely revolutionary. And so these breakthroughs in the day that I've been here is permeating your industry, permeating life sciences, some of the most exciting work that I've seen in a very long time.
And so today we announced a very important partnership. Before I get there, let me say a couple of things. We announced of course some of the work that we've done with all of you. We've brought an accelerated computing platform. But on top of that are a suite of libraries that we're very proud of. Parabricks for gene sequencing. MONAI, an open-source platform for medical imaging. And one of the most important ones, BioNeMo, a platform that understands geometric structures like molecules and proteins and these platforms are used by many of you in the room and I'm very grateful for that. The several announcements that we made, before I say that, there are so many partners here, a couple hundred different companies here. They're pharmas or hospitals or startups. There are startups in robotics, surgical robotics. There's startups in robotics labs. There are agentic AI companies here helping with healthcare diagnostics and also some really cool work being done in AI scientists. You know, part of being at the JP Morgan event is that large companies meeting small companies and so many of them are doing revolutionary work and I'm so grateful that all of you have chosen the NVIDIA platform to work on and so our platforms ended up achieving some really great results. MONAI has been downloaded some six million times. We achieved the Guinness World Record of gene sequencing. You know, that's cool. I've always wanted to be in the Guinness World Record, so that's cool. The vast majority of the world's structure prediction models are built on NVIDIA today. And so, I'm really proud of that.
And on top of the BioNeMo platform are a whole bunch of pre-trained models. You use it for data processing. You use it for training these biological models. You use it for fine-tuning. But we also have a whole bunch of pre-trained models and a bunch of data that you can start with and you could apply your data to fine-tune your molecules and some of the models that we've created and participated and jointly developed are really proud of. AlphaFold for protein design, Resynthesis for molecular synthesis, Kermit for predicting toxicity, EVO 2 a DNA foundation model. These foundation models start of course, we're trying to learn the representation of proteins so that we can create a foundation model for protein. But the moment that time where we're going to learn a representation and create a foundation model for cells is just around the corner. The multi-protein systems, CodonFM for RNA design, just really terrific progress. And so today we announced a large number of partners that are working with us in BioNeMo and I'm very grateful for that. And we also announced a really landmark partnership with one of the most important companies in the world and a really great friend, Dave Ricks and Lilly.
And so I won't tell you more about it. I'm going to let the two of us describe it together. But Dave, he's been the CEO of Lilly since 2017. And in the context of the history of Lilly, 150 year old company, he's been in the company for 30 years. That's so he's basically a new college grad in the context of Lilly. And I'm anxious to learn from him, you know, how he learned so much in just a short period of time. But he's been with Lilly now 30 years and as you know he has helped guide and help build Lilly into one of the most valuable companies in the world and surely one of the most important companies in the world. And so if you could join me in welcoming my guest David Ricks.
Thank you, Dave, for being here. So, it turns out the two of us have a, I can't imagine more different upbringing. So, Dave was an Eagle Scout and I went to a juvenile delinquent school.
Just to be clear, I was placed there as an accident. It wasn't...
It was a very good school. It was very affordable. And so 2017 you became CEO.
My question is this. What was Lilly like and what was its vision for the future of life sciences and then what is it like today and your vision for the future of life sciences in the context of...
Yeah. Well, so in 2017, we were almost 140 years old or just a year older than that and had a pretty well established company. We had been through a difficult period of time in our industry and we'll get into this a little bit later. You know, we are plagued by as many people in the room know the cycles of innovation that are actually longer than the reward cycles on average. So you have this peak trough, you know, you're peaking and troughing and in the troughs of course you either don't survive, you end up as a back end of a hyphen on another company or you learn how to survive and you do difficult things. And I think it's true that Lilly is the only major pharma company that has not merged or combined ourselves to get through such a period. So, and we've been through a few of them. So, I think we've developed a bit of a DNA for scraping by in these tough times. And one of the things that really stuck with me, I was a leader on the executive committee at the time. We went through our last cycle with 2010, 11, 12, was, you know, how you have to be so relentless on invention because that time cycle is brutal. When you have a patent cliff and you have nothing to replace it, you have to rescale your whole company. And imagine doing this every 10 or 15 years, it's impossible to get momentum. People leave, you lose inherent knowledge, etc. So, we set about to do that. And I think that problem can be attacked two ways. One, you can say, okay, we'll just have better ideas. But as you may know, that's difficult to sustain. Even if you have a good idea, everyone thinks you're smart. They think about that about Lilly now. But it's hard to sustain. So we of course worked on that, more invention. But the more important thing we did was speed up invention. So we systematically took apart the R&D process literally like every step and there's hundreds of steps from candidate selection to first approval and we squish the time. And so still today probably Lilly is about 40% faster than the next scaled pharma company on average across all therapeutic areas and so we can invent things inside the patent cycle.
Then you can start to grow.
Yeah. So I think that was a notion in 2017 which we've actualized.
Along the way. Then of course we did a lot of things to drive more invention. We were lucky enough to be early on an idea called GLP-1 in 2006. We launched the first one in the world.
We fiddled around. Is it doing well?
If it weren't for you, it would be the most important idea in the stock market. Yeah. But never heard of it.
Pretty good idea for mankind and we fiddled around with it for like 12 years. And to our credit our...
Sounds like a chip, GLP-1.
GLP-1. Is there a two coming?
There's a three coming. So there is a GLP-2 that's actually a real protein.
It's a protein. It sort of does the opposite. But you don't want that. What we've done is combined two hormones.
Very small market. I gotta tell you.
That's right. That's right.
Extremely small. I don't know healthcare, but I'm guessing sign up on GLP-2 now.
Because tirzepatide is two proteins fused together and that's a technology we've developed too. And there's a triple acting one coming next year 27. So that'll be pretty interesting. But anyway, we caught that wave. Our competitor innovated into obesity. We chased and executed I think quite well to become the market leader last year. So we not only now have an engine that's faster, we have a tremendous success cycle with this one idea. The challenge before us now is like how do we find another success cycle before that one runs out, hopefully a lot sooner than it runs out and you know sort of get exit velocity which really no major company in our industry has had. They've just had these cycles. So yeah, you know, you said something earlier and I have several of my employees here and they hear me say this all the time. You know, the fundamental purpose of the CEO is to create the conditions by which, you know, our job isn't to create ideas. Our job is to create the conditions by which great ideas can systematically arise. And what you just described is essentially Lilly with wisdom that you have to create the conditions by which these great inventions can happen.
One of the, you know, we've spoken before about the early days of Nvidia getting into AI and I would say that without exception probably the single most important thing that we did was investing in the infrastructure, the creating the conditions by which amazing AI researchers can come to our company and great inventions could arise. You know, today we're quite silent about this but people are increasingly discovering that over the course of the last 10 years. One of the best decisions that I made in the beginning was, and at the time we couldn't afford it. But nonetheless I built one of the largest supercomputers in the world for Nvidia to use. And at the time I had nobody to use it but because Nvidia had very few AI researchers at the time but the problem is if you don't have AI researchers and therefore you don't have the instrument, if you don't have the instrument you'll never have the AI researchers because they need, scientists need the instrument necessary for their science. If you want to attract the world's greatest particle physicists, if you don't have a particle collider, you're going to have a very hard time doing that work. And so I came to the conclusion that the right answer is to create the conditions for great AI work to be done and I built that supercomputer and we attracted some amazing people and today we do frontier work in physical AI and biological AI and robotics AI and the work that we did in self-driving cars is revolutionary. And so which led our conversation to that AI is starting to make enough progress that we might be able to apply it to tackle some of the most extraordinary challenges of humanity which is biology.
And the two of us thought it'd be amazing if we have the largest computer company in the world and computer science company in the world partner with the world's largest life science company. And it's as if we created this lab where we are essentially one company combining all of the best amazing things together which led us to our announcement today.
Yeah. And so, so exciting. Yeah. Today we announced...
Yeah. And so as you know there's like four parts of this. We bought a bunch of chips and we're building, I think it'll be finished this week actually, the largest dedicated on-prem biology supercomputer in the world. That'll be in Indianapolis.
Which is the world center for biological discovery by the way.
You know I was going to say earlier because of the selection of the headquarters of Lilly it is exactly the reason why it has never merged with anyone.
Maybe so that could be true. That could be true. There's a lot of artifacts with that.
I, you know, I think that seriously there's something to like if you manage to make it all the way to...
Everybody knows what I mean.
If you manage to make it all the way...
And you're a discovery chemist or biologist, you really aren't going to leave, right? I mean, you made it all that way. You're going to stay. So, we have great retention. Anyway, we're building this huge... So, we have massive compute. Dave's invited me to visit him many times.
Yeah. He's never... Well, actually, you did come in 2018. I don't know if you remember that.
Actually, that just a sidebar like that was the first time we met and that you gave a talk to our management team, a room about this size about AI in 2018.
And I guess I was lucky enough to...
100% of the audience I could see. I could see it in their eyes. Nonsense.
Yeah. Honestly, it blew my mind. What you said blew my mind to the point where I spent several months like reading about some of these things which sounded like science fiction. And now it's fact. It's incredible what you built and it's great to partner with Nvidia on this. So supercomputer, we're going to put together a research team here in the Bay Area.
So if there's any smart scientists or computer scientists who want to work at the center it'll be a joint Lilly NVIDIA LA AI lab. We're going to develop new data because anyone who knows about this problem we have is you can train models but if the data is an inch thin you don't get very far and biology is largely unknown. So we need to create massive experimentation for the purpose of training machines and then we need to build on the work that Nvidia has done with the models you've created to ladder up to even more sophisticated predictions.
That's right. So super exciting and...
And so you know one of the important things to understand about AI is it's easy to think about it as a multi-layer cake. I mean at the lowest level you need some energy for these computers and then of course you have the chips and the computers. You have the infrastructure layer. That infrastructure layer is about, you know, how do you cause this computer, it's kind of like the operating system of AI. How do you cause this computer to learn, process the data, learn? How do you guardrail it? How do you fine-tune it? How do you teach it? And so that layer is kind of like the HR department of AI, you know, it's like...
Only they tell better jokes.
Exactly. And so that's the infrastructure layer and then the layer above that is the models. And the layer above that are the domain specific data and flywheel. And one of the areas that's really exciting is the idea that we would be able to train models, synthesize proteins or chemicals, go put it into a robotics lab, collect more data, take that data back into our model. And this flywheel of data generation or protein generation using AI, collecting ground truth experimentation and then putting that scientific flywheel on steroids. That future is just incredibly exciting. So we're systematically bringing together some of the brightest minds in the field in the domain expertise of drug discovery and some of the brightest minds in computer science and together, you know, between the two of us in our co-innovation lab, we'll have essentially every layer of that cake and put together hopefully a blueprint for what is possible in the future of drug discovery. I was going to tell you this. I was going to tell you earlier. I was going to tell you the story 40 years ago. So I guess 43 years ago I was the first generation of engineers where the concept of computer aided design was invented before my generation. My manager, that generation of engineers, they would design a chip and it would be a hit but they would go several years without another hit. And they would design another chip and then it would be a hit and they would go several years.
Exactly. And our generation came along and we were able to represent the transistor, the logical gates, the functionality. We learned the representation and described it in software. We did 43 years ago. Finally a computer could represent your domain and accurately simulate it. It's right to do as much in silico as possible. And so today we design massive chips. 15,000 people come together to design a computer system. And we just taped out Vera Rubin. Billions of dollars of R&D goes into it. The computer when it's done is three tons, you know, a million and a half parts. And six brand new chips, 220 trillion transistors. I mean, it's gigantic systems. 15,000 people coming together and that computer is going to work first time. It does exactly what we expected. And every single generation we build upon the last generation. Every single generation we build upon the last generation. I think your time has come for that.
Finally the computer, you know, a million times, a million times, a quadrillion times of computing power later. I think finally we might be able to represent one of the hardest problems in the world which is the human biology. And I'm really hoping that your industry moves from drug discovery, which is kind of like wandering around the forest looking for truffles, which literally...
Sidebar, we used to have a department of soil discovery. Seriously, where we would send people into tropical forests and scoop up soils and bring them back to Indianapolis. And refine out what became antibiotics. In fact, the most important antibiotic in the world still is vancomycin, the last resort antibiotic in the hospital. We couldn't run hospitals without it. It was discovered in Borneo in a soil sample by Lilly.
Now, how do they know to go there?
They would just go to exact because fungus produce as tools to fight bacteria and so they dig in the soil and discover these naturally occurring defense mechanisms, refine them into drugs.
But that is true drug discovery. Okay, when you think about it, right? That's like you could get malaria doing that job. Like that was in the 50s. That was, you know, 70 years ago.
Well, see, now your job gets to be just like my job. You never have to leave your office.
There we go. Yeah. Do it all in silico. Yeah.
Yeah. So, I think that's what really excites me. The concept that we could have computer-aided drug design. And that this is going to be of course the science behind it is going to be incredible. It's incredibly complex. The human biology is incredibly complex and it's going to be still a decade yet. But I really think that this announcement between us, you know, could be a blueprint.
Start of something. Yeah. Start something. I hope you're right. I mean, we need it, right? I think if we think of all the unmet needs left at a conference at JP Morgan, you're reminded of like the incredible pipeline of the industry, but most of that is sort of like chipped out of stone, you know, it's like hard work. And especially occasionally we can harness systems like when monoclonal antibodies came around we basically hijacked like biology's way of making solutions to infections and other things but mostly what we do is like really empirical and like chipping away and very difficult. Each small molecule discovery is like a work of art. There's only one of them and you know one brush stroke and it's wrong. If we can make that an engineering problem versus you know this sort of discovery artisanal drug making problem think of the impact on human life.
The first chips that I worked on, my first generation of computers were quite empirical. There are some rules that we used and we apply some intuition. We have no clue how it's going to perform. We know that it's going to perform better than the last one we built because we made mistakes and so we could at least correct those mistakes and we'll have more transistors and more everything. So it's going to do better.
However, the functionality of it unknown. Its full functionality unknown. Its performance completely estimated. Today we know the functionality of the systems we created down to every single bit. The performance we could measure. We simulate everything.
And so we know exactly what we're going to produce. I really believe that that could revolutionize the future of...
It would make that come true even for a part of human biology.
Yeah, that's right. And so tell me, GLP-1, you were saying that it's really in its third phase.
And where does it go from here? What happens? What other things can you use it for? How could it change everything?
Yeah. Well, I think as it turns out, you know, obesity is like a key linchpin driver of chronic disease in adults and like a lot of things in evolution, most the ailments of today weren't concerns in our evolutionary past, but we evolve much slower than our environment is and so we're maladapted to the modern world. And obesity is one of those things. We grew up in a world of scarcity. Sounds like you grew up in a world of scarcity, but most people here didn't. Our ancestors did. And so...
There was no need for an off switch on hunger.
The word is abundance now.
And we live in a world of abundance of intelligence, abundance of productivity, abundance of those kind of things. And I don't know what you were thinking about, but my life is surrounded by abundance.
But unfortunately, we're all in caloric abundance now. And there was no need for us to wire in an off switch evolutionarily. So, that's why we have obesity. And that leads to many...
Many chronic diseases that ail people and probably shorten life. So, I think by having a pretty effective set of tools going to get more effective, and I'll get back to that in a second, to really help people manage that to create an off switch synthetically in a way.
We can downregulate a lot of diseases, over 200 chronic diseases.
