Jensen Huang details NVIDIA's $4 trillion trajectory, driven by extreme co-design that shards AI across thousands of GPUs to bypass Amdahl's Law and Moore's Law slowdowns. He outlines four scaling laws—pre-training, post-training, test time, and agentic—predicting future growth via "token factories" that could value NVIDIA at $10 trillion. Addressing power constraints, Huang advocates for graceful data center degradation and aligns global CEOs to secure HBM4 memory and EUV lithography. While asserting AGI exists today, he argues professions will expand as AI automates tasks, leaving humanity's unique compassion and creativity intact. Ultimately, his philosophy of continuous knowledge transfer ensures NVIDIA's legacy thrives beyond any single leader, fostering a future where disease ends and travel reaches light speed. [Automatically generated summary]
The following is a conversation with Jensen Huang, CEO of NVIDIA, one of the most important and influential companies in the history of human civilization.
NVIDIA is the engine powering the AI revolution.
And a lot of its success can be directly attributed to Jensen's sheer force of will and his many brilliant bets and decisions as a leader, engineer, and innovator.
This is Alex Friedman Podcast.
And now, dear friends, here's Jensen Huang.
You've propelled NVIDIA into a new era in AI, moving beyond its focus on chip scale design to now rack scale design.
And I think it's fair to say that winning for NVIDIA for a long time used to be about building the best GPU possible.
And you still do.
But now you've expanded that to extreme co-design of GPU, CPU, memory, networking, storage, power, cooling, software, the rack itself, the pod that you've announced, and even the data center.
So let's talk about extreme co-design.
What is the hardest part of co-designing a system with that many complex components and design variables?
So first of all, the reason why extreme co-design is necessary is because the problem no longer fits inside one computer to be accelerated by one GPU.
The problem that you're trying to solve is you would like to go faster than the number of computers that you add.
So you added, you know, 10,000 computers, but you would like it to go a million times faster.
Then all of a sudden, you have to take the algorithm, you have to break up the algorithm, you have to refactor it, you have to shard the pipeline, you have to shard the data, you have to shard the model.
Now, all of a sudden, when you distribute the problem this way, not just scaling up the problem, but you're distributing the problem, then everything gets in the way.
This is the Amdahl's law problem, where the amount of speed up you have for something depends on how much of the total workload it is.
And so if computation represents 50% of the problem, and I sped up computation infinitely, like a million times, you know, I only sped up the total workload by a factor of two.
Now, all of a sudden, not only do you have to distribute the computation, you have to, you know, shard the pipeline somehow, you also have to solve the networking problem because you've got all of these computers are all connected together.
And so distributed computing at the scale that we do, the CPU is a problem, the GPU is a problem, the networking is a problem, the switching is a problem, and distributing the workload across all these computers are a problem.
It's just a massively complex computer science problem.
And so we just got to bring every technology to bear.
Otherwise, we scale up linearly or we scale up based on the capabilities of Moore's Law, which has largely slowed because Denard's scaling has slowed.
Plus, you have a completely disparate disciplines here.
I'm sure you have specialists in each one of these, high-bandwidth memory, the networking, the NV-Link, the NIX, the optics and the copper that you're doing, the power delivery, the cooling, all that.
I mean, there's like world experts in each of those.
How do you get them in a room together to figure out?
There's the first question, which is, what is extreme co-design?
We're optimizing across the entire stack of software from architectures to chips to systems to system software to the algorithms to the applications.
That's one layer.
The second thing that you and I just talked about is goes beyond CPUs and GPUs and networking chips and scale-up switches and scale-out switches.
And then, of course, you've got to include power and cooling and all of that because, you know, all these computers are extremely, extremely power-hungry.
They do a lot of work and they're very energy efficient, but they, in aggregate, still consume a lot of power.
And so that's one, the first question is: what is it?
The second question is, why is it?
And we just spoke about the reason, you know, you want to distribute the workload so that you can exceed the benefit of just increasing the number of computers.
And then the third question is: how is it?
How do you do it?
And that's kind of the miracle of this company.
You know, when you're designing a computer, you have to have an operating system of computers.
When you're designing a company, you should first think about what is it that you want the company to produce.
You know, I see a lot of companies' organization charts and they all look the same: hamburger organization charts, software organization charts, and car company organization charts, they all look the same.
And it doesn't make any sense to me.
You know, the goal of a company is to be the machinery, the mechanism, the system that produces the output.
And that output is the product that we'd like to create.
It is also designed, the architecture of the company should reflect the environment by which it exists.
It almost directly says what you should do with the organization.
My direct staff is 60 people.
You know, I don't have one-on-ones with them because it's impossible.
You can't have 60 people on your staff if you're going to get work done.
So, as you mentioned, NVIDIA is this company that's adapting to the environment.
So, which point can you say, did the environment change?
You began adapting sort of secretly in the early days from GPU for gaming, maybe the early deep learning revolution, to we're now going to start thinking of it as an AI factory.
We started out as an accelerator company, but the problem with accelerators is that the application domain is too narrow.
It has the benefit of being incredibly optimized for the job.
Any specialist has that benefit.
The problem with intense specialization is that, of course, your market reach is narrower, but that's even fine.
The problem is the market size also dictates your RD capacity.
And your R D capacity ultimately dictates the influence and impact that you can possibly have in computing.
And so when we first started out as an accelerator, very specific accelerator, we always knew that that was going to be our first step.
We had to find a way to become accelerated computing.
But the problem is when you become a computing company, it's too general purpose and it takes away from your specialization.
I connected two words that are actually have fundamental tension.
The better computing company we become, the worse we become as a specialist.
The more of a specialist, the less capacity we have to do overall computing.
And I connected those two words together on purpose, that the company has to find that really narrow path, step by step by step, to expand our aperture of computing, but not give up on the most important specialization that we had.
So the first step that we took beyond acceleration was we invented the programmable pixel shader.
So that was the first step towards programmability.
It was our first journey towards moving into the world of computing.
The second thing that we did was we created, we put FP32 into our shaders.
That FP32 step, IEEE compatible FP32, was a huge step in the direction of computing.
It was the reason why all of the people who were working on stream processors and other types of data flow processors discovered us.
And they said, hey, all of a sudden, we might be able to use this GPU that's incredibly computationally intensive.
And it's now compliant with IEEE.
I can take my software that I was writing previously on CPUs and I can see about using the GPU for that.
And which led us to put C on top of FP32, what's called, we call CG.
That CG path took us to eventually CUDA, CUDA, step by step by step.
Well, putting CUDA on GeForce, that was a strategic decision that was very, very hard to do because it cost the company enormous amounts of our profits and we couldn't afford it at the time.
But we did it anyways because we wanted to be a computing company.
A computing company has a computing architecture.
A computing architecture has to be compatible across all of the chips that we build.
So we invented this thing called CUDA, and it expanded the aperture of applications that we can accelerate with our accelerator.
The question is, how do we attract developers to CUDA?
Because a computing platform is all about developers.
And developers don't come to a computing platform just because it could perform something interesting.
They come to a computing platform because the installed base is large.
Because a developer, like anybody else, wants to develop software that reaches a lot of people.
So the install base is, in fact, the single most important part of an architecture.
The architecture could attract enormous amounts of criticism.
For example, no architecture has ever attracted more criticism than the x86, you know, as a less than elegant architecture.
But yet, it is the defining architecture of today.
It gives you an example that, in fact, so many RISC architectures, which were beautifully architected, incredibly well designed by some of the brightest computer scientists in the world, largely failed.
And so I've given you two examples where one is, you know, one is elegant, the other one's barely aesthetic.
And so there were other architectures at the time.
CUDA came out, OpenCL was here.
There were, you know, there's several other competing architectures.
But the thing that the decision that we made that was good was we said, hey, look, ultimately, it's about installed base and what is the best way we could get a new computing architecture into the world.
By that timeframe, GeForce had become successful.
We were already selling millions and millions of GeForce GPUs a year.
And we said, you know, we ought to put CUDA on GeForce and put it into every single PC, whether customers use it or not, and use it as a starting point of cultivating our installed base.
Meanwhile, we'll go and attract developers and went to universities and wrote books and taught classes and put CUDA everywhere.
And eventually people discover, and at the time, the PC was the primary computing vehicle.
There was no cloud.
And we could put a supercomputer in the hands of every researcher in school, every scientist, every engineering school, every student in school.
And eventually something amazing will happen.
Well, the problem was CUDA increased our cost of that GPU, which is a consumer product, so tremendously, it completely consumed all of the company's gross profit dollars.
And so at the time, the company was probably worth, I don't know, at the time, was it like $8 billion or something, six, seven billion dollars or something like that.
After we launched CUDA, I recognized that it was going to add so much cost, but it was something we believed in.
Well, I had to make it clear to the board what we're trying to do.
And the management team knew our gross margins were going to get crushed.
So you could imagine a world where GeForce would carry the burden of CUDA, and none of the gamers would appreciate it, and none of the gamers would pay for it.
You know, they only pay a certain price, and it doesn't matter what your cost is.
And so we increased our cost by 50% and that consumed, and we were a 35% gross margin company.
And so it was quite a difficult decision to make.
But you could imagine that someday this could go into workstations and it would go into supercomputers and in those segments, maybe we can capture more margin.
So you could reason your way into being able to afford this, but it still took a decade.
The management team will reason about it, all the people that we spent a lot of time reasoning about it.
The thing that the next part of it is probably a skill thing, which is, you know, oftentimes in leadership, the leadership stays quiet or they learn about something and then they do some manifesto and it's a brand new year and somehow at the end of the year, next year, we're going to have a brand new plan, big, huge layoff this way, big, huge organization change this way, new mission statement, brand new logos, you know, that kind of stuff.
We've just never, I never do things that way.
When I learn about something and it's starting to influence how I think, I'll make it very clear to everybody near me that, you know, this is interesting.
This is going to make a difference.
This is going to impact that.
And I reason about things step by step by step.
Oftentimes, I've already made up my mind, but I'll take every possible opportunity, external information, new insights, new discoveries, new engineering revelations, new milestones develop.
I'll take those opportunities and I'll use it to shape everybody else's belief system.
And I'm doing that literally every single day.
I'm doing that with my board.
I'm doing that with my management team.
I'm doing that with my employees.
I'm trying to shape their belief system such that when I come the day I say, hey, let's buy Mellanox, it's completely obvious to everybody that we absolutely should.
On the day that I that I said, hey guys, let's go all in on deep learning.
And let me tell you why.
I've already been laying down the bricks to different organizations inside the company.
Every organization and everybody, many of the people might have heard everything.
Most of the company heard here's, of course, pieces of it.
And on the day that I announce it, everybody's kind of bought into many pieces of it.
And in a lot of ways, I like to announce these things.
And I imagine that the employees are kind of saying, you know, Jensen, what took you so long?
And in fact, I've been shaping their belief system for some time.
And therefore, leadership, sometimes it looks like you're leading from behind.
But you've been shaping their, you know, to the point where on the day that I declared it, 100% buy-in.
But that's what you want.
You want to bring everybody along.
You know, otherwise we announce something about deep learning and everybody goes, what are you talking about?
You know, you announce something about, let's go all in on this thing.
And your management team, your board, your employees, your customers, they're kind of like, where's this coming from?
You know, this is insane.
And so GTC, in fact, if you go back in time, you look at the keynotes, I'm also shaping the belief system of my partners and the industry.
And I'm using that to shape the belief system of my own employees.
And so by the time that I announce something, like, for example, we just announced Grok.
I've been talking about the stepping stones for two and a half years.
You guys go back and, oh my gosh, they've been talking about it for two and a half years.
And so I've been laying the foundation step by step by step.
So when the time comes, you announce it, everybody's, you know, what took you so long.
We don't, as it turns out, we're a computing platform company.
And so nobody can buy anything from us.
That's the weird thing.
You know, we vertically design, vertically integrate to design and optimize, but then we open up the entire platform at every single layer to be integrated into other companies' products and services and clouds and supercomputers and OEM computers.
And so the amazing thing is, I can't do what I do without having convinced them first.
And so most of GTC is about manifesting a future that by the time that we, my product is ready, they're going, what took you so long?
So I think you've outlined four of them with pre-training, post-training, test time, and agentic scaling.
What do you think when you think about the future, deep future and the near-term future, what are the blockers that you're most concerned about that keep you up at night that you have to overcome in order to keep scaling?
Well, we can go back and reflect on what people thought were blockers.
So in the beginning, we were the first, the pre-training scaling law.
People thought, well, rightfully so, that the amount of data that we have, high-quality data that we have, will limit the intelligence that we achieve.
And that scaling law was an important, very important scaling law.
The larger the model, the correspondingly more data results in a better, with a results in a smarter AI.
And so that was pre-training.
And Ilyas, Suscova, Ilya said, we're out of data or something like that.
Pre-training is over or something like that.
The industry panicked, you know, that this is the end of AI.
And of course, that's obviously not true.
We're going to keep on scaling the amount of data that we have to train with.
A lot of that data is probably going to be synthetic.
And that also confused people, you know, and what people don't realize is they've kind of forgotten that most of the data that we are training, that we teach each other with, inform each other with, is synthetic.
You know, it's synthetic because it didn't come out of nature.
You created it.
I'm consuming it.
I modify it, augment it.
I regenerate it.
Somebody else consumes it.
And so we've now reached a level where AI is able to take ground truth, augment it, enhance it, synthetically generate an enormous amount of data.
And that part of post-training continues to scale.
And so the amount of data that we could use that is human-generated will be smaller and smaller and smaller.
The amount of data that we use to train model is going to continue to scale to the point where we're no longer limited.
Training is no longer limited by data.
It's now limited by compute.
And the reason for that is most of the data is synthetic.
Then the next phase is test time.
And I still remember people telling me that inference, oh, yeah, that's easy.
Pre-training, that's hard.
These are giant systems that people are talking about.
Inference must be easy.
And so inference chips are going to be little tiny chips.
And they're not like NVIDIA's chips.
Oh, those are going to be complicated and expensive.
And we could make, and this is, and in the future, inference is going to be the biggest market.
And it's going to be easy.
And we're going to commoditize it.
And everybody can build their own chips.
And that was always illogical to me because inference is thinking.
And I think thinking is hard.
Thinking is way harder than reading.
Pre-training is just memorization and generalization and looking for patterns and relationships.
You're reading and reading versus thinking, reasoning, solving problems, taking unexplored experiences, new experiences, and breaking it down into decomposing it into solvable pieces that we then go off either through first principle reasoning or through previous examples, prior experiences,
or just exploration and search and trying different things.
And that whole process of test time scaling, inference is really about thinking.
And it's about reasoning.
It's about planning.
It's about search.
And so how could that possibly be compute light?
And we were absolutely right about that.
So test time scaling is intensely compute intensive.
Then the question is: okay, now we're at inference and we're at test time scaling.
What's beyond that?
Well, obviously, we have now created one agentic person.
And that one agentic person has a large language model that we've now developed.
But during test time, that agentic system goes off and does research and bangs on databases and it goes and uses tools.
One of the most important things it does is spins off and spawns off a whole bunch of sub-agents, which means we're now creating large teams.
It's so much easier to scale NVIDIA by hiring more employees than it is to scale myself.
And so the next scaling law is the agentic scaling law.
It's kind of like multiplying AI.
Multiplying AI, we could spin off agents as fast as you want to spin off agents.
And so, you know, you now have four scaling laws.
And as we use the agentic systems, they're going to create a lot more data.
They're going to create a lot of experiences.
Some of it we're going to say, wow, this is really good.
We ought to memorize this.
That data set then comes all the way back to pre-training.
We memorize and generalize it.
We then refine it and fine-tune it back into post-training.
Then we enhance it even more with test time, you know, and the agents and agentic systems, you know, put it out into the industry.
And so this loop, the cycle, is going to go on and on and on.
It kind of comes down to basically intelligence is going to scale by one thing, and it's compute.
But there's a tricky thing there that you have to anticipate and predict, which is some of these components requires different kinds of hardware to really do it optimally.
So you have to anticipate where the AI innovation is going to lead.
For example, make sure that you're experts with sparsity.
And then the last part is to have an architecture that's flexible, that can adapt and move with the wind.
And one of the benefits of CUDA is that it's, you know, on the one hand, an incredible accelerator.
On the other hand, it's really flexible.
And so that balance, incredible balance between specialization, otherwise we can't accelerate the CPU, versus generalization so that we can adapt with changing algorithms.
That's really, really important.
That's the reason why CUDA has been so resilient on the one hand, and yet we continue to enhance it.
We're at CUDA 13.2.
And so we're evolving the architecture so fast that we can stay with you know with with the with the modern algorithms.
For example, when Mixture Experts came out, that's the reason why we had MVLink 72 instead of MVLink 8.
We could now take an entire 4 trillion, 10 trillion per annum model and put it in one computing domain as if it's running on one GPU.
People probably didn't notice.
Said it, but if you look at the architecture of the Grace Blackwall racks, it was completely focused on doing one thing: processing the LLM.
All of a sudden, one year later, you're looking at a Verarubin rack.
It has storage accelerators.
It has this incredible new CPU called Vera.
It has Verarubin and NV-Link 72 to run the LLMs.
It also has this new additional rack called Grok.
And so this entire rack system is completely different than the previous one.
And it's got all these new components in it.
And the reason for that is because the last one was designed to run MOE large language models, inference, and this one is to run agents.
No matter what happens, at some point, in order for that large language model to be a digital worker, let's just use that metaphor.
Let's say that we want the LLM to be a digital worker.
What does it have to do?
It has to access ground truth.
That's our file system.
It has to be able to do research.
It doesn't know everything.
And I don't want to wait until this AI becomes universally smart about everything, past, present, and future before I make it useful.
And so therefore, I might as well let it go do research.
It's obviously, if it wants to help me, it's got to use my tools.
You know, a lot of people would say, you know, AI is going to completely destroy software.
We don't need software anymore.
We don't even need tools anymore.
That's ridiculous.
Let's use a thought experiment.
And you could just sit there and enjoy a glass of whiskey and think about all these things, and it would become completely obvious.
Like if I were to create the most amazing road, the most amazing agent that we can imagine in the next 10 years, let's say be a human or robot.
If that human or robot were to be created, is it more likely that the human or robot comes into my house and uses the tools that I have to do the work that it needs to do?
Or does this hand turn into a 10-pound hammer in one instance, turns into a scalpel in another instance, and in order to boil water, it beams microwaves out of its fingers?
Or is it more likely just to use a microwave?
And the first time it goes up to the microwave, it probably doesn't know how to use it.
But that's okay.
It's connected to the internet.
It reads the manual of this microwave, reads it, instantly becomes an expert, and so uses it.
And so I think I just described, in fact, almost all of the properties of OpenClaw.
That it's going to use tools, that it's going to access files.
It's going to be able to do research.
It has IO subsystem.
And when you're done reasoning about it through it in that way, then you say, oh my gosh, the impact to the future computing is deeply profound.
And the reason for that is, I think we've just reinvented the computer.
And then now you say, okay, when did we reason about that?
When did we reason about OpenClaw?
If you take the OpenClaw schematic that I used at GTC, you will find it two years ago.
Literally two years ago at GTC, I was talking about agentic systems that exactly reflect OpenClaw today.
And of course, the confluence of many things had to happen.
First of all, we needed Claude and GPT and all of these models to reach a level of capability.
So their innovation and their breakthroughs and their continued advances was really important.
And then, of course, somebody had to create an open source project that was sufficiently robust, and sufficiently complete, and that we can all put to work.
And I think OpenClaw did for agentic systems what ChatGPT did for generative systems.
So part of it is also the humans that represent the thing.
And part of it is memes.
Because we're all trying to figure it out.
There's really serious and complicated security concerns about when you have such powerful technology, how do you hand over your data so they can do useful stuff?
But then there are scary things associated with that.
And we, as a civilization, as individual people and as a civilization, figuring out how to find that right balance.
Power is a concern, but it's not the only concern.
But that's the reason why we're pushing so hard on extreme co-design so that we can improve the tokens per second per watt, orders of magnitude, every single year.
And so in the last 10 years, Moore's Law would have progressed computing about 100 times in the last 10 years.
We progressed and scaled up computing by a million times in the last 10 years.
And so we're going to keep on doing that through Extreme Co-Design.
So energy efficiency per watt completely affects the revenues of a company.
It affects the revenues of a factory.
And we're just going to push that to the limit so that we can keep on driving token costs down as fast as we can.
You know, our computer price is going up, but our token generation effectiveness is going up so much faster that token cost is coming down.
It's coming down an order of magnitude every year.
You've talked about small module nuclear power plants.
There's all kinds of ideas for energy.
How much does it keep you up at night, the bottlenecks in the supply chain of AI, like ASML with EUV lithography machines, TSMC with advanced packaging, like COAS, and SK Henix with high-bandwidth memory?
No company in history has ever grown at a scale that we're growing while accelerating that growth.
It's incredible.
And it's hard for people to even understand this.
In the overall world of AI computing, we're increasing share.
And so supply chain upstream and downstream are really important to us.
I spent a lot of time informing all the CEOs that I work with what are the dynamics that's going to cause the growth to continue or even accelerate.
It's part of the reasons why to the entire right-hand side of me were CEOs of practically the entire IT industry upstream and practically the entire infrastructure industry downstream.
And they were all, there were several hundred CEOs.
And I don't think there's ever been keynotes where several hundred CEOs show up.
And part of it is I'm telling them about our business condition now.
I'm telling them about the growth drivers in the very near future and what's happening.
And I'm also describing where we're going to go next so that they could use all of this information and all of the dynamics that are here to inform how they want to invest.
And so I inform them that way like I inform my own employees.
And then, of course, then I make trips out to them and make sure that, hey, listen, I want you to know this quarter, this coming year, this next year, these things are going to happen.
And if you look at the CEOs of the DRAM industry, the number one DRAM in the world was DDR memory for CPUs in data centers.
About three years ago, I was able to convince several of the CEOs that even though at the time HBM memory was used quite scarcely, you know, and barely by supercomputers, that this was going to be a mainstream memory for data centers in the future.
And at first it sounded ridiculous, but several of the CEOs believed me and decided to invest in building HBM memories.
Another memory was rather odd to put into a data center is the low power memories that we use for cell phones.
And we wanted them to adapt them for supercomputers in the data center.
And they go, cell phone memory for supercomputers?
And I explained to them why.
Well, look at these two memories, LPDDR5, HBM4.
The volumes are so incredible.
All three of them had record years in history.
And these are 45-year companies.
And so, you know, that's part of my job is to inform and shape, inspire, you know.
Because we changed the system architecture from the original DGX1 that you remembered to MVLink 72 rack scale computing.
What does that mean?
What does that mean to software?
What does that mean to engineering?
What does that mean to how we design and test?
And what does that mean to the supply chain?
Well, one of the things that it meant was we moved supercomputer integration at the data center into supercomputer manufacturing in the supply chain.
If you're doing that, you also have to recognize you're going to move and if your total footprint of whatever data center you're going to build, let's say you would like to have 50 gigawatts of supercomputers that are running simultaneously, and it takes one week to manufacture that 50 gigawatts of supercomputers,
then each week in the supply chain, the supercomputers are going to need a gigawatt of power.
And so we're going to need the supply chain to increase the amount of power it has to build, test, to build and test the supercomputers in the supply chain before I ship it.
Well, MVLink 72 literally builds supercomputers in the supply chain and ships them two, three tons at a time per rack.
They used to come in parts and we used to assemble them inside the data center.
But that's impossible now because MVLink 72 is so dense.
And so that's an example.
And I would have to go into, you know, I fly into the supply chain, go meet my partners and say, hey, I said, guess what?
So here's what I'm going to do with this is the way we used to build our DGXs.
We're going to build them this way.
This is going to be so much better because we're going to need them for inference.
The market for inference is, you know, coming.
The inflection point for inference is coming.
It's going to be a big market.
And so I first explain to them what's going on, why it's going to happen.
And then I ask them to make several billion dollars of capital investments each.
And because they, you know, they trust me and I'm very respectful of them.
And I give them every opportunity to question me.
And I spend time to explain things to people and I reason about it.
I draw them pictures and I reason about it in first principles.
And by the time I'm done with them, there's no what to do.
One of the areas, Lex, that I would love us to talk about and just get the message out.
You know, our power grid is designed for the worst case condition with some margin.
Well, 99% of the time, we're nowhere near the worst case condition because the worst case condition is a few days in the winter, a few days in the summer, and extreme weather.
Most of the time, we're nowhere near the worst case condition, and we're probably running around, call it 60% of peak.
And so 99% of the time, our power grid has excess power, and they're just sitting idle.
But they have to be there sitting idle because just in case, when the time comes, hospitals have to be powered and infrastructure has to be powered and airports have to run and so on and so forth.
And so the question that I have is whether we could go and help them understand and create contractual agreements and design computer architecture systems, data centers, such that when they need the maximum power for infrastructure in society, that the data centers would get less.
And during that time, we either have a backup generator for that little part of it, or we just have our computers shift to workload somewhere else, or we have the computers just run slower.
You know, we could degrade our performance, reduce our power consumption, and provide for a slightly longer latency response when somebody asks for an answer.
And so I think that that way of using computers, of building data centers, instead of expecting 100% uptime and these contracts that are really, really quite rigorous, it's putting a lot of pressure on the grid to be able to, now they're going to have to increase from their maximum.
The end customer puts requirements on the data centers that they can never not be available.
Okay, so that the end customer expects perfection.
Now, in order to deliver that perfection, you need a combination of backup generators and your grid power supplier to deliver on perfection.
And so everybody's got to have six nines.
Well, I think, first of all, right now, we ought to have everybody understand that when the customer asks for these things, you got somebody, you have somebody in your data center operations team disconnected from the CEO.
I bet the CEO doesn't know this.
I'm going to talk to all the CEOs.
The CEOs are probably not paying any attention to the contracts that are being signed.
And so everybody wants to sign the best contract, of course.
And they go down to the cloud service providers and the contract, the two contract negotiators that are, I could just see them now, you know, negotiating these multi-year contracts.
Both sides want the best contract.
As a result, The CSPs then have to go down to the utilities and they expect the six nines.
And so I think the first thing is just make sure that all of the customers, the CEOs of the customers, realize what they're asking for.
Now, the second thing is we have to build data centers that gracefully degrade.
And so if the power, if the utility of the grid tells us, listen, we're going to have to back you down to about 80%, we're going to say that's no problem at all.
We're just going to move our workload around.
We're going to make sure that data is never lost, but we can reduce the computing rate and use less energy.
The quality of service degrades a little bit.
For the critical workloads, I shift that somewhere else right away.
So I don't have that problem.
And so, you know, whoever, whichever data center still has 100% uptime.
And the third thing is we need the utilities to also recognize that this is an opportunity.
And instead of saying, look, it's going to take me five years to increase my grid capability, if you're willing to take power of this level of guarantee, I can make them available for you next month and at this price.
And so if utilities also offered more segments of power delivery promises, then I think everybody will figure out what to do with it.
But there's just way too much waste in the grid right now.
You've highly lauded Elon and XAI's accomplishment in Memphis in building Colossus supercomputer, probably in record time in just four months.
It's now at 200,000 GPUs and growing very quickly.
Is there something that you could speak to the understanding about his approach that's instructive broadly to all the data center creators that enable that kind of accomplishment?
His approach to engineering, his approach to the whole management of construction, everything.
And then that becomes an engineering question often.
And yes, I think when you get the ground truth of actually, I remember one of the times I was hanging out with him, he literally is going through the entire process of how to plug in cables into a rack.
He was working with an engineer on the ground that's doing that task.
And he's just trying to understand what does that process look like so it can be less error prone.
And just building up that intuition from every single task involved in putting together the data center, you start to immediately get a sense at the detailed scale and then at the broad system scale of where the inefficiencies are.
And so you can make it more and more and more efficient.
Plus you have the big hammer of being able to say, let's do it totally different and remove all possible blockers.
Well, first of all, co-design is an ultimate systems engineering problem.
And so we approach the work that we do from that principle.
The other thing that we do, and this is a philosophy that a thought, a state of mind, I guess, a method that I started 30 years ago, and it's called the speed of light.
A speed of light is not just about the speed.
Speed of light is my shorthand for what's the limit of what physics can do.
And so everything that we do is compared against the speed of light.
Memory speed, math speed, power, cost, time, effort, number of people, manufacturing cycle time.
And when you think about latency versus throughput, when you think about cost versus throughput, cost versus capacity, all of these things, you test against the speed of light to achieve all of these different constraints separately.
And then when you consider it together, you know you have to make compromises because a system that achieves extremely low latency versus a cheap system that achieves very high throughput are architected fundamentally differently.
But you want to know what's the speed of light of a system that achieves high throughput?
What's the speed of light of a system that achieves low latency?
And then when you think about the total system, you could make trade-offs.
And so I force everybody to think about what's this, what the first principles, the limits, the physical limits for everything before we do anything.
And we test everything against that.
And so that's a good frame of mind.
I don't love the other methods, which is continuous improvement.
The problem with continuous improvement, first of all, you should engineer something from first principles at the speed, you know, with speed of light thinking, limited only by physical limits and physics limits.
And after that, of course, you would improve it over time.
But I don't like going into a problem and somebody says, hey, you know, it takes 74 days to do this today, right now, and we can do it for you in 72 days.
You know, I'd rather strip it all back to zero.
And say, first of all, explain to me why it's 74 days in the first place.
And let's know, let's think about what's possible today.
And if I were to build it completely from scratch, you know, how long would it take?
Oftentimes, you'd be surprised and might come to six days.
Now, the rest of the six days to 74 could be very well reasoned and compromises and cost reductions and all kinds of different things.
But at least you know what they are.
And now that you know that six days is possible, then the conversation from 74 to six, surprisingly much more effective.
In such incredibly complex systems that you're working with, is simplicity sometimes a good heuristic to reach for?
I mean, if I can just, I mean, the pod, the Verarubin pod that you announced is just incredible.
We're talking about seven chips, seven chip types, five purpose built rack types, 40 racks, 1.2 quadrillion transistors, nearly 20,000 NVIDIA dies, over 1,100 Ruben GPUs, 60 exaflops, 10 petabytes per second of scale bandwidth.
I mean, so you have the, and then even the NVL72 rack alone is 1.3 million components, 1,300 chips, 4,000 pounds crammed into a single 19-inch wide rack.
And have to work together and report directly to you.
This is wonderful.
You've recently traveled to China.
So it's interesting to ask you: China's been incredibly successful in building up its technology sector.
What do you understand about how China is able to, over the past 10 years, build so many incredible world-class companies, world-class engineering teams, and just this technology ecosystem that produces so many incredible products?
50% of the world's AI researchers are Chinese, plus or minus.
And they're mostly in China, still.
We have many of them here, but there's amazing researchers still in China.
Their tech industry showed up at precisely the right time.
At the time of the mobile cloud era, their way of contributing was software.
And so this is a country's incredible science and math, really well-educated kids.
Their tech industry was created during the era of software.
They're very comfortable with modern software.
China is not one giant economic country.
It's got many provinces and cities with mayors all competing with each other.
That's the reason why there's so many EV companies.
That's the reason why there's so many AI companies.
That's the reason why there's so many every company you could imagine, they all create some of them.
And as a result, they have insane competition internally.
And what remains is an incredible company.
They also have a social culture where it's family first, friends second, and company third.
And so the amount of conversation that goes back and forth between they're essentially open source all the time.
So the fact that they contribute more to open source is so sensible because they're probably, what are we protecting?
You know, my engineers, their brothers are in that company.
Their friends are in that company, and they're all schoolmates, you know, the schoolmate concept.
It's a, you know, one schoolmate, your brother for life.
And so they share knowledge very, very quickly.
And so there's no sense keeping technology hidden.
You might as well put it on open source.
And so the open source community then amplifies, accelerates the innovation process.
So you get this rapid, incredible great talent, rapid innovation because of open source and just, you know, the nature of friends.
And insane competition among the company, what emerges is incredible stuff.
And so this is the fastest innovating country in the world today.
And this is something that has everything that I've just said is fundamental to just how the kids were grown, the fact that they have excellent education, the fact that parents want them to do well in school, the fact that their culture is that way.
These are just the thing about their country.
And they showed up at precisely the time when technology is going through that exponential.
And thank you for releasing open source Nematron 3 Super, which you can also use inside Perplexity to look stuff up, which is 120 billion parameter open weight MOE model.
What's your vision with open source?
So you mentioned China with DeepSee, Comminimax, with all these companies really pushing forward the open source AI movement.
And NVIDIA is really leading the way in close to state-of-the-art open source LLMs.
First, If we're going to be a great AI computing company, we have to understand how AI models are evolving.
One of the things that I love about Nemotron 3 is it's not just a pure transformer model, it's Transformer and SSMs.
And we were early in developing the conditional GANs, which that progressive GANs, which led step-by-step to diffusion.
And so the fact that we're doing basic research in model architecture and in different domains gives us visibility into what kind of computing systems would do a good job for future models.
And so it is part of our extreme co-design strategy.
Second, I think we rightfully recognize that on the one hand, we want world-class models as products, and they should be proprietary.
On the other hand, we also want AI to diffuse into every industry and every country, every researcher, every student.
And if everything is proprietary, it's hard to do research and it's hard to innovate on top of around with.
And so open source is fundamentally necessary for many industries to join the AI revolution.
NVIDIA has the scale and we have the motives to not only skills, scale, and motivation to build and continue to build these AI models for as long as we shall live.
And so therefore, we ought to do that.
We can open up, we can activate every industry, every researcher, you know, every country to be able to join the AI revolution.
There's a third reason, which is from that to recognizing that AI is not just language.
These AIs will likely use tools and models and sub-agents that were trained on other modalities of information.
Maybe it's biology or chemistry or laws of physics or fluids and thermodynamics.
And not all of it is in language structure.
And so somebody has to go make sure that weather prediction, biology, AI, AI for biology, physical AI, all of that stuff stays, can be pushed to the limits and pushed to the frontier.
We don't build cars, but we want to make sure every car company has access to great models.
We don't discover drugs, but I want to make sure that Lilly has the world's best biology AI systems so that they can go use it for discovering drugs.
And so these three fundamental reasons, both in recognizing that AI is not just language, that AI is really broad, that we want to engage everybody into the world of AI and then also co-design of AI.
You're originally from Taiwan and have a close relationship with TSMC.
So I have to ask TSMC, I think, also is a legendary company in terms of the engineering teams, in terms of the incredible engineering work that they do.
What do you understand about TSMC culture and their approach that explains how they're able to achieve this singular unmatched success in everything they're doing with semiconductors?
You know, first of all, the deepest misunderstanding about TSMC is that Their technology is all they have.
That somehow they have a really great transistor.
And if somebody shows up, another transistor game over.
It's the technology, and of course, I don't mean just the transistor and metallization systems, the packaging, the 3D packaging, the silicon photonics, all of the technology that they have.
That technology is really what makes the company special.
Their technology makes the company special.
But their ability to orchestrate the demands, the dynamic demands of hundreds of companies in the world as they're moving up, shifting out, increasing, decreasing, pushing out, pulling in, changing from customer to customer, way for starting, wafer stopping, emergency way for starts.
You know, all of this dynamics of the world's complexity as the world is shape-shifting all the time.
And somehow they're running a factory with high throughput, high yields, really great cost, excellent customer service.
They take their promises seriously.
When you're wafers, because they know that they're helping you run your company, when the wafers, when the wafers were promised to show up, the wafers show up, you know, so that you could run your company appropriately.
And so their system, their manufacturing system, is completely miraculous.
I would say then the second thing is their culture.
This culture is simultaneously technology-focused on one hand, advancing technology, simultaneously customer service-oriented on the other hand.
A lot of companies are very customer service-oriented, but they're not very technology-excellent.
They're not at the bleeding edge of technology.
A lot of companies who are at the bleeding edge of technology, but they're not the best customer service-oriented company.
And so it just depends on somehow they've balanced these two and they're world-class of both.
And then probably the third thing is the technology that I most value in them that they created is this intangible called trust.
Our single most important property as a company is the installed base of our computing platform.
Our single most important thing today is the installed base of CUDA.
Now, the reason why, 20 years ago, of course, there was no installed base.
But what makes, and if somebody came up with a GUDA or TUDA, it wouldn't make any difference at all.
And the reason for that is because it's never been just about the technology.
The technology, of course, was incredible visionary.
But it's the fact that the company was dedicated to it, stuck with it, expanded its reach.
It wasn't three people that made CUDA successful.
It was 43,000 people that made CUDA successful.
And the several million developers that believed in us, that trusted that we were going to continue to make CUDA 1, 2, 3, 13, that they decided to port and dedicate their software on top of it, their mountain of software on top of it.
And so the installed base is the number one most important advantage.
That install base, when you amplify it with the velocity of our execution at the scale that we're talking about, no company in history had ever built systems of this complexity, period.
And then to build it once a year is impossible.
And that velocity combined with the installed base in the developer's mind is just going to now take the developer's mind.
From the developer's perspective, if I support CUDA, tomorrow it will be 10 times better.
I just have to wait six months on average.
Not only that, if I develop it on CUDA, I reach a few hundred million people, computers.
I'm in every cloud.
I'm in every computer company.
I'm in every single industry.
I'm in every single country.
So if I create an open source package and I put it on CUDA first, I get these both attributes simultaneously.
And not only that, I trust 100% that NVIDIA is going to keep CUDA around and maintain it and improve it and keep optimizing the libraries for as long as they shall live.
You could take that to the bank.
And that last part, trust.
You put all that stuff together.
If I were a developer today, I would target CUDA first.
I would target CUDA most.
And that's the reason that I think in the final analysis is our first, that's even our first core advantage.
Our second one is our ecosystem.
The fact that we vertically integrated this incredibly complex system, but we integrated horizontally into every single company's computers.
We're in the Google Cloud, we're in Amazon, we're in Azure.
We're ramping up AWS like crazy right now.
We're in new companies like Core Weave and NScale.
We're in supercomputers at Lilly.
We're in enterprise computers.
We're at the edge in radio base stations.
It's just crazy.
One architecture is in all these different systems.
We're in cars, we're in robots, we're in satellites.
We're out in space.
And so the fact that you have this one architecture in the ecosystem is so broad.
It basically covers every single industry in the world.
So mentally, you're actually, when you're thinking about a single unit of compute, you're like, literally, when you go to bed at night, you're thinking now about a collection of racks, so pods, not individual chips.
What do you think about the space angle that Elon has talked about doing compute in space for solving some of the it makes some of the energy issues in terms of scaling energy easier?
It's the right place to do a lot of imaging because those satellites have really high-resolution imaging systems and they're sweeping the earth continuously now.
And you want centimeter scale imaging that is done continuously for the world.
So you'll basically have real-time telemetry of everything.
You don't want to beam that back down to Earth.
It's just petabytes and petabytes of data.
You ought to just do AI right there at the edge.
Throw away everything you don't need, you've seen before, didn't change, and then just keep the stuff that you need.
And so AI ought to be done at the edge.
Obviously, we have 24-7 solar if we put it at the polars.
But, you know, there's no conduction, no convection.
And so, you know, you're pretty much just radiation.
And, uh, but, you know, space is big.
I guess we're just going to put big giant radiators out there.
Yeah, there's a lot of low-hanging fruit here on Earth that we can utilize for the AI scaling.
Quick pause.
Quick 30-second thank you to our sponsors.
Check them out in the description.
It really is the best way to support this podcast.
Go to lexfriedman.com/slash sponsors.
We got Perplexity for curiosity-driven knowledge exploration, Shopify for selling stuff online, Element for electrolytes, Finn for customer service AI agents, and Quo for a phone system like calls, texts, contacts for your business.
Choose wisely, my friends.
And now back to my conversation with Jensen Kwong.
Do you think NVIDIA may be worth $10 trillion at some point?
Let's ask it this way: what does the future of the world look like where that's true?
I think that NVIDIA's growth is extremely likely and in my mind, inevitable.
And let me explain why.
We're the largest computer company in history.
That alone should beg the question, why?
And the reason, of course, two reasons.
First, two foundational technical reasons.
The first reason is that computing went from being a retrieval-based file retrieval system.
Almost everything is a file.
We pre-write something, we pre-record something.
We draw something, we put it on the web, we put it in a file, and we use a recommender system, some smart filter to figure out what to retrieve for you.
And so we were a pre-recording, human pre-recording, and file retrieving system.
That's what a computer is, largely.
To now, AI computers are contextually aware, which means that it has to process and generate tokens in real time.
So we went from a retrieval-based computing system to a generative-based computing system.
We're going to need a lot more processing in this new world than in the old world.
We need a lot of storage in the old world.
We need a lot of computation in this new world.
And so that's the first part of it.
We fundamentally changed computing and the way how computing is done.
The only thing that would cause it to go back is if this way of computation, this way of computing generating information that's contextually relevant, situationally aware, that is grounded on new insight before it generates information, this computation-intensive way of doing computing would only go back if it's not effective.
So, for the last 10, 15 years while working on deep learning, if at any single moment I would have come to the conclusion that, you know what, this is not going to work out.
I think this is a dead end, or it's not going to scale, it's not going to solve this modality, not going to be used in this application, then, of course, I would feel very differently about it.
But I think the last five years has given me more confidence than the last 10 years, the previous 10 years.
The second idea is computers, because it was a storage system, it was largely a warehouse.
We're now building factories.
Warehouses don't make much money.
Factories directly correlate with a company's revenues.
And so the computer did two things.
Not only did it change the way it did it, its purpose in the world changed.
It's no longer a computer, it's a factory.
A factory is used for generation of revenues.
We're now seeing not only is this factory generating products, commodities that people want to consume, we're seeing that the commodities are so interesting, so valuable to so many different audiences that the tokens are starting to segment like iPhones.
You have free tokens, you have premium tokens, and you have several tokens in the middle.
And so, intelligence, as it turns out, you know, it's a scalable product.
There's extremely high intelligence products, tokens that are used for specialized things.
People be willing to pay, you know, the idea that somebody's willing to pay $1,000 per million tokens is just around the corner.
It's not if, it's only when.
And so, so now we're seeing that the commodity that this factory makes is actually valuable and is revenue-generating and profit-generating.
Now, the question is: how many of these factories does the world need?
How many tokens does the world need?
And how much is society willing to pay for these tokens?
And what would happen to the world's economy if the productivity were to improve so substantially?
What would happen?
Are we going to discover new drugs, new products, new services?
And so, when you take these things in combination, I am absolutely certain that the world's GDP is going to accelerate in growth.
I'm absolutely certain the percentage of that GDP that will be used for computation will be 100 times more than the past because it's no longer a storage unit, it's a product generation unit.
And so, when you look at it in that context, and then you back into what is NVIDIA's, what does NVIDIA, what does NVIDIA do, and how much of that new economics, new industry would we have to benefit to address?
I think we're going to be a lot lot bigger.
And then, the rest of it to me is: you go, is it possible for NVIDIA to be a $3 trillion revenues company in the near future?
The answer is, of course, yes.
And the reason for that is because it's not limited by any physical limits.
There's nothing that I see that says, you know, gosh, $3 trillion is not possible.
And as it turns out, NVIDIA supply chain is the burden is shared by 200 companies.
And the fact that we scale out on the backs of, with the partnership of this ecosystem, the question is, do we have the energy to do so?
And surely we will have the energy to do so.
And so all of these things combined, that number is just a number, you know?
And I still remember NVIDIA was a, NVIDIA was a, the first time we crossed a billion dollars, I was reminded of a CEO who told me, you know, Jensen, it's theoretically impossible for a fabulous semiconductor company to exceed a billion dollars.
And I won't bore you with why, but of course it's illogical and there's a lot of evidence we're not.
And then somebody told me, you know, Jensen, you'll never be more than $25 billion because of some other company.
Somebody told me that you'll never be, you know, because.
And so those aren't first principle reason thinking.
And the simple way to think about that is what is it that we make and how large is the opportunity that we can create?
Now, NVIDIA is not in the market share business.
Almost everything that I just talked about don't exist.
That's the part that's hard.
You know, if NVIDIA was a $10 billion company trying to take NVIDIA share, then it's easy to see for shareholders that, oh yeah, if they could just take 10% share, they could be this much larger.
But it's hard for people to imagine how large we could be because there's nobody I could take share from.
And so I think that that's one of the challenges for the world is the imagination of the future.
But I got plenty of time and I'll keep reasoning about it and I'll keep talking about it and every single GTC will become more and more real, you know, and then more and more people talk about it.
Yeah, this view of token factories, essentially, this token per second per watt and every token having value.
Like it's an actual thing that brings value and it brings different kinds of value, different amounts of value to different people, but it's value.
That's the actual product that really could be loosely thought of as the token.
And so you have a bunch of token factories and it's very easy, first principles, to imagine a future, given all the potential things that AI can solve, that you're going to need an exponential number more of token factories.
Yeah, there's something truly, as you know, something truly special happening from about December where people really woke up to the power of Cloud Code, of Codex, of OpenClaw.
I mean, I've embarrassed to admit that on the way here in the airport, this is the first time I've done this in public.
I was programming quote unquote by talking to my laptop.
And I was embarrassed because I was pretending like I'm talking to a human colleague.
I'm not sure how I feel about the future where everybody is walking around talking to their AI, but it's such an efficient way to get stuff done.
That's the part that I think most people don't realize is the person who's going to be chatting with them, texting them most, is their claws or lobster.
I read that you attribute a lot of your success to your ability to work harder than anyone and withstand more suffering than anyone.
So we can list many of the things that entails, I mean, dealing with failure, the cost of engineering problems we've talked about, the human problems, uncertainty, responsibility, exhaustion, embarrassment, the near-death company moments that you've mentioned.
But also the pressure.
Now, as the CEO of this company that economies and nations strategize around, plan their financial allocations around, plan their AI infrastructure around.
How do you deal with this much pressure?
What gives you strength given how many nations and peoples depend on you?
I'm conscious about the fact that NVIDIA's success is very important to the United States.
We generate enormous amounts of tax revenues.
We establish technology leadership for our nation.
Technology leadership is important for national security.
National security, not just in one aspect of national security, all aspects of national security.
When our country is more prosperous, we could do a better job with domestic policies and helping social benefits.
Because we're generating so much reindustrialization in the United States, we're creating mountains of jobs.
We're helping shift how we build things back to the United States in so many different plants, chips, computers, and of course these AI factories.
I'm completely aware that, and I have the benefit, and this is a real gift with mainstream investors, teachers, policemen who have somehow, for whatever reason, invested in NVIDIA or because they watched Jim Kramer bought some stock and now are millionaires.
And I am completely aware of that circumstance.
I'm aware of the circumstance that NVIDIA is central to a very large network of ecosystem partners behind us and downstream from us.
And so the way I deal with that is exactly what I just did.
I reason about what is it that we're doing?
What is it causing?
What's the impact that has on other people benefit positively or even through great burden, for example, to supply chain?
And the question is, therefore, what are you going to do about it?
And almost everything that I feel, I break it down.
I reason about, okay, what's the circumstance?
What has changed?
What's hard?
And what am I going to do about it?
And I break it down, decompose the problem.
And the decomposition of these circumstances turns it into manageable things that I can do.
And the only thing that I, after that, I could do is, did you do it?
Did you either do it or did you get somebody else to do it?
And if you didn't do it, you reasoned that you need to do it and you didn't do it and you didn't get anybody else to do it, then stop crying about it.
And so I'm fairly tough on myself.
But I also break things down so that I don't panic.
I can go to sleep because I've made the list of things that needed to be done.
And I've made sure that everything that could put our company in harm's way, could put my partners in harm's way, put our industry in harm's way.
I've told somebody.
Everything that I feel could put anybody in harm's way, I've told someone.
And I've told that someone who could do something about it.
And so I've gotten it off my chest or I'm doing something about it.
But you basically allow yourself to be pulled by the light of the future.
Forget the past and just keep working towards that.
I mean, you did say there's this kind of famous thing you said that if you knew how hard it would be to build NVIDIA, it turned out to be, what is it, a million times more hard than you anticipated, that you wouldn't do it.
That is, by the way, what I was trying to explain is that there's a incredible superpower of being have the mind of a child.
And I say to myself, oftentimes, when I look at something and almost everything, my first thought is, how hard can it be?
And so you get yourself into that mode.
How hard could it be?
And nobody's ever done it.
It looks gigantic.
It's going to cost hundreds of billions of dollars.
It's going to take, you know, all this.
And you just go, yeah, but how hard could it be?
You know, how hard could it be?
And, and so, so you got to get yourself into that state of mind.
You don't want to, you don't want to actually oversimulate everything and all the setbacks and all the trials and tribulations and all the disappointments.
You don't want to simulate all that in advance.
You don't want to know that.
You want to go into a new experience thinking it's going to be perfect.
It's going to be great.
It's going to be incredibly fun.
And then while you're there, you know, you need to have, you need to have endurance.
You need to have grit so that when the setbacks actually happened, and those setbacks are going to surprise you, the disappointments, disappointments are going to surprise you.
You know, the embarrassments are going to surprise you.
The humiliations are going to surprise you.
You just can't.
Now you just got to turn on the other bit, which is just forget about it.
And to the extent that, to the extent that my assumptions about the future and why the future is going to manifest, so long as those assumptions and that input doesn't change or didn't change materially, then I should expect that the output won't change.
And so my simulated output of the future is still going to happen.
And if it's still going to happen, I'm still going to go after it.
I believe it's going to, you know, and so there's a combination of two or three human characteristics.
The ability to go into an experience fresh-minded, the ability to forget the setbacks, the ability to believe in yourself, you know, to believe what you believe and stay true to that belief.
But you're constantly reevaluating.
This combination of three, four, five things, I think is really important for resilience.
And, you know, you're now one of the wealthiest people on earth, one of the most successful humans on earth.
Is it harder to be humble and to be able to, do you feel the effect of money and power and fame in making it harder for you to sort of be wrong in your own head enough to hear out an opinion of somebody else when it disagrees with you and learn from them, those kinds of things?
You have this way about you of when you're explaining stuff, I can feel you actually reasoning on the spot about it with a constant open-mindedness where you could, I could feel like I could steer your thinking.
Even those meetings, knowing that there's people around you where you declared one idea and it was shown that that idea was wrong and be able to admit that and to grow from that, that's very difficult on a human level.
I think their perspective makes sense, and I could see where they're coming from because I don't love AI slop myself.
You know, all of the AI-generated content increasingly looks similar, and they're all beautiful.
And I'm empathetic towards what they're thinking.
That's just not what DLSS 5 is trying to do.
I showed several examples of it.
But DLSS 5 is 3D conditioned, 3D-guided.
It's ground truth, structured data-guided.
And so the artist determined the geometry.
We are completely truthful to the geometry, maintain so in every single frame.
It's conditioned by the textures, the artistry of the artist.
And so every single frame, it enhances, but it doesn't change anything.
Now, the question is, the question about enhancing, DLSS 5 also lets, because the system is open, you could train your own models to determine, and you could even in the future prompt it.
You know, I want it to be a tune shader.
I want it to look like this kind of thing, you know, so you can give it even an example.
And it would generate in the style of that, all consistent with the artistry, you know, the style, the intent of the artist.
And so all of that is done for the artist so that they can create something that is more beautiful, but still in the style that they want.
I think that they got the impression that the games are going to come out the way the games are, shipped the way they do, and then we're going to post-process it.
That's not what DLSS is intended to do.
DLSS is integrated with the artist.
And so it's about giving the artist the tool of AI, the tool of generative AI.
I would say Doom from the intersection of the cultural implication as well as the industry, turning a PC into a gaming device.
That was a very important moment.
Now, of course, flight simulation companies were before it.
But they just didn't have the popularity that Doom did to have made the industry turn the PC from an office automation tool into a personal computer for families and gamers and things like that.
And so Doom was really impactful there.
From an actual game technology perspective, I would say Virtual Fighter.
I think accurately that the AGI timeline question rests on your definition of AGI.
So let me ask you about possible timelines here.
Let's this ridiculous definition perhaps of what AGI is, but an AI system that's able to essentially do your job.
So run, no, start, grow, and run a successful technology company that's worth a good one or a one?
No, it has to be worth more than a billion, more, more than a billion dollars.
So, you know, you know how hard it is to do all of those components.
So how far are we away from that?
So we're talking about OpenClaw that does all the incredibly complex stuff that are required to, first of all, innovate, to find customers, to sell to them, to manage, to build a team of some agents, some humans, all that kind of stuff.
And the reason for that is this: you said a billion, and you didn't say forever.
And so, for example, it is not out of the question that a claw was able to create a web service, some interesting little app that all of a sudden, you know, a few billion people used for 50 cents.
And then it went out of business again shortly after.
Now, we saw a whole bunch of those type of companies during the internet era.
And most of those websites were not anything more sophisticated than what OpenCloud could generate today.
Well, by the way, it's happening right now, right?
You know that when you go to China, You're going to see a whole bunch of people teaching their getting their clause to try to go out and look for jobs and do work, make money.
And I'm not actually, I wouldn't be surprised if some social thing happened or somebody created a digital influencer, super, super cute, or some social application that feeds your little Tomagotchi or something like that, and it becomes an out of the blue, an instant success.
A lot of people use it for a couple of months and it kind of dies away.
Now, the odds of 100,000 of those agents building NVIDIA is 0%.
And then the one part that I won't do, and I want to make sure we all do, is to recognize that people are really worried about their jobs.
And I just want to remind them that the purpose of your job and the tasks and the tools that you use to do your job are related, not the same.
I've been doing my job for 33 years.
I'm the longest running tech CEO in the world, 34 years.
And the tools that I've used to do my job has changed continuously in the last 34 years and sometimes quite dramatically, you know, over the course of a couple, two, three years.
And the one story that I really want to make sure that everybody hears is the story, the first job that computer scientists said, AI researchers said was going to go away was radiology because computer vision was going to achieve superhuman levels.
And it did.
Computer vision was superhuman in 2019, maybe a little bit later, 2020.
And the people that are currently programmers and software engineers, I think they're at the cutting edge of understanding intuitively how to communicate with agents using natural language in order to design the best kind of software.
So over time, they'll converge, but I think there's still value in getting, I think, learning how to program, like learning what programming languages are, the old kind of programming.
What are good practices for programming languages?
What are design principles for programming languages for large software systems?
And the reason for that, Lex, you know, I was just saying for the audience, I think the goal of specification, the artistry of specification, the goal and the artistry of it, is going to depend on what problem you're trying to solve.
When I'm thinking about giving the company strategies and formulating corporate directions and things that we should do, I describe it at a level that is sufficiently specific that people generally understand the direction and it's actionable.
It's so specific enough that they can take action on it.
But I underspecify it on purpose so that enable 43,000 amazing people to make it even better than I imagined.
And so when I'm working with engineers, when I'm working with people, I think about what problem am I trying to solve?
Who am I working with?
And the level of specification, the level of architecture definition relates to that.
And so everybody's going to have to learn where in the spectrum of coding they want to be.
Writing a specification is coding.
And so you might decide to be quite prescriptive because there's a very specific outcome you're looking for.
You might decide that this is an area you want to be much more exploratory.
And so you might underspecify and enable you to go back and forth with the AI to even push your own boundaries of creativity.
And so this artistry of where you are in the spectrum, this is the future of coding.
But just to linger on it outside of coding, I think a lot of people, rightfully so, Are worried about their jobs, have a lot of anxiety about their jobs, especially in the white-collar sector.
Um, I don't think any of us know what to do with tumultuous times that always come when automations and new technology arise.
And I just, first of all, I think we all need to have compassion and the responsibility to feel sort of the burden of what the actual suffering feels like for individual people and families that lose their job.
I think whenever you have transformative technology like that's coming with artificial intelligence, there's going to be a lot of pain.
And I don't know what to do about that pain.
Hopefully, it creates much more opportunities for those same people for the same kind of job as the tooling evolves and makes them more productive and makes them more fun.
Hopefully, as it does in the programming, I've been having so much fun programming, I have to say.
Like, I've never had this much fun.
So, hopefully, it makes their job automates the boring parts and makes the creative parts the ones that the human beings are responsible for.
But still, there's going to be a lot of pain and suffering.
So, my first recommendation before, and this is now how I deal with anxiety.
In fact, we just talked about it earlier: enormous anxiety about the future, enormous anxiety about the pressure, enormous anxiety about uncertainty.
I first break it down, and then I'm going to tell myself, okay, there are some things you can do something about, there's some things you can't do anything about, but for the stuff that you can do something about, let's reason about it and let's go do it.
If we were to hire a new college graduate today, and I have a choice between two, one that has no clue what AI is, and one that is expert in using AI, I would hire the one who's expert in using AI.
If I had an accountant, a marketing person, the one that is expert in using AI, supply chain, customer service, a salesperson, business development, a lawyer, I would hire the one who is expert in using AI.
And so, I would advise that every college student, every teacher should encourage their student to go use AI.
Every college student should graduate and be an expert in AI.
And everybody, if you're a carpenter, if you're an electrician, go use AI.
Go see what it can do to transform your current job.
Elevate yourself.
If I were a farmer, I would absolutely use AI.
If I were a pharmacist, I would use AI.
I want to see what it could do to elevate my job so that I could be the innovator to revolutionize this industry myself.
And so that would be the first thing that I would do.
And then I would also help them.
It is the case that the technology will dislocate and will eliminate many tasks because it will automate it.
If your job is the task, if your job is the task, then you're very highly going to be disrupted.
If your job's purpose includes certain tasks, then it's vital that you go learn how to use AI to automate those tasks.
And then there's the world of spectrum in between.
And by the way, the beautiful thing about AI, so the chatbot versions, is you can break down, you have anxiety and you can break down the problem by talking to it.
Like I've recently, it's really just incredible how much you can think through your life's problems and through, and I don't mean like therapy problems.
I mean like very practically, okay, I'm worried about my, literally, I'm worried about my job.
What are the skills?
What are the steps I need to take?
How do I get better at AI?
Everything you just said, you can literally ask and it's going to give you a point-by-point plan.
And that's the, you know, of course, the conditions by which that causes anxiety or nervousness or whatever emotion.
I believe that AI will be able to recognize those and understand those.
I don't think my chips will feel those.
And therefore, how that anxiety, how that feeling, how that excitement, how that, how that, you know, all of those feelings manifest in human performance.
For example, extremely amazing human performance, athletic performance, you know, average or lesser than average.
That entire spectrum of human performance that comes out of exactly the same circumstances for different people, manifesting in different outcome, manifesting in different performance.
I don't think there's anything about anything that we're building that would suggest that two different computers being presented with all of exactly the same context would perform, of course, it would produce statistically different outcomes, but it's not because it felt different.
Yeah, the subjective, boy, there's something truly special about the subjective experience that we humans feel.
Like I mentioned to you, I was pretty nervous talking to you.
Like I mentioned to you that the hope, the fear, the anxiety, and just life itself, the richness of life, how amazing everything is, how deeply we fall in love, how deeply our hearts get broken, how afraid we are of death, and how much pain we feel when our loved ones pass away.
All of that, the whole thing.
It's very hard to think AI being able to, a computational device being able to do that.
But there's so many mysteries about this whole thing that we're yet to uncover that I am open to be surprised.
I've been surprised a lot over the past few months and few years.
Scaling can create some incredible miracles in the space of intelligence has been truly marvelous to watch.
And it's just really important to break down what is intelligence.
And the word, that word we use all the time, it's not a mysterious word.
Intelligence has a meaning, you know, and it's a system that, you know, it's something that we do that includes perception and understanding and reasoning and the ability to do plan.
And, you know, that loop, that loop is fundamentally what intelligence is.
Intelligence is not one word that is exactly equal to humanity.
And that's, I think, really important to separate the two.
We have two words for that.
I'm not, I don't over-fantasize about, and I don't over-romanticize about intelligence.
Intelligence is, and people have heard me say it before.
And I'm surrounded by intelligent people more intelligent than I am in each one of the spaces that they're in.
And yet, I have a role in that circle.
It's actually kind of interesting.
They're more educated than I am.
They went to better schools than I did.
They're deeper than in any fields that they're in.
All of them.
I have 60 of them.
They're all superhuman to me.
And somehow I'm sitting in the middle, orchestrating all 60 of them.
And so you got to ask yourself: what is it about a dishwasher that allows that dishwasher to sit in the middle of superhumans?
Does that make sense?
And so, but that's my point.
My point is: intelligence is a functional thing.
Humanity is not specified functionally.
It's a much, much bigger word.
And our life experience, our tolerance for pain, our determination, those are different words in intelligence.
And so the thing that I want to help the audience understand, if I could give them one thing, is intelligence is a word that we've elevated to very high form over time.
Compassion, generosity, all of the things that you say just now, I believe those are superhuman powers.
And that now intelligence is going to be commoditized because we've spoken about it.
The most important thing is your education.
Now, even when they said the most important thing is your education, when you went to school, there's more than just knowledge that you gained.
And so, but unfortunately, our society had put everything into one single word.
And life is more than one word.
And I'm just telling you, my life would suggest that being lower on the intelligence curve than everybody around me doesn't change the fact I'm the most successful.
And so, and I think, I think that kind of is, I'm trying hopefully to inspire everybody else.
That don't let this democratization of intelligence, this commoditization of intelligence, you know, cause you anxiety.
This is not a once-in-a-once in a lifetime experience suggests that it has been experienced by many people, just not one person.
This is a once-in-a-humanity experience, what I'm going through.
NVIDIA is one of the most consequential technology companies in history.
We're doing very important work.
I take it very seriously.
And so some of the things that, of course, are practical things, like how do we think about succession planning?
And I'm famous in saying that I don't believe in succession planning.
And the reason for that isn't because I'm immortal.
The reason for that is because if you're worried about succession planning, if you're worried, all that anxiety of succession planning, then what should you do about it?
Then you break it all the way back down.
The most important thing you should do today, if you care about the future of your company, post-you, is to pass on knowledge, information, insight, skills, experience as often and continuously as you can, which is the reason why I continuously reason about everything in front of my team.
Every single meeting is about a reasoning meeting.
Every moment I spend inside a company, outside the company, is about passing on knowledge to people as fast as I can.
Nothing I learn ever sits on my desk longer than a fraction of a second.
I'm passing that information, that knowledge.
Oh my gosh, this is cool.
Before I even finish learning all of it myself, I've already pointed it to somebody else.
Get on this.
This is so cool.
You're going to want to learn this.
And so I'm constantly passing knowledge, empowering people, elevating the capability of everybody around me so that the outcome that I seek, that I hope for, is that I die on the job.
And also, if I could just say one more thing, and thank you for all the interviews that you do, the depth, the respect that you go through with, and the research that you do to reveal, you know, for all of us, the amazing people that you've interviewed over the years.
I've enjoyed them immensely.
And as an innovator, to have created this long-form, unbelievable, and yet, you know, it's just captivating.
Thank you for listening to this conversation with Jensen Kwong.
To support this podcast, please check out our sponsors in the description, where you can also find links to contact me, ask questions, give feedback, and so on.
And now, let me leave you with the words from Alan Kay: The best way to predict the future is to invent it.