NVIDIA CEO Jensen Huang 01:29:47

i will say David， i would love to have nvidias full production team every episode it was nice not having to worry about turning the cameras on and off and making sure that nothing bad happened myself！

while we were recording this yeah just the gear i mean the drives that came out of the camera alright！

uh red cameras for the homestudio starting next episode yeah good all right lets do it， citydown say is straight another storyon welcome to this episode of acquiredthepodcast about great technology companies and the stories and playbooks behind them im Bengilbert！

um！

David Rosenthall and we are your hosts listeners just so we dont Barry the lead this episode was insanely cool for David， and i yeah after researching Nvidia for something like five hundred hours over the last two years we flew down to Nvidia headquarters to sit down with Jenson himself and Jenson of course， is the founder in CEO of Nvidia the company powering this whole AI explosion at the timer for recording， Nvidia is worth one point， one trillian dollars and is the sixth most valuable company in the entire world and right now is a crucible moment for the company expectations are set high i mean sky high they have about the most impressive strategic position and lead against their competitors of any company that weve ever studied but heres the question that everyone is wondering will nvideos insane prosperity continue for years to come is ai gonna be the next trillian dollar technology wave how sure are we of that and if so can Nvidia actually maintain their ridiculous dominance as this market comes to take shape so jenson takes us down memory lane with stories of how they went from graphics to the data center to ai how they survived multiple near death experiences he also has plenty of advice for founders， and he shared an emotional side to the founder journey toward the end of the episode yeah！

i got new perspective on the company and on him as a founder and a leader just from doing this despite you know we thought we knew everything before we came in advance and uh it turned out we didnt turns out the protagonist actually knows more yes？

or i olistners join the slack there is incredible discussion of everything about this company ai the whole Lego system and a bunch of other episodes that wedone recently going on in there right now so that is acquired dot fm slash slack we would love to see you and without further do this show is not investment advice Steven and i may have investments in the companies we discuss and the show is for informational and entertainment purposes only onto jenson so jenson this is acquired so we want to start with storytime so we want to win the clock all the way back to i believe it was nineteen 97 youre getting ready to ship the reva one twenty 8， which is one of the largest graphics chips ever created in the history of computing it is the first fully threed accelerated graphics pipeline for a computer yeah and you guys have your running at month of cache left and so you decide to do the entire testing in simulation rather than ever receiving a physical prototype you commission the production run site on scene with the rest of the companies money yeah bedding it all right here on the river one twenty 8 no comes back and of the 32 directx blend modes it supports aid of them and you have to convince the market to buy it and you get a convince developers not to use anything but those ate blend modes walk us through what that the other twenty four werenthat important OK？

so we we first was that the plan all a lot like when when did you realize， we should have？

which what i realized i didnt learn about it until was too late we should have implemented all three two yeah but but it we built we built and so we had to make the best of it that was really an expegnary time remember when twenty was mb， three NV1， NV2， erebased on forward texture mapping no triangles but curves and teslated the curves and because we were rendering higher level objects we essentially avoid using z buffers and we thought that that was going to be a good rendering approach and turns out to them completed the wrong answer and so what revote run？

28 was was a reset of our company now remember at the time that we started the company three we were the only consumer three d graphics company ever created and we we were focused on transforming the pc and to an accelerate pc because at the time， windows was really a software rendered system and so anyways river one twenty eight was a reset of our company because by the time that we realize we had gone down the wrong road Microsoft had already rolled out direct x it was fundamentally in compatible with Nvidia Architecture， thirty competitors have already shown up uh even though we were the first company at the time that we were found it so the world was a completed different place the question about what to do as a company strategy at that point i would say that we made a whole bunch of wrong decisions， but on that day that mattered we made a sequence of extraordinarily good decisions and that time nineteen 97 was probably nvidias best moment and the reason for that was our backswerup against the wall we were running out of time， were running out of money and for a lot of them plays running out of hope and the question is what do we do well the first thing that we did was we decided that love direct access now here when i got to fight， it lets go figure out a way to build the best thing in the world for it and reval one 28 is the worldsfirst fully accelerated hardware accelerated pipeline for the renderine 3D and so transform the projection every single element all the way down to the frame buffer was completely hard works already we implemated out a a texture cache we took the bus limit the frame buffer limit to as big as as a physics could afford a time we made the biggest chip that anybody had ever imagine building we used the fasses memories basically， if we build that chip， there could be nothing that could be faster and we also chose a costpoint that is substantially higher than the highest price that we think that any of our competors will be willing to go if we build it right， we acsoliated everything we implement， everything uh in direct x that we knew of and we build it as large as we possibly could then obviously nobody can build something passer than that today in a way you kind of do that here at Nvidia two you were a consumer products company back then right？

there was end， consumer who were gonna have to pay the money to buy thats right？

but we observe that there was a segment of the market where people were because at the time that the pc industry was still coming up and it wasngood enough everybody was clamoring for the next fastest thing and so， if your performance was ten times higher this year， then what was available？

theres a whole large market of enthusiaswho who we believe would would have gone after it and we were absolutely right that the pc industry had a substantially large enthusiast market that would by the best of everything to this day is kind of remains true and for certain segments of market word technology is never good enough like three d graphics when we chose the right technology 3D graphics is never good enough and we call it back that threed gives us sustainable technology opportunity because its never good enough answer your technology can keep getting better we chose that uh we also meet the decision to use this technology called emulation that was a company called icos and on the day that i call them they were just shutting the company down because they had no customers and i said hey， look， uh， ill buy what you have inventory and uh you know uh no promises are necessary and the reason why we needed that emulator is because if you figure out how much money that we have we taped out a chip and we got a back from the fab and we started working on our software by the time that we found all the bugs because we did the software， then we take down the chip again well， we would have been out of business already yeah and so i know your competitor is would cut up well not to mention we would have been out of business or sense who cares of exactly， they serve you gonna be out of business anyways that plan obviously wasnt plan the plan that companies normally go through， which is you know build a trip right the software fix the bugs tab out a new trip so unso forth that method wasnt work and so the question is if we only had six months and you get the tape out just one time， then obviously you gonna take out a perfect trip so i so i remember having conversation with their leaders and they said but just how do you know is gonna be perfect as it， i know is gonna be perfect because of its not will be out of business and so lets make a perfect we get one shot we essentially virtually prototype the chat by buying this emulator and guide in the solver team road are software the entire stack and ran in on this emulator and just setting the lad waiting for windows to paint you know and it was like it sixty thousand for a frame result easily， i actually think that was an hour perframe something like that and so we resist sit there and watch a pain and so on the day that we decided to take out i assume that the trip was perfect and everything that that we could have tested we tested in advance and told everybody this is that we want to take out the trip is going to be perfect well， if you are going to take out a trip and you know whats perfect then what else would you do thats actually， the good question if you knew that you hit enter， you take that a chip and you knew was going to be perfect then what else would you do while the answer obviously go to production and marketing blits yeah！

yeah and developer kick everything off！

take everything off because you got a perfect chap and so we got in our head that we have a perfect chap how much this was you and how much of this was like your co founders？

the rest the company in the board was everybody telling you you were crazy！

no everyway was clear we had no shot they did not doing it would be crazy because the otherwise you want yeah youre gonna deal out this the same anyways so anything aside from that is crazy so it seemed like a fairly logical thing and quite frank it right now to some describing it every year prior thinking yeah its pretty sensible well worked yeah and so we take that out when directly to production so is the lesson for founders out there when you have conviction on something like the revo one twenty eight or ub could a go back the company on it and this keeps working for you so it seems like your lesson learn from this is yes keep pushing all the chips in because so far its working every time no how do you think about that？

no， no when you push your chips and um i i know its gonna work notice we assume that we taped out a perfect trip the reason why we taped out a perfect trip this because we emuated the whole trip before we take it out we develop the entire software stack we ran qa on all the drivers and all the software we ran all the games we had we ran every vga application we had and so when you push your chips and what youre really doing is your when you get the form， youre saying im going to take everything in the future， all the risky things on i pull in advance and that is probably the lesson and to this day everything that we can prefetch everything in the future that we can simulate today uh we prefetch it and we talk about this a lot we just talk about this on our cosko episode you want to push your chips in when you know its gonna work so every time we see you make a better the company move you already simulated it you know yeah yeah yeah do you feel like that was the case with cuda？

uh， yeah in fact before there was cuda， there was a cg right and so we were already playing with a concept of how do we create an abstrationlayer above our chip that is expressible in a higher level language and higher level expression in and how can we use our gpu for uh things like ctreconstruction image processing？

we were already down that path， and so there were some positive feedback and something to id of positive feedback that that we think that that general purpose computing could be possible and if just look that the pipeline of a programle shader， it is a processor and is a highly parallone， it is a massibly threaded， and it is the only processor in the world that does that， and so there are a lot of characteristics about programle shading that would suggest that CUDA has a great opportunity to succeed。

and that is true if there was a large market of machine learning practitioners， who would eventually show up and want to do all this great scientific computing and accelerated computing but at the time when you were starting to invest， what is now something like ten thousand percent years， no in building that platform， no did you ever feel like oh， man we might have invested a head of the demand for machine learning since were like a decade before the whole world is realizing it i guess yes and no you know when we saw deep learning！

when we saw alexnet and realized its incredible effectiveness and computer vision， we had the good sense if you will to go back the first principles announce you know what is it about this thing that made it so successful when a new software technology， your new algorithm classalone and somehow leap frogs thirty years， a computer vision work you have to take a step back and ask yourself， but why and fundamentally is is scalable and of its scalable what other problems can it solve and there were several observations and we made in the first observation of course， is that if you have a whole lot of example data， you could teach this function to make predictions well， what we basically done is discovered a universal function approxmatter because the dimensionality could be as high， she wanted to be， and because each layer is trained one layer at a time， theres no reason why you cant make very very deep neural networks OK， so now you just reason your way through right， okay， so now i go back to twelve years ago and you could just imagine the reasoning im going through my head that weve discovered the universal functional proxmatter in fact， we might have discovered with a couple more technologies and universal computer the and you continue attention to the image net competition yeah！

yeah， leading up to this yeah， yeah！

and the reason for that is because we were already working on computer vision at the time and we were trying to get CUDA to be a good computer vision system or mostly algorithms are were creative for computer vision arent a good fit for CUDA and so we were sender trying to figure it out of us and Alex net shows up and so that was incredibly intriying its so effective that it makes us take us that back and ask yourself wiseof happening so by time that you reason you wait through this you you go well， what are the kind of problems in a world where universal functional proximator？

yeah， pistons right well， we know that most of our algorithms start from principled sciences OK， you want to understand the causality and from the causality you create a simulation algorithm that allows us the scale well， for a lot of problems we kind of dont care about the causeality we just care about the predictability of it like do i really care for what reason you prefer this to the paced over that i donre really care the causality i just want to know that this is the one you were a predicted do i really care that the fundamental cause of somebody who buys a high dog， buys catch up a mustor it doesnt really matter it only matters that i can predict it it applies to predict the movies， predicting music it applies to predicting quite frankly whether we understand thermal dynamics， we understand radiation from the sun， we understand cloud effects， we understand ocean and complex， we understand all these different things we just want to know whether we should were switter or not isthat right yeah！

its a causality for a lot of problems in the world doesnt matter we just want to emulate the system and predict the outcome in it can be an incredibly lookative markets if you can predict what the next best performing uh feed item to serve into a social media feed turns out thats a huge if i was gonna go with that love the examples people two paste catch up existing movies when you realize as you realize hang hang on second。

a universal functional proxmatter a machine learning system you know something that learning from examples could have tremendous opportunities because just the number of applications is quite enormous and everything from obviously we use that talk about commerce， autoway， decides and so you realize that maybe this could affect a very large part of the worlds industries almost every piece of softwin a world would eventually be programved this way and if thats the case then how you build a computer and how you build a trip in fact can be completely changed and realizing that the rest of it is just comes what you know？

do you have the courage to put your chips behind it so thats where we are today um and thats where Nvidia is today but im curious in the you know theres couple years after alexnet and this is when Benna， i were getting into the technology industry in the venture industry ourselves i started at Microsoft in 22。

yeah， yeah， so right after alexnap。

but before anyone was talking about machine learning and even the main stream engineering community there were those couple years there where to a lot of the rest of the world these looked like science projects yeah， the technology companies here in silicon valley， particularly the social media companies they were just realizing huge economic value out of this the Googles， the Facebook did flex， yes， etc yeah and obviously that led a lot of things including open ai a couple years later， yeah， but during those couple years when you saw just that huge economic value unlockhere in silicon value how are you feeling during those times？

the first thought was of course reasoning about how we we should change our computing stack the second thought is where can we find earliest possibilities of use if we were to go build this computer， what would people use it to do and we were fortunate that working with the worlds？

universities and researchers was was innate in our company because we were already working on CUDA and CUDA early adopters were researchers because we democratize supercomputing you know， CUDA is not just used as you know for AI CUDA is used for almost all fields of science everything from the regular dynamics to imaging ctreconstruction to um a sysmake processing to you know whether simulations quantum chemistry the list goes on right and so the number of applications of CUDA in research was very high and so when a time came and we realized a deep learning could be really interesting it was natural for us to go back to the researchers and find every single ai research your on the planet and say how can we help you advance your work and that included yangle can and Andrew Ang and jev hinton and thats how i met all these people and and i used to go to all the ai conferences and thats where you know i met Alia susk over there for the first time yeah and so was really about at that point where the systems are we can build the softest we can build to help you be more successful to advance to research because at the time it look like a toy， but we had confidence that even gang the first time i made good fellow the gan was as like 32 by 32 and it was just the you know blurry image of a cat you know but how far can i go and so we believed in it we believed it one you could scale deep learning because obviously its trained layer by layer， and you could make the data sets larger and you could make the models larger and we believe that if you made that larger and larger if we get better and better， yeah， kind of sensible and i think the discussions in the engagement with researchers was the exact positive feedback system that we needed how would go back to research it was bouts where it all happened when open ai was founded in twenty out 15 yeah！

we now is such a an important moment thats obvised today now but at the time， i i think most people even people in tech for like what is this yeah， yeah， we were you involved in it， it all like you know because you were so connected to the researches to alia taking that talent out of Google Facebook to me point， but yeah reseeding the research community yeah and opening it up um was such an important moment were you involved in it at all i wasnt involved in the founding of it but i knew a lot of the people there and um a inline of course。

i i knew and a Peter beale was there and illion was there and 啊 we have we have some great employees today that were there in the beginning and i knew that they needed the amazing computer that we were building and were building the first version of the dgx which you know today when you see a hopper its 70 pounds 35 thousand parts， ten thousand amps， but gtx the first version that we built was a used internally and i deliver the first one to open ai yeah that was a fun day， but most of our success was aligned around in the beginning adjust about helping the researchers get to the next level i knew it wasnt very useful in its current state， but i also believe that in a few clicks it could be really remarkable， and that believes system came from the interactions with all these amazing researchers， and it came from just seeing the incremental progress at first， the papers were coming out every three months and then then papers today are coming out every day right， so you could just monitor the archive papers and i took an interest in learning about the progress of deep learning and to the best mibility read these papers and you could just see that progress happening you know in real time exponentially in real time it even seems like within the industry from some researchers。

we spoke with it seemed like no one predicted how useful language models would become when you just increase the size the models they thought oh， there has to be some algorithmic change that needs to happen but once you cross that tenbillion parameter market certainly once you cross the hundredbillion， they just magically got， much more accurate， much more useful！

much more lifelike were you shocked by that the first time you saw a truly large language model and do remember that feeling well my first feeling about the language model was how clever it was to just mask out words and and make a predict the next word its self supervised learning at its best we have all this text you know， i know what the answer is i just make you guess it and so my first impression of Bert was really how clever was and now the question is how can you scale that you know？

the first observation on almost everything is interesting and then and then try to understand too itively white words and then the next step of courses from first principles how would you extrapolate that yeah and so obviously we knew that Bert was going to be a lot larger now one of the things about these line， which models is its encoding information isnthat right is compressing information and so within the world languages on text theres a fair amount of reasoning thats encoded in it and we describe a lot of reasoning things and answer you were to say that a few step reasoning is somehow learnable from just reading things i wouldnbe surprised no you know for all a lot of us we get our common sense and we get our our reasoning ability by reading， and so why when they machine learning model also learning some other reasoning capabilities from that and from reasoning capabilities， you could have emergent capabilities right immersionabilities are consistent within tuitively from reasoning and so some of it could be predictable but still itstill amazing the fact that its sensible doesnt make it anyless amazing right i could visualize literally the entire computer um and an all the boot modules in a self driving car and the fact that is still keeping lanes makes me insanely happy and so i even remember that from my first operating systems class in college when i finally figured out all the way from programming language to the electrical engineering classes bridged in the middle by that os class im like oh！

i think i understand how the volumen computer works soup to nuts and its still a miracle yeah！

yeah！

yeah！

exactly yeah！

yeah when you put all together is still a miracle yeah now is a great time to talk about one of our favorite companies static seg and we have some tech history for you yes！

so in our Nvidia part three episode we talked about how the AI research teams a Google and Facebook job incredible business outcomes with cuttingedge ml models and these models powered features like the Facebook newsfeed googleads and the YouTube next video recommendation in the process transforming Google in Facebook into the juggernots that we know today and while we talked all about the research， we didnt touch on how these models were actually deployed yeah！

the most common way to deploy new models was through expermentation ab testing when the research team created a new model， product engineers would deploy the model to a subset of users and measure the impact of the model on core product metrics great experimentation tools transformed the machine learning development process they d risk releases since each model could be released to a small set of users theyspetupreleasecycles researchers could suddenly get quick feedback from realuserdata， and most importantly they created a pragmatic data driven culture since researchers were rewarded for driving actual product improvements and over time these experimentation tools gave Facebook in Google a huge edge because they really became a requirement for leading ml teams yep so now youprobably thinking well。

thats great for Facebook in Google， but my team cant build out our own internal experimentation platform well， you dont have to thanks to static， so static was literally founded bytexfacebook engineers who did all this theyve built a best in class experimentation feature flagging and product analyxed platform thats available to anyone and surprise surprise a ton of ai companies are now using stat seg to improve and deploy their models including and tropic yep so whether youbuilding with ai or not stat signcan help your teamship faster and make better data driven product decisions they have a very generous free tier and a special program for ventureback companies。

simple pricing for enterprises and no seatbased fees if youre in the acquiredcommunity， theres a special offer you get five million free events a month and white glove onboarding supportsovisitcom slash acquired and get started on your data driven journey we are some questions we want to ask you uh summer cultural about Nvidia。

but um others are generalizable to company building broadly and the first one that we wanted to ask is uh we heard that you have forty plus direct reports and that this org chart works a lot differently than a traditional company org chart do you think theres something special about Nvidia that makes you able to have so many direct reports not worry about kotling or focusing on career growth of your executives and youlike no youjust here to do your freakin best work and the most important thing in the world now go a is that correct and b is there something special about Nvidia that enables that i dont think its something special one video i think that we had the courage to build a system like this and video is not build like a military is not build like a like the arm forces where you have you know generals and kernels you you we just were not set up like that were not set up in a commanding control and information distribution system from the top down were really built much more like a computing stack and a computing stack the lower slayer is our architecture and then theres our chip and then theres our software and and on top of it there are all these different modules and each one of these layers of modules are people and so the architecture of the company to me is a computer with a computing stack with um uh people managing different parts of the system and who reports to whom your titles not related to anywhere you are in the stack it just happens to be whois the best that running that module on that function on that layer it is in charge and that person is the pilot in command thats so thats one characteristic and have you always thought about the company this way even from the earliest days yeah， pretty much yeah and the reason for that is because your organization should be the architecture of the machinery of building the product right yeah， thats what a company is yeah and yet everybodys company looking exactly the same but they all feel different things hows i make any sense you do you see all the same yeah you know？

how you make four i chicken versus how you feel burgers versus？

how you make you know Chinese fragrice is different and so， why would the machinery？

why would the process be exactly the same and so its not sensible to me that if you look at the art charts of most companies， it all kind of looks like this and then that you have one group thats for your business and you have another for another business you have another for another business and theyre all kind of supposedly autonomous and so none of that stuff makes any sense to me just depends on what is that that were trying to build and what is the architecture the company that best suits to go build it thats so thats number one in terms of information system and how do you enable collaboration we kind of wiredup like a neural network and the way that we say is the theres a phrase in the company called mission is the boss and so we figure out what is the mission of what is the mission and we go wire up the best skills and the best teams and best resources to achieve that mission and it cuts across the entire organization in a way that doesnt make any sense but its looks like a little bit like a year or network you know when you see mission daming mission like nvidias mission hopper yeah OK！

so its not like further accelerated computing gets like worshipping DJs cloud uh build hopper or somebody else is uh build a system for hopper？

somebody is uh build cuda for hopper， somebodys job is build cudy and for cuda for hopper， somebodys job is the mission right is is so you know your mission is to do something what are？

the tradeoffs associated with that versus the traditional structure the downside is the pressure on the leaders is fairly high and the reason for that is because in a commanding control system， the person who you reports to have more power than you and the reason why they have more power than you is because theyre closer to the source of information that you are hmm in our company the information is dissiminated fairly quickly to a lot of different people and usually add a team level so for example， just now wasn i was in our robotics meeting and were talking about certain things and were making some decisions and there are new college grads in a room theres three visepresidents in a room theres two east ops in a room and at the moment that we decide together we reason through some stuff we made a decision everybody hearted is exit exactly the same time so nobody has more power than anybody else hmm does it make sense the new college grad learn at exact in the same time as the staff and so the the executive staff and the leaders at that work for me and myself you earned the right to have your job based on your ability to reason through problems and helping other people succeed and and its not because you have some privileged information that i knew the answer was three point seven and only i knew you know where everybody knew when we did our most reason episode in video part three that we？

we just released we started to this thought exercise um especially over the last couple years， your product shipping cycle has been very impressive， especially given the level of technology that you are working with and the difficulty of this all we sort said like could you imagine apple shipping to iPhones a year？

and we said for illustrative as for illustrative purposes not pick on apply what a large tech company？

a large set two flagship products or their flagship product twice per year yeah or you know two wdcc year yeah there seems to be something can like you cant really imagine that whereas that happens here are there other companies either current or historically that you lookuptoadmire maybe took some of this inspiration from in the last thirty years ive read my fair share of business books and as in everything you read you youre supposed to yous supposed to first all enjoy it right enjoy it be inspired by it but not to adopt it thats not the whole point of these books the whole point of these books is to share their experiences and and you youre supposed to ask you know what does it mean to me？

in my world and what is it mean to me？

in the context， what ive going through what is this mean to me？

in the environment that im in what is this mean to me and what im trying to achieve even what is us mean to advance in the age of our company in the capabilities of our company and so yours wai ask yourself what is a mean to you and then from that point being informed by all these different things that were learning uh Wes supposed to come up with our own strategies you know what i just described this kind of how i go about everything yous supposed me inspired and learn from every everybody else and and the education is free you know when somebody talks about a new product， yous supposed to go listen to yous supposed to ignore it you supposed to learn from it and i could be a competitor， could be a json industry it could be nothing to do with us now the more where we learn from uh was happening on the world of the better but then you use supposed to come back and ask yourself you know what does this mean to us yeah you dont just want to imitate thats right yeah！

i love this team up of learning， but not imitating and learning from a wide array of sources theres this sort of um unbelievable third element i think to what Nvidia is become today and thats the data center its certainly not obviously i cant reason from alexnet and your engagement with the research community and and you know social media feedback matters to yeah， you deciding and the company deciding all weve gonna go in a fiveyear， all in journey on the data center yeah， yeah， how did that happen？

yeah！

our journey to the data center happened i would say almost years ago im always being asked in me what what are the challges that the company could see some day and and ive always felt that the fact that nvidias technology is plugged into a computer and that computer has the sit next to you because it has to be connected to a monitor that will limit our opportunity someday， because there are only so many desktop pcs that plugagp you into， and theres only so many crts and and the time lcds that we could possibly drive so the question is when it be amazing if our computer doesnt have to be connected to the viewing device that that the separation of it um made a possible for us to compute somewherelse and one of our engineers came a showed to me one day and it was really capturing the frame buffer encoding it into video and streaming it um to a receiver device separating computing from the viewing in many ways of the cloud Gaming game double in fact double when we started GFM， we knew that GFN was going to be um a journey that would take a long time because you youre fighting！

youre fighting all kinds of problems including including the speed of light and latency everywhere you look thats right for listen GFN g force now！

g force now！

yeah yeah！

g force now and and we were working on all force as your first cloud product thats right now and and look at look at g force now was nvidias first data center product and our second data center product was remote graphics putting our gps in in the worlds enterprise data centers， which then led us to our third product， which combined could apples rgpu， which became a super computer， which then work towards you know more， more， more and the reason why so important is because the disconnection between where nvidias uh computing is done versus where is enjoyed if you can separate that your market opportunity explodes yeah， yeah and was completely true and so were no longer limited by the physical constraints of the Desktop pc sitting by your desk um you know and and were not limited by one gpu per person and and so it doesnmatter where it is anymore and so that was really the great observation its a good reminder i you know the data center segment of nvidias business to me has become synoms with a howsai going and thats a false equivalence and its interesting that you were only this ready to sort of explode in ai in the data center because you had three plus previous products where you learned had a build data set computer exactly even though those markets werent these like gigantic world changing technology shifts the way the ai is yeah thats how you learn yeah thats right you want to pave the way to future opportunities you cant wait until the opportunity is sitting in front of you for you to reach off word and so you have to anticipate in our job of cos to look around corners and and anticipate where will opportunities be someday and even if im not exactly sure what and when how do i position the company to be nearet to be just standing kind of near under the tree and we can do a diving cache when the apple false he guess i am saying yeah！

but you gotta be close enough to do the diving catch pre wind twenty open ai if you havenbeen Ling this grounwork in the data center， yeah， you wouldnt be powering open ai right now！

but the idea that computing will be mostly done away from the viewing device that the vast majority computing will be done awake from the computer itself that insight was good in fact， cloud compuiting everything about today s computing is about separation of that and by putting it in a data center， we can overcome this latency problem meaning youre not going overcome speed， speed and end is only a hunt twenty mills seconds or something like that its not that long from a data center to a interanywhere are the planet yeah yeah and so we do i see and literally across the planet yeah right， so we could solve that problem approxmilly something like that i would forget the number， but its is seventy millix seconds， hard million seconds but its not that long and so my point is if you could remove the obstacles everywhere else then speed a light should be you know particularly finding and you could build data centers as large like and you could do amazing things and and this low tiny device that we use as a computer or you know your tv as a computer whatever computer they all they can all instantly become amazing and so that insight you know fifteen years ago was a good one so speakingof the speed of light in finiband yeah！

like da David like begging me to go here can i feel like the same time you totally saw that infinaband would be way more useful way sooner than anyone else realized yeah， acquiring melinox， i think you uniquely saw that this was required to train large language models and you were super aggressive and acquiring that company why did you see that when no one else that well？

uh， there are several reasons for that first um if you want to be a data center company build building the processing trip is in the way to do it a data center is distinguished from a Desktop computer versus a cell phone not by the processor in it yeah， a Desktop computer in a data center uses the same cpuse uses same GPUs apparently right， very close and so its not the chip itnot the processing should but this describes it but its the networking of it is the infrastructure of it is the you know how the the computing is distributed how security is provided how networking is done you know so on so forth and so so it those characteristics are associated with melinox not Nvidia and so the day that i concluded that really Nvidia wants to be a you know build computers of the future and computers of the future you can be data centers embodied and data centers then we， then we want to be data center in the company， then that we really need to get into networking and so that was one the second thing is observation that whereas cloud compiting started in the hyperscale， which is about taking commoding components a lot of users and virtualizing many users uh on top of one computer ai is really about distributed computing where one job one training job um is orchistrated across millions of processors， and so its the inverse of hyperscale almost and the way that you design a hyperscale computer with with oft shell commotive eathernet， which is just fine for hadoop is just fine for search queries is just fine for all of those things is but now when youre sharing a model across now from your sharing a model across right and so now that observation says that the type of networking you want to do is not exactly Ethernet and the way that we do not working for super computing is really quite ideal and so the combination of those two ideas um i you know convince me that that melinoxes is absolutely the right the right company because they were theyre the world leading high performance networking company and and we work with them in so many different areas in in a high performance computing already plus i i really like the people um！

uh the the isrole teams were class uh we have some three thousand people there now and it was one of the best particularly decisions i ever made when we were researching particularly part three of our Nvidia series we talked to a lot of people and many people told us the melinox acquisition is one of if not the best of all time yeah！

any technology company i think so too yeah yeah and its so disconnected from the work that we normally do it was surprising to everybody but frame this way you were you were standing nearwhere the action was yeah so you could figure out as soon as that apple sort of becomes available to purchase like oh！

lmsarabout to blow up im gonna need that everyones gonna need that i think i know that before anyone else does yeah you want to position yourself near opportunities you dont have to be that perfect you know is yes you want to position yourself near the tree and even if you dont catch the the apple before it hits the ground so long sure the first one to pick it up， you want a position shes the close to the opportunities now and so thats kind of a lot of my work is positioning the company near opportunities and and um i having the the uh the company having the skills to to umm monetize each one of the steps along the way so that we can be sustainable what you just said remindme of a great uh afrism from a buffet and mongerwhich is its better to be approximately right than exactly wrong yeah there you go oh yeah thats a good ones good whats good ones yeah fine yeah right listeners we are here to tell you about a company that literally couldnbe more perfect for this episode cruise o yes！

crusoe as you know by now is a cloudprovider built specifically for ai workloads and powered by clean energy annvidia is a major partner of crucial their data centers are filled with a one hundreds and h one hundreds as you probably know withtherizing demand for ai theres been a huge surge in the need for highperforming gpuesleading to a noticeable scarcity of Nvidia GPUs in the market， Criso has been ahead of the curve and is among the first cloudproviders to offer nvidias H100 scale they have a very straightforward strategy create the best aicloudsolution for customers using the very best GPU hardware on the market the customers ask for a like Nvidia and invest heavily in an optimized Cloudsoftware Stack yup to illustrate they already have several customers already running large scale generative AI workloads on clusters of Nvidia H100 GPUs。

which are interconnected with 32 gigabitinfinaband and leveraging Crusos network attached block storage solution and because their cloud is run on wasted stranded or clean energy they can provide significantly better performanceper dollar than traditional cloudproviders yep ultimately this results in a huge winwin。

they take what is otherwise a huge amount of energy waste that causes environm and use it to power massive ai workloads and its worth noting that through their operations crusal is actually reducing more emissions than they would generate in fact in 2022， crusocaptured over 4 billion cubic feed of gas， which led to the avoidance of approximately five hundred thousand metric towns of CO2 missions， thats equivalent to taking about a hundred thousand cars off the road amazing if you？

your company or your portfolio companies could use lower cost and more performance infrastructure for your ai workloads go to crusoclouddot com slash acquired thatcru sowecloud dot com slash acquired or click the link in the shownotes i want to move away from Nvidia if youre OK with it and ask you some questions since we have a lot of founders that listen to the show sort of advice for company building hmm the first one is when youre starting a startup in the earliest days， your biggest competitor is uh you dont make anything people want like your companies likely to die just in some just because people dont actually care as much as you do about where the right you know in the later days you actually have to be very thoughtful about competitive， strategy and im curious what would be your advice to companies that you know have product market fit that are starting to grow there in interesting growing markets um where should they look for competition and how should they handle it well。

there are all kinds of waste the thing about competition we prefer to position ourselves in a way that serves a need that usually has an emerged ive heard？

uh， you were others in a video， i think hes a pray billion dollar thats exactly right yeah！

its our way of saying theres no market yet but we believe there will be one and and usually， when your position there， everybody is trying to figure out why are you here？

right？

because when we first got into automotive， because we believe that in the future， the cars can be largely software， and if is gonna be largely software， um a a really incredible computer is necessary and so so when we positioned ourselves there most people i i i so remember one of the one of the ctos told me you know what cars can not tolerate the blue screen of death as i dont think anybody can tolerate that， but it doesnt change the fact that someday every car will be a software define car and i think of fifteen years later were we were were largely right and so often times theres nonconsuction and we like to navigate our company there and by doing that um by the time that you uh that the market emerges it is very its very likely there are that many competors shape that way and so we were early in PC Gaming and today， uh Nvidia is very large in PC Gaming uh we uh imagined what a what a uh design war station would be like and today just by reward station on the planet uses of Nvidia technology uh we re reimagine um how supercomputing how to be done and who should who should benefit from supercomputing that we would democratize it and look today and videos in an accelerated computing is is um quite large and we reimagine how sour would be done and today its call machine learning and how computer would be done we call a ai and so we reimagine these kind of things uh try to try to do that about a decade in events and so we spend about a decade in zero billion or markets and today i spent a lot time when omniverse and omniverses a you know classic example of us your billing door business and theres like forty customers now yeah！

Amazon， bmw yeah！

no！

its cool its cool so lets say you do get this great tenyear lead but then other people figured out you got people nipping at your heels what are some structural things that someone whobuilding？

a business can do to sort of stay ahead and you can just keep her petal to the metal and say were going to outwork um were going to be smarter and like that works to some extent， but those are tactics what strategically can you do to sort of make sure that you can maintain that lead oftentimes if you created the market？

you ended up， having you know what what people describe as modes because if you build your product， right and its enabled， uh an entire ecosystem around you to help serve that end market if essentially creative， the platform is a its a product base platform， sometimes， as a service base platform， somethames a technology base platform but if you are you are early there and you you were mindful about helping the ecosystem， um succeedwithyou you ended up， having this network of networks and all these developers and all these customers were who are built around you yeah and that network is essentially remote and so you know i i dont love thinking about it in the context of a moat， um and the reason for that is because you now focused on building stuff around your castle i tend to like thinking about things in the context of building a network and that network is about enabling other people to enjoy the success of the final market you know that youre not the only company that enjoys it， but youre enjoying it with a whole bunch of other people， including yeah！

im so glad you brought this up because i want to ask you um in my mind at least in sounds like in years to Nvidia is absolutely a platform company of which there are very few meaningful platform companies in the world， i think its also fair to say that when you started for the first few years， you were a technology company and not a platform company every example， i can think of a company that try to start as platform company fails he got a start as a technology first， when did you think about making that transition to being a platform like your first graphic cards were technology they werned there was no CUDA， there was no platform yeah！

what you observe is not wrong however， inside our company we were always a platform company and the reason for that is because from the very first day of our company， we had this architecture called UDA is the UDA of CUDA CUDA is compute unified yeah device architecture thats right and the reason for that is because what weve done， what we what essentially did in the beginning even though reva one twenty eight only had computer graphics the architecture described accelerators of all kinds and we would take that architecture and developers would program to it in fact， nvidias first strategy， business strategy was we were going to be a game console inside the pc and a game console needs developers， which is the reason why Nvidia a long time ago， one of our first employees was a developer relations person and so its the reason why we knew all the game developers and all the three d developers and we knew we years was the original business plan to like sort of build direct x yeah compete with an in Nintendo and Sega。

as like info PCS。

original Nvidia Architecture was called direct NV directive media yeah， and direct x was an API that made a possible for operance system to directly addirect hardware yeah hardware yeah！

but direct forgn direct when you started Nvidia right and thats what and made your strategy wrong three yeah we had direct Nvidia yeah！

ha， and which in nineteen ninety five became you know well direct x came out so this is an important lesson you we were always a developer already to company the initial attempt was we will get the developers to build on direct nv。

and then theyll build for our chips， and then will have a platform and yeah， exactly， what played out is Microsoft or he had all these developer relationships so exactly you learned the lesson the hard way of like yeah， yes！

we just got a thats a sliders of did back in the day theyre like oh， that could be a developer platform will take that thank you you know no！

but they had a lot they did a very differently and and they did a lot of things right we did a lot of things wrong but but what have you were competing against Microsoft in the 97， thats yeah that against Nvidia today yeah has no its a lot different， but i appreciate that but but we were we were nowhere near near competing with them if you look now， when couta came along and there was opencial， there was to work dex um but theres there still another uh extension of you will and that extension is could a yeah and that could extension allows a trip that got paid for running direct dex and open gl to create an install base for cuda yeah and so thats， CUDA yeah if your computing platform everything is got to be compatible where the only accelerator on the planet where every single accelerator is co Archie architectually compatible with the others Nana has ever existed there are literally a couple of hundred million right 250 million， 3 hundred million installed base of active CUDA GP is being used in world today， and there are architatially compatible how would you have a computing platform？

if if you know， mb thirty and nv thirty would five and a thirty nine and nv forty， theyre all different right at thirty years its all completely compatible and so thats the only unnegotiable role in our company everything else is negotiable i mean i guess uh cuda was a rebirth of uda！

but understanding this now uda going all the way back yeah really is all the way back to all the chips even yeah， yeah！

yeah and in fact it uda is goes all the way back to all of our chips today wow for the record i didnt help any of the the founding ceos that that are listening i got a you know， while you were asking that question what what lessons， what i import um i i dont know， i mean there the characteristics of successful companies in successful cos i think are a fairly well described theyre whole bunch of them i just think starting successful companies are insanely hard its just insanely hard and when i see these amazing companies again， build i have nothing but ameration respect because i i just know that its insanely hard， and i think that everybody did many similar things there are some good smart things that people do there some dumpthings that you can do um but but you could do all the right smart things and still fail you could do a whole bunch of done things and i did many of them and still succeed so obviously thats not exactly right you know just i think skills are are the things are you can learn along the way， but an important moments， certain circumstances have to come together and and i do think that that the market has to you know be one of the agents yeah to help you succeed its not enough obviously because a lot of people still fail do you remember any moments in Nvidia history were you like oh！

we made a bunch of wrong decisions， but somehow we got saved because you know it takes the sum of all the lock and all the skill yeah in order to succeed do you remember any moments？

where i should thought that you starting with rewr river one twenty was that spot on uh river one 28 a a as i mission that the number of smart decisions we made， which are smart to this day how we design shifs is exactly the same to this day because gosh you know nobys， ever done it back then and we pulled every trick in the book in a desperation because we had no other choice well， guess what does the way things are to be done and now everybody doesnt that way right everybody doesnt because why should you do things twice if you can do it once why tab out a chip seven times if we get tapeed out one time right， and so the most efficient， the most cost effective， the most competitive， um uh speed is technology right， speed is performance， time to market is performance， all of those things apply so why do things twice if we could do one once yeah and so real one twenty eight made a lot of great decisions and how we spec products um how we how we think about market needs and unlack of and how do we judge markets and all of this may we make some amazing amazingly good decisions yeah we were you know back against the wall we only have one more shot to do it but once you pull out all the stops and you see we are capable of why would you put stops in exactly like good keep stops out all the time thats right everytime thats right is it pretty say theymay be on the next side of the equation thinking back to nineteen 97 that that was the moment where consumers tipped to really really valuing 3D graphical performance in game oh！

yeah！

so for example， lock lets listen time what lock um if if i car Mac hand um， i decided to use acceleration because remember doing was completely software rendered and the Nvidia philosophy was that although general purpose computing as a as a fabulous thing is gonna enable software in it and everything um we felt that there were， there were applications that would be possible or would be costly if it wasnacsolided， should be acsoliated and three graphics was one of them， but it wasnt the only one and was just happens to be the first one and a really great one and i still remember the first time we met John he was quite emphatic about using CPUs and and a software render was really good i mean quite frankly we look at look at doom uh the performance of doom was really hard to achieve， even with exhilarators of the time you know if you didnt filter， if you didnhave to do bold in your filtering um it did a pretty good job the problem a doom there was you needed car Mac to program it yeah you need car Mac to program it exactly it was it was a genius piece a code and um but none the last software renders did a really good job in but in and if he had decided to go to open gl in accelerate accelerate for quake， frankly you know what would be the killer app that putis here right and so carmac and swinny both between a unreal and quake created the first two uh killer applications for for consumer three d yeah and so i i i old them a great deal i wanna come back real quick to you need some you told these stories and you like well。

i dont know what founders can take from that i i actually do think um you know if you look at all the big tech companies， today perhaps with the exception of Google all they did all start and understanding this now by you by addressing developers planning to build a platform and tools for developers um you know all of the apple not as on but with AWS as how AWS started， so i think that actually is a lesson you point like that wont guarantee success by any means right， but thatll get you hanging around a tree if the apple falls yeah as many good ideas as we have。

um you donhave all the worlds good ideas and and the benefit of having developers is you get to see a lot of good ideas yeah yeah well！

as we we start to drift toward the end here， we spend a lot of time on the past and i want to think about the future a little bit ill sure you spend a lot of time on this being on the cutting edge of ai you know were moving into an error where the productivity that software can accomplish when a person is using software can massively amplify the impact in the value that theyre creating， which has to be amazing for humanity in the long run in the short term its going to be inevitably bumpy as we sort of figure out what that means what do you think some of the solutions are as ai gets more and more powerful and better at accelerating productivity uh for all the displacsed jobs that are gonna come from it all first of all we have keep ai safe and there is a couple of different areas of ai safety um thats really important！

obviously i and robotics and self driving car theres a whole field of ai save DM we dedicator ourselves the functional safety and active safety in all kinds of different different areas of safety um when to apply human and loop when is it OK for human not to be in the loop a a a you know how do you get to a point where where um uh increasingly human doesnhave to be in the loop but human largely in the loop yeah in the case of information save d obviously bias false information and appreciating the the rights of artison and creators um that that whole area uh deserves a lot of attention and and yousing some of the work that weve done instead of scraping the internet um， we we partered with getty and share stock to create commercially fair way of applying artificial intelligence， shared to the eye yeah in the area of our large language models in the in the future of increasingly greater agency， ai clearly the answer is bros long as its sensible and i think its going to be sensible for a long time is a human in the loop of the ability for an ai to selflearn and improve and change out the while i i in the digital form， i should be avoided and um， i we should collect data， we should carry the data， we should train the model， we should you know test the model， validate the model before we release it on the wild again， so humanisin the loop yeah， there are a lot of different industries that have already demonstrated how to build systems that are safe and good for humanity and obviously the way uh auto pilot works for for a plane and to pilot system and then air traffic control and um you know， we dont see in diversity and and all of the basic philosophy is of designing safe systems um apply uh as well and self driving cars and so also for things and so i i think theres a lot of models of of creating safe ai and and i think we need to apply them with respect to automation my feeling is that and well see， but it is more likely that ai is gonna create more jobs and in a near term the question is whats the definition， a near term and the reason for that is is um uh the first thing that that happens with productivity is prosperity and prosperity when the companies get get more successful， they are more people because they want expand into more areas and so the question is if you think about a company and say OK if we improve the productivity， the need they need they need fewer people well， thats because the company has no more ideas but thats not true about your company um if you become more productive and the company become more profitable usually， they are more people to expand into new areas and so long as we believed that there more areas to expand into that the the the more ideas and drugs this drug discovery there more ideas in transportation there more ideas in retail there more ideas in entertainment that theres more ideas in technology so long as we believe that there are more ideas the prosperity of the industry， which comes from improve productivity results in hiring more people more ideas now you go back in history， we can fairly say that todays industry is larger than the industry what the worlds industry is a thousands years ago and the reason for that is because obviously humans have a lot of ideas and i think that theres plenty of ideas yet for prosperity and plenty of ideas that can be beget from productivity improvements but that my senses i dislikely degenerate jobs now obviously that net generation of jobs doesnt guarantee that any one human doesnget fired OK， i mean that is obviously true and and its more likely that someone um will lose a job to someone else some other human theyuses an ai you know and not not likely to an ai but to some other human， theyuses an ai and so i think the the first thing that everybody should do is learn how to use ai so that they can augnment their own productivity and every company should argument their own productivity to be more productive so that they get up more prosperity higher more people and so i think jobs will change my guesses that will actually have higher employment will create more jobs i think industrys will be more more productive um and many of the industry that are currently suffering from the lack of lack of labor uh workforce is likely to uh use ai to get themselves off the free and and get back to growth in prosperity so i see it a little bit differently but i do think that jobs will be affected um and i id encourage everybody just to learn ai this is a property this a version of um something we talk about a lot on acquiredwe call it the merits coralary to moreslaw after Mike merits from a ha yeah from a sequid um！

yeah of course， yeah yeah the great story behind it is that uh when Mike was taking over for down Valentine with with dog， even sitting in looking at sequence returns， name is looking at fun three or four i think was four maybe that had this is going and he is like how are we ever gonna top that you know i cant i cant you know dons gonna have us beat were never gonna beat that they thought about any realize that well， as compute gets cheaper and it can access more areas of the economy because it gets cheaper and can it get adopted more widely well， then the markets that we can address should get bigger yeah and ai your argument is basically exactly i will do the same thing excycle i just gave you exactly the same example that in fact。

productivity doesnt result in us doing less productivity usually， results in us doing more hmm everything we do will be easier， but weend up doing more yeah because we have infinite nvision you know the the world has infinite so so if a company is more profitable。

they tend to hard more people to do more job yeah thats true technology is a lever and the the place where the idea kind of falls down is that like that we would be satisfied yeah like yeah humans have never ending ambition no， he humans will always expand and consumer energy and uh attempt to pursue more ideas that is always been true of every version of our species yeah over time now is a great time to share something new from our friends。

a blink est and go one that is very appropriate to this episode yes！

so personal storytime i a few weeks ago was scouring the web to find jensonsfavoritbusiness books， which was proving to be difficult i really wanted blink is to make blinkes of each of those books so you could all access them and i think i found one or two in random articles， but that just wasnt enough so finally before i gave up as a last resort， i asked an ai chat bot specifically barred to provide me a list and site the sources of Jensons favor business books and miraculously it worked bardfoundbooks that Jensen had called out in public forums over the past several decades so if you click the link in the shownotes or go to blink， as dot com slash Jenson， you can get the blinks of all five of those booksplus， a few more that Johnson specifically told us about leader in the episode yes！

and we also have an offer from blinkistinggo one that goes beyond personal learning blinkys has handpicked a collection of books related to the themes of this episode so techinnovation leadership the dynamics of acquisitions these booksofferthemental models to adapt to a rapidly changing technology environment。

and just like all other episodes blinkies is giving acquiredlisteners， an exclusive fifty percent discount on all premium content this gives you keyinsights from thousands of booksetyourfingertips， all condensed into easytodigest summarys and if yourefounder， a team lead or an lnd manager， blinkest also includes curated， readingless and progress tracking features， all overseen by a dedicated customersuccess manager to help your team flourish as you grow yes！

so to claim the whole free collection unlock the fifty percent discount an explorer blink is enterprise solution simplyvisitblinkist dot com， slass Jenson and used the promocode Jenson biggest and their parent company go one aretruly awesome resources for your company and your teams as they develop from small startup to enterprise are thanks to them and seriously the soffer is pretty awesome go take em up on it we have a few waiting around questions we want to ask you can， then we have a can！

then we have a very far think that fast what of far an easy one based on all these conference rooms we see in name around here favorite scifybook dive never read of scifi book before no come on yeah whats like the obsession with star track and oh！

this is your watching tv show yeah OK favorite sifi tv sales i will start tracks my favorite yeah yeah start tracks my favor its not feature out there on the way and good to good competter name visuisan excellent yeah yeah what car is your daily driver these days and related question to these head super oh， i know is one of my favorite cars um and also favorite memories you guys might not know this but but uh， i lore and i got engaged um i Christmas， one year and we Joe back in my my brand new super and we totalled it we were this close to the end think i didnt but but nonetheless it was my fault wasnt wasnthesupersport but but it its a remark i i love the one time when it wasnt the super sfall， yeah i love that car im driven these days for first curious anothers but um， uh， uh， im driven in the mercities eqs is great ah i am yeah great card thanks！

yeah using Nvidia technology！

yeah has yeah werin were in the in the uh， the the yeah were the central computer yeah sweet i know we already talked a little bit about business books but one or two favorites that youve taken something from click Christians and i think has that the series is the best i mean theres this no no two ways about it and and the reason for that is b is because its so intuitive and so sensible its it its approachable but i read a whole bunch of them and i read just by all of them i really enjoyed any and growth books theyre all really good custom favorite characteristic of dawn Valentine grumpy but endearing and what he said to me the last time as he decided to invest in our company， says if you lose my money， ill kill you god sanded and then over the course of of the decades uh， uh， the years of followed uh when something is nice， written about us in mercury news， um it seems like he wrote it in a crayon he you know hell， say hell， say good job done and just right right over the newspaper and just good job darnies males a tenant i i hope we ikept kept them but anyways you could tell hes a hes a real sweetart and and um！

um but but he cares about the companies this special character yeah is increble what is something that you believe today that forty year old Johnson would have push back on and said no i disagree theres plenty of time 嗯？

yeah theres plenty of time if you prioritize yourself， uh properly and you make sure that you， you？

you dont let outlook be the controller of your time theres plenty of time plenty of time in the day plenty of time do anything to achieve state like to understand just dont do everything priortize your life make sacrifices dont let outlook control what you do everyday notice i was late to our meeting and the reason for that but time i looked up i oh my gosh you know， Ben and David are waiting you know yeah， its already we have time yeah！

exactly thats so didns not this from being a great Jack no！

but you have to prioritize your time really carefully and dont let outlook the determining that love that what are you afraid of if anything im afraid of the same things today that i was i was a in in the very beginning of this company， which is letting the employees down you know you have a lot of people who join your company because they believe in your hopes and dreams and and they adopted it as their hopes and dreams and and you you want to be right for them， you want to be successfor them for them you want them to be able to build a great life as as well as help you build a great company and be able to build a great career you want them to have to enjoy all of that and these days i want them to be when i enjoy the the things ive had the benefit of enjoying and all the great success ive enjoyed i want them to be will enjoy all the that and so so i think i think the the greatest period is that that you let them down what point did you realize that you werengonna have another job that like this was it i just i dont change shops you know if it wasnbecauseof Chris and Curtis convincing me to do doing video i would still be a else i logic today im certain of it wow， yeah but really yeah yeah im sort of it i would keep doing what im doing and at the time that i was there i was completed dedicated and focused on on helping allicylogic be the best company could be and i was lsylogics best ambassador ive got great friends that to this day uh that ive known from from ls i logic i its a company i i loved uh then i love dearly today i know exactly why what um uh the revolutionary impacted had on chip design and system design in computer design in my estimation one of the most important companies that that ever came to Silicon valley and change to everything about how computers were made i it put me in the in the epic of some of the most important events in computer industry had led me to meeting Chris and Curtis and Devecto Shin and John Robin Stein and you know some of the most important people in the world and at frank that i always with the other day and just i mean the list goes on and and so i i alisci logic was really important to me and and i would still be there i i would you know who knows what else i logic would become if i were still there right and and so thats kind of how my my mind works um paring the ai of the world yeah， who exactly i mean i might be doing the same thing im doing that the sense from yeah！

remember back to part one of our series on Nvidia but until until im fired this is this is my last job i love yeah， yeah， i got this instead um lsi logic might have also changed your um perspective in philosophy about computing to the sense i we got from the research was that when right out of school and when you first went a md first right yeah！

you believed that like kind of a version of that was it the Jerry standards real men have fabs like you you need to do the whole stack like you gotta do everything and that lsa logic changed you what lsi logic did was was a realized that you can express um transistors in logical gates and chip functionality in high level languages that by raising the level up abstraction and what is now call high level design it was coined by a Harbi Jones whos on a nnvideus board and i met met him out way back in the early days of synopsis but but during that time there was this belive that you can express chip design in high level languages and by doing so you could take advantage of optimizing compilers in optimization logic and tools um and and be a lot more productive that logic was so sensible to me and i was twenty one years old time and i i want to pursue that vision now frankly that that idea happened in an um uh machine learning had happened in also， for programming it i want to see it happened in digital biology so that we can we can think about uh biology and a most higher level language uh probably a large language model um would be the the way to make it， make it representable that transition was so revolutionary i thought that was the best thing ever happened to the industry and i was per i was really happy to be part of it now was i ground url and so so i i saw one industry um change revolutionize another industry and if not for lsi logic doing the work that it did i set up this shortly after then， why would the computer industry be weird is today yeah its a really really terrific i was i was a at the right place at the right time to see all that i was super cool yeah and it sounded like the co of alsa logic put a good word in for you yeah without i dont know how to write a business plan。

which it turns out is not actually important no！

no， no it turns out that making a financial forecast that nobody knows is going to be right around turns out not to be that important， but the important things that a business plan probably could have teased out i i think that the the art of writing a business plan out be much much shorter and a forces you the contents you know what what is the true problem years trying to solve what is the unmet need that you？

you believe will emerge and what is it that you gonna do that is sufficiently hard that when everybody else finds out is a good idea that theyre not gonna swarm it and you know make you upsleep and so that has to be sufficiently hard to do um there， there are whole bunch of those skills that are involved in just you know product and positioning and pricing and go to market and you know all that kind of stuff but those are skills and you can learn those things easily the stuff that is really really hard is the essence what i described i did that OK but i i know idea how to write the business plan and um i and i was fortunate that wolf corrigan was so pleased with me and the work that i did when i was at elsea logic he called up down Valentine and told done you know investing this kid and um i hes gonna come your way and and uh， uh， so so i was you know i was always set up for success from that moment and got it got a song ground yeah as long as he didnt lose the money i think secreted OK， OK？

yeah we！

we， i i think we probably are one of the best investments theyve ever made have they held through today the vc partner i still on the board mark stew， yeah！

mark hell， yeah！

yeah！

yeah！

all these years the two founding vcas are still on the board sutter hell and sequence， yeah！

tenge cocks and more Stevens i dont think that ever happens， yeah may where singular in that in that circuits dance， i believe theyvad value this whole time uh been inspiring this whole time i uh uh gave great wisdom and uh uh great support uh but they they also uh were they will tell you yeah not yeah but they theyre entertained you know by the company and spared by the company and enrich by the company and so they stayed with it then i and im im really grateful well？

in app in our final question for you， its twenty twenty three 嗯， thirty years university of the founding of Nvidia if you were magically thirty years old again？

today twenty three and you were going to Dennis with your two best friends or the two spardest people you know and youtalking about starting a company what are you talking about starting i wouldndo it i know and the reason for that is really quite simple ignoring that company that we would start first of all of not exactly sure the reason why when do it and it goes back to wide so hard is building a company and building a video turned out to have been a million times harder than i expected it to be any of us expected it to be and at that time if we realize the paint and suffering and just have vulnerable you， youre gonna feel um and the challenges are you gonna endure i the embarrassment in the shame and you know the list of all the things that that go wrong i dont think anybody would start it company nobody in their right line would do it and i think that thats kind of the the superpower of a entrepreneur they dont know how hard this and they only ask themselves how hard can i be and to this day i i trick my brain into thinking how hard can i be because you have to still yeah you wake up the warning yeah how hard can i be everything that were doing how hard can i be on reversed how hard can i be you know in terms of the super that youre um planning a retier anytime see you know no youre still thinking that you the couldnchoose to say like wow， this is too hard the trick is to work youre still the trick is to work im still enjoying myself immensely and im adding a little bit of value but but the the thats thats really the trick of an entrepreneur you have to get yourself to believe that its not that hard because its way hard than you think and so if i go taking all of my knowledge now and i go back and i said im gonna endoor that whole journey again i think its too much it is just too much do you any suggestions on any kind of support system or a way to get through the emotional trama that comes with building something like this have family and friends and all were colleages we have here uh im surrounded by people who been here for thirty years Chris been here for thirty years and jefishers been here thirty years been here thirty years and joiner and Brian and been here you know twenty five some years and probably longer than that and you know Joe Grechos been here thirty years im surrounded by these people that never one time gave up and they never one time gave up on me and thats the entire ball wax you know and and to be able to go home and and i i have your family be fully committed to everything that youre trying to do when um uh sick or then theyre theyre proud of you and pride the company and you cant need that you need the unwavering support of people around you， you know Jim gathres and the more you know the the tenge coxes， the marks deals and you know Harvey Jones and all the the early people of our company to build millers they are are not one time gave up on the company in an awson and you can not you need that you are not kind of need that you need that and im pretty sure that almost every successful company and entrepreneurs that that have gone through some difficult challenges they they had that support system around them i can only imagine how meaningful that i mean i know how meaningful that is in any company but for you give it i i feel like the Nvidia journey is um particularly amplify on these dimensions right and like not normal you know you went through two two if not three youd need percent plus drawdowns in the public markets yeah they have investors who stuck with you yeah from day one through that must be just like so much support yeah yeah it is incredible and you hate that any other stuff happened and and most of you you know most of it is is out of your control but you know eighty percent for it its an executive thing no， no matter how you look at it and i forget exactly but i mean we we traded down at about a couple of two three billion dollars in market value for a while because of the decision we made in going into cuda all that work and your believes system has to be really really strong you know you have to really really believe it really really want it otherwise its just too much to endure i mean because you know everybody questioning you and employees are questionyou， but employees of questions right um people outside are questioning you， and its all embarrassing and its like you know when your stock price gets hit isnbearsing no matter how you think about it and its hard to explain you know and so there is no good good answers any that stuff you know the cos are human and companies are built of humans and these challenges are hard to endure and then had an appropriate common on our uh worst reason episode on you all were uh uh we were talking about you know the current situations i think you said for any other company。

this would be a been a precarious spot to be in but for Nvidia that this is kind of all that yeah you know you guys are familiar with these large swings an amplitude yeah！

the thing that that the keep in mind is at all times i what is the market opportunity that that youre engaging and that help that informs your size you know was i was told a long time ago that nvideo can never be larger than a billion dollars obviously as an underestimation， under under imagination of the size of the opportunity yeah， it is the case that no trip company can ever be so big and so but if youre not a trip company then then why that wise that apply to you yeah and this is the extraordinarily thing about technology right now is technology is a tool and its only so large whats whatunique about our current circumstances today is that were in the manufacturing of intelligence， were in the manufacturing of work world does ai and the world of tasks doing work productive， generative ai work， generative intelligent work that market size is enormous as measurement trillions one way to think about that is if you build a trip for a car how many cars are there in how many chips would they consume thats one one way to think about that however， if you if you build a a system that i whenever needed a assistant in the driving of the car um and you know whats the value of a autonomous show fer um， every now and then and so now the the market uh obviously the problem becomes much larger， the opportunity becomes larger um you know what would be like if we， if we were does the magically conjure up um a show for for everybody uh who has a car and you know how big is that market and obviously obviously that thats a much much larger market and so the technology industry is at the uh you know where what we discovered what Nvidia is discovered what some of the discovered is a byseparating ourselves from being a trip company um but the building on top of a trip in your now in the ad company， the the market opportunity has has grown by probably a thousand times you know dont be surprised if technology companies become much larger in the future， because because what you produce i， it is something very different and and that thats the kind of the the uh the the way to think about you know how large can your opportunity how large can you be has everything to do with the size of the opportunity thank you！

so much thank you oh， David that was awesome so fun olistners we want to tell you that you should totally sign up for our email list of course， it is notifications when we drop a new email， but weve added something new we are including little tidbits that we learn after releasing the episode including listener， corrections， and we also have been sort of teasing what the next episode will be so if you want to play the little guessing game， along with the rest of the acquiredcommunity signeupacquireddatafmslash email 2 huge thank you to blinkist stat， seg and cruise o all the links in the shownotes are available to learn more and get the exclusive offers for the acquiredcommunity from each of them， you should check out ack two， which is available at any podcast player as these main acquiredepisodes get longer and come out uh you know， once a month instead of uh， once every couple weeks， its a little bit more ferrarity these days weve been upleveling our production process and that takes time yes ack two has become the place to get more from David and i and weve just got some awesome episodes coming up that we are excited about if you want to come deeper into the acquiredkitchen becoming lp acquireddatafm slash lp once every couple months or so webe doing a call with all of you on zoom just for lps to get the inside scope of whatgoing on an acquiredland and get to know David and i a little bit better and once a season youll get to help us pick a future episode so thats acquireddatafmslash lp anyone should join the slack acquiredatafm slash slack god weve got a lot of things now David！

i know the hamburger bar on our website is expand！

expanding and now thats how you know were becoming enterprise you to have a mega menu a menu of menus if you will what is the acquiredsolution that we can sell thats true we gotta find that all right with that listeners acquiredatafm slash slack to join the slack and discuss this episode acquiredatafm slash store to get some of that sweet merch that everyone is talking about and with that listeners we will see you next time well！

see you next time who got the truth is you isnyou with you who got the truth now 哼。