Latest / From Spark to New Venture / Jensen Huang--From Denny’s to the AI Revolution
Transcript
- 0:00SPEAKER_00: What's up everyone and welcome to the podcast.
- 0:03SPEAKER_00: My name's Alex.
- 0:04SPEAKER_00: And I'm Tyler.
- 0:04SPEAKER_00: And today we're going to be talking about Jensen Huang, who's the co-founder and CEO of NVIDIA.
- 0:09SPEAKER_00: And obviously most people have probably heard of NVIDIA at this point, especially if you're into gaming, computers, even just like all the AI stuff that's been happening recently.
- 0:17SPEAKER_00: Yeah, I feel like even if you don't really know what NVIDIA does, you've definitely heard the name.
- 0:22SPEAKER_00: Right, and that and that's actually what made us interested in talking about Jensen Huang.
- 0:26SPEAKER_00: Because when you look at NVIDIA today, you see this massive technology company that's involved in gaming, AI, data centers, all that type of stuff.
- 0:36SPEAKER_00: But NVIDIA didn't originally start as an AI company.
- 0:40SPEAKER_00: Not even close, really.
- 0:41SPEAKER_00: Yeah, it it really started around computer graphics, and Jensen's story before NVIDIA is actually really interesting too.
- 0:49SPEAKER_00: So he was born in Taiwan, then he came to the US as a kid, he got a job at Denny's, funny enough, and then he went to college at where where was it?
- 1:00SPEAKER_00: Uh Oregon State University.
- 1:02SPEAKER_00: Oregon State University.
- 1:03SPEAKER_00: And then he became an extra uh electrical engineer, and then he started NVIDIA when he was around 30 years old.
- 1:08SPEAKER_00: Which is a pretty big jump from Denny's to running Nvidia.
- 1:12SPEAKER_00: Yeah, dude.
- 1:13SPEAKER_00: No, definitely.
- 1:14SPEAKER_00: So we're basically gonna go through how that happened and more importantly, what made Huang successful as an entrepreneur.
- 1:23SPEAKER_00: So we gotta start back from the beginning.
- 1:27SPEAKER_00: Born in Taiwan, 1963, when he was still young, his family moved to Thailand because of his father's job and whatnot.
- 1:35SPEAKER_00: But eventually his parents decided that Jensen and his older brother should go to the US for school.
- 1:41SPEAKER_00: And he was only nine years old when they sent him over to the U.S.
- 1:44SPEAKER_00: Nine years old?
- 1:46SPEAKER_00: Yeah.
- 1:47SPEAKER_00: Dude, that's wild.
- 1:48SPEAKER_00: I couldn't imagine moving at only nine.
- 1:51SPEAKER_00: Yeah, no, it's it's crazy, especially when you think about how different everything would have been.
- 1:55SPEAKER_00: Like a new country, new culture, and obviously English wasn't the language he he grew up speaking.
- 2:01SPEAKER_00: He and his brother uh ended up at a school in Oneida, Kentucky.
- 2:06SPEAKER_00: And uh apparently that's actually where some confusion from his parents about what kind of school they were sending him to.
- 2:14SPEAKER_00: Uh, that's not exactly where you'd expect the future CEO of Nvidia American story to start, really.
- 2:20SPEAKER_00: No, dude, not at all.
- 2:21SPEAKER_00: But eventually, Jensen's parents immigrated to America and the family reunited in Oregon.
- 2:28SPEAKER_00: And Jensen was apparently a really strong student because he graduated from um Aloha High School when he was only 16 years old.
- 2:35SPEAKER_00: He graduated at 16?
- 2:37SPEAKER_00: Yeah, dude.
- 2:37SPEAKER_00: This this guy's just a stud.
- 2:39SPEAKER_00: And here's another completely random fact about him.
- 2:43SPEAKER_00: He was also a really good ping pong player when he was young.
- 2:47SPEAKER_00: So we've got Nvidia CEO, an electrical engineer, and a ping pong player.
- 2:52SPEAKER_00: Pretty pretty much.
- 2:54SPEAKER_00: So when does the whole Denny's thing happen?
- 2:57SPEAKER_00: So around this time, uh, before becoming an engineer, and obviously way before he was CEO, he was uh washing dishes at Denny's and also worked as a busboy and waiter.
- 3:10SPEAKER_00: And this is something he actually still talks about to this day.
- 3:13SPEAKER_00: Wow, really?
- 3:13SPEAKER_00: Yeah, because I mean you think he'll look back and say Stanford or engineering taught him everything about running a company, but he's always talked about how working in restaurants taught him a lot, a lot about dealing with people.
- 3:26SPEAKER_00: And if you've never if you ever worked in a restaurant, you kind of understand what he's talking about.
- 3:29SPEAKER_00: Yeah, you're dealing with customers, coworkers, people getting pissed off, hungry.
- 3:33SPEAKER_00: Exactly, yeah, exactly.
- 3:35SPEAKER_00: Everything happens at once, really.
- 3:37SPEAKER_00: Yeah, you got you gotta communicate with people and figure things out quickly.
- 3:41SPEAKER_00: And Jensen has connected that experience to his uh his leadership.
- 3:44SPEAKER_00: I think that's actually important to his entrepreneurship story as a whole because being a CEO isn't just being the smartest engineer in the room.
- 3:52SPEAKER_00: You gotta communicate, convince people to work for you, you gotta convince investors to give you money, all that type of stuff.
- 3:59SPEAKER_00: And eventually you have to convince customers that what you're building is worth buying.
- 4:04SPEAKER_00: So even though Denny's has nothing to do with computer chips, some of the skills still carried over.
- 4:09SPEAKER_00: Precisely.
- 4:11SPEAKER_00: Eventually, Jensen, after working at Denny's, he went to Oregon State University and studied electrical engineering.
- 4:19SPEAKER_00: He then graduated in 1984 and started working in the semiconductor industry.
- 4:25SPEAKER_00: He worked at AMD, which is funny enough now, because AMD and NVIDIA are the two biggest competitors in in graphics, and they're they're seen as rivals.
- 4:36SPEAKER_00: Wow, so he started at AMD and started NVIDIA.
- 4:41SPEAKER_00: Yeah, he also worked at LSI Logic, another semiconductor company, and while he was working there, he continues uh he continued his education.
- 4:49SPEAKER_00: In 1992, he earned his master's degree in electrical engineering from Stanford.
- 4:54SPEAKER_00: So by this point, he's got the engineering education, but he also has actual industry experience.
- 5:01SPEAKER_00: Probably helps a lot more when you're about to start a semiconductor company.
- 5:05SPEAKER_00: Oh, definitely, dude, for sure.
- 5:07SPEAKER_00: And then that that brings us to 1993.
- 5:11SPEAKER_00: Jensen Huang, and I think I'm saying this right, but Chris Malachowski, Chowski, and Curtis Preem decide to start NVIDIA while Jensen was around 30 years old.
- 5:22SPEAKER_00: And Nvidia isn't making AI chips at this point.
- 5:26SPEAKER_00: Nope.
- 5:26SPEAKER_00: And that's probably the most important thing to remember about this entire story.
- 5:29SPEAKER_00: AI wasn't the original idea.
- 5:32SPEAKER_00: They believed that computer graphics were going to become much more important.
- 5:36SPEAKER_00: I mean, just think about computers in 1993 compared to computers now.
- 5:40SPEAKER_00: Yeah, gaming wasn't anything like it is now.
- 5:42SPEAKER_00: Not even close.
- 5:43SPEAKER_00: Completely different.
- 5:44SPEAKER_00: No, no Valorant, no Fortnite, no Call of Duty, no Steam, no Twitch.
- 5:50SPEAKER_00: I mean, esports wasn't this, you know, massive industry that people talk about.
- 5:56SPEAKER_00: I mean, even 3D graphics are still pretty new.
- 5:58SPEAKER_00: But Jensen and his co-founders basically believed that people were eventually going to demand way better graphics from computers.
- 6:05SPEAKER_00: Turned out to be a pretty good prediction.
- 6:06SPEAKER_00: Yeah, absolutely.
- 6:08SPEAKER_00: But here's the thing NVIDIA didn't just start and immediately dominate.
- 6:12SPEAKER_00: The company had serious problems early on.
- 6:15SPEAKER_00: The chip industry is brutal because developing a chip costs a ton of money and takes super long time.
- 6:21SPEAKER_00: So if you spend all that money developing the wrong technology, you're screwed.
- 6:25SPEAKER_00: Exactly.
- 6:26SPEAKER_00: And NVIDIA actually made some early technological bets that didn't work the way they wanted.
- 6:31SPEAKER_00: The company had to change direction.
- 6:33SPEAKER_00: And this is one of the first things about Jensen that I really think shows his entrepreneurial mindset.
- 6:38SPEAKER_00: He wasn't right about everything.
- 6:40SPEAKER_00: Which I feel like gets ignored when people talk about successful entrepreneurship.
- 6:44SPEAKER_00: A hundred percent.
- 6:45SPEAKER_00: We always hear we always hear like the successful decision decisions after somebody becomes a billionaire.
- 6:51SPEAKER_00: But there were decisions that almost killed the company too.
- 6:54SPEAKER_00: The important part was that NVIDIA adapted.
- 6:57SPEAKER_00: So when does NVIDIA really start taking off?
- 7:00SPEAKER_00: When the first big high marks in with NVIDIA comes in 1999, NVIDIA releases the GeForce 256, which is was marketed as NVIDIA's were the world's first GPU, basically, or gr graphics processing unit.
- 7:20SPEAKER_00: Which obviously is a term everybody knows now if they build PCs.
- 7:24SPEAKER_00: 100%.
- 7:25SPEAKER_00: If you're building a gaming PC today, your GPU is one of the first things you're thinking about.
- 7:29SPEAKER_00: You're probably signing between different GForce cards or AMD cards, figuring out how much VRAM you need, what resolution you're playing at, what frame rate you're trying to get.
- 7:39SPEAKER_00: But back then this market was still developing.
- 7:42SPEAKER_00: The idea behind a GPU is basically that instead of relying only on your main processor or CPU, you have specialized hardware that's really good at handling graphics calculations.
- 7:52SPEAKER_00: And games started becoming way more complicated.
- 7:55SPEAKER_00: You had more detailed environments, textures, lighting, 3D models, all of that.
- 8:00SPEAKER_00: So better games basically created more demand for NVIDIA's products.
- 8:04SPEAKER_00: Exactly.
- 8:05SPEAKER_00: And NVIDIA kept pushing graphics performance forward.
- 8:08SPEAKER_00: Gaming became a huge part of the company's identity, and that's why for a long time, when people heard NVIDIA, they thought gaming.
- 8:16SPEAKER_00: I mean, that's still what I think of first.
- 8:18SPEAKER_00: Oh, same, absolutely.
- 8:20SPEAKER_00: But then NVIDIA made a decision that looking back might be even more important than what it did for gaming.
- 8:27SPEAKER_00: In 2006, NVIDIA introduced CUDA.
- 8:32SPEAKER_00: Okay, so explain CUDA like I'm five.
- 8:34SPEAKER_00: Right.
- 8:35SPEAKER_00: Basically, CUDA allowed programmers to use NVIDIA GPUs to do things other than just create graphics.
- 8:43SPEAKER_00: Okay.
- 8:44SPEAKER_00: GPUs are really good at doing a lot of calculations at the same time, right?
- 8:48SPEAKER_00: That's called parallel computing.
- 8:51SPEAKER_00: So imagine you've got a massive job that can be broken into thousands of smaller jobs.
- 8:57SPEAKER_00: Instead of doing every single one in order, you can have a bunch of processing units working on pieces of the problem at the same time.
- 9:06SPEAKER_00: That's really useful for graphics because your computer is constantly calculating tons of visual information.
- 9:12SPEAKER_00: But researchers realized I only use that ability for graphics.
- 9:17SPEAKER_00: Okay, okay.
- 9:18SPEAKER_00: You can use that computing power for scientific research, simulations, data processing, and eventually artificial intelligence.
- 9:28SPEAKER_00: So CUDA basically helped turn the video graphics cards into more general computing tools.
- 9:34SPEAKER_00: Exactly.
- 9:35SPEAKER_00: And that is just crazy timing because CUDA came out in 2006.
- 9:40SPEAKER_00: Think about that.
- 9:40SPEAKER_00: Yeah, that's way before any of the current AI stuff going on.
- 9:44SPEAKER_00: Way before it.
- 9:45SPEAKER_00: There was no Chat GPT, no Claude.
- 9:48SPEAKER_00: Generative AI wasn't something normal people were talking about every day.
- 9:53SPEAKER_00: Nvidia was investing in that idea that GPUs could have a much bigger purpose than gaming, and it took years for the true value of that decision to become obvious.
- 10:02SPEAKER_00: So when does AI actually enter the picture with NVIDIA?
- 10:07SPEAKER_00: The next really important moment that happens in 2012.
- 10:11SPEAKER_00: There was a neural network called AlexNet.
- 10:16SPEAKER_00: And AlexNet performed extremely well in this major image recognition competition called ImageNet.
- 10:23SPEAKER_00: The researchers use NVIDIA GPUs to train it, and this helped demonstrate just how useful GPUs could be for deep learning.
- 10:30SPEAKER_00: So why are GPUs so good for AI?
- 10:33SPEAKER_00: It's a similar reason, they're good for graphics, right?
- 10:36SPEAKER_00: Training an AI model requires an insane amount of math.
- 10:39SPEAKER_00: A lot of calculations can be can be oh my god, can be performed in a parallel, excuse me.
- 10:45SPEAKER_00: But GPUs are designed to handle that type of workload.
- 10:49SPEAKER_00: So this technology that NVIDIA spent years developing for graphics suddenly became extremely useful for training neural networks.
- 10:58SPEAKER_00: That's actually wild when you start to think about it, because you're making chips so somebody can make better graphics in a video game, and then years later the same basic type of technology is helping train AI.
- 11:10SPEAKER_00: Yeah, yeah, and I I I think that's one of the coolest parts of Jensen Huang's story.
- 11:15SPEAKER_00: He didn't predict every single thing that would happen.
- 11:18SPEAKER_00: It's not like he sat down in 1993 and said in 30 years everyone's gonna be using AI.
- 11:23SPEAKER_00: True.
- 11:24SPEAKER_00: But he recognized that accelerated computing had applications beyond graphics and NVIDIA kept investing in it.
- 11:32SPEAKER_00: But Nvidia obviously didn't stop caring about gaming.
- 11:35SPEAKER_00: No, definitely not.
- 11:36SPEAKER_00: And another big gaming moment comes in 2018, when NVIDIA introduces its RTX platform.
- 11:42SPEAKER_00: One of the big selling points was real-time ray tracing, which ray tracing is basically a way of simulating how light behaves in the real world.
- 11:51SPEAKER_00: It can create more realistic lighting, reflections, shadows, stuff like that.
- 11:55SPEAKER_00: While absolutely destroying your framework though.
- 11:57SPEAKER_00: Yeah, dude, absolutely, tell me about it.
- 11:59SPEAKER_00: It's it's crazy.
- 12:00SPEAKER_00: You need you need top-of-the-line stuff.
- 12:02SPEAKER_00: Sounds about right.
- 12:03SPEAKER_00: Yeah.
- 12:04SPEAKER_00: But that's another example of NVIDIA pushing toward more demanding, you know, graphics.
- 12:12SPEAKER_00: And they also develop technologies like DLSS that uses AI to help improve gaming performance and image quality.
- 12:20SPEAKER_00: So these different sides of NVIDIA start overlapping.
- 12:24SPEAKER_00: So graphics technology was benefiting from AI heavily, and AI was benefiting from hardware originally developed around graphics.
- 12:32SPEAKER_00: And then we get to the current AI boom.
- 12:35SPEAKER_00: Generative AI becomes massive.
- 12:38SPEAKER_00: Suddenly, you've got companies trying to build increasingly powerful AI systems, and to do that, they need an enormous amount of computing power.
- 12:46SPEAKER_00: Which is exactly what Nvidia sells.
- 12:48SPEAKER_00: Boom.
- 12:49SPEAKER_00: And this is where Jensen's long-term decisions really pay off.
- 12:54SPEAKER_00: Because NVIDIA didn't just wake up one morning, one morning, and decide to start making AI hardware.
- 13:00SPEAKER_00: It had spent years developing GPUs.
- 13:02SPEAKER_00: It had CUDA, it had developers who already knew how to work with the technology, and it had experience with accelerated computing.
- 13:10SPEAKER_00: So when demand for AI computing exploded, NVIDIA already had a massive head start.
- 13:16SPEAKER_00: So basically the biggest decision might have happened years before everybody even cared about AI.
- 13:22SPEAKER_00: I mean, yeah, dude, that that's how I look at it.
- 13:25SPEAKER_00: And that's probably the main entrepreneurship lesson I'd I gathered from researching Jensen.
- 13:31SPEAKER_00: Sometimes you're investing in something without knowing exactly what the final market is going to look like.
- 13:36SPEAKER_00: So what do you think actually makes Jensen different from other entrepreneurs?
- 13:41SPEAKER_00: I think there are probably three big things I would say.
- 13:45SPEAKER_00: First, he's willing to take risks.
- 13:47SPEAKER_00: Starting any company is risky.
- 13:49SPEAKER_00: Starting a semiconductor company is another level.
- 13:52SPEAKER_00: You need engineers, huge amounts of investments, you have crazy competitors, I mean, technology just changes insanely fast.
- 14:01SPEAKER_00: And one bad product can send you back years.
- 14:04SPEAKER_00: Yeah.
- 14:05SPEAKER_00: Years.
- 14:06SPEAKER_00: And second, he's willing to adapt.
- 14:08SPEAKER_00: Nvidia has gone through multiple eras, graphics, gaming, you know, general purpose GPU computing, AI, data centers, and the company keeps changing.
- 14:19SPEAKER_00: But it's still connected to that original idea of high performance computing.
- 14:23SPEAKER_00: Exactly.
- 14:24SPEAKER_00: And then the third thing is probably patience.
- 14:27SPEAKER_00: CUDA came out in 2006.
- 14:30SPEAKER_00: I mean, we were born in 2006.
- 14:32SPEAKER_00: That's that's been some time.
- 14:34SPEAKER_00: The big deep learning breakthrough with Alexnet came years later, and the massive generative AI boom came even later than that.
- 14:42SPEAKER_00: That's that's a really long time to invest in a technology.
- 14:45SPEAKER_00: Especially because most businesses are not worried about what they're going to make in the quarter.
- 14:50SPEAKER_00: Right.
- 14:51SPEAKER_00: And that's and that's what makes Jensen interesting from an entrepreneurship standpoint.
- 14:56SPEAKER_00: He's thinking about what computing might look like years into the future.
- 15:00SPEAKER_00: Now, obviously, not not every bet NVIDIA has made has worked.
- 15:06SPEAKER_00: And that that's important too, but the company has been willing to make those bets.
- 15:11SPEAKER_00: There's also something pretty unusual about Jensen himself.
- 15:14SPEAKER_00: He's still the CEO.
- 15:16SPEAKER_00: Yeah, I mean, that's another crazy part.
- 15:17SPEAKER_00: I mean, he co-founded NVIDIA in 1993 and has remained the CEO ever since.
- 15:22SPEAKER_00: I mean, just think about how much technology has changed during that that time period.
- 15:26SPEAKER_00: When NVIDIA started, the modern internet, it barely even existed.
- 15:30SPEAKER_00: No smartphones?
- 15:31SPEAKER_00: No smartphones.
- 15:32SPEAKER_00: No social media.
- 15:33SPEAKER_00: Nothing like modern social media.
- 15:35SPEAKER_00: I mean, gaming was completely different, AI was not even a thought, and Jensen has managed to lead the same company through all of those technological changes.
- 15:45SPEAKER_00: I mean, that's pretty rare.
- 15:47SPEAKER_00: So, I mean, I would say I guess the big question is, what can somebody actually learn from Jensen Huang?
- 15:53SPEAKER_00: Because obviously the lesson can't just be start NVIDIA.
- 15:55SPEAKER_00: Yeah, it's a little late for that one.
- 15:57SPEAKER_00: I mean, yeah, exactly.
- 15:58SPEAKER_00: I th I think one lesson is that your original idea doesn't necessarily determine what your company eventually becomes.
- 16:05SPEAKER_00: I mean, NVIDIA started around the computer graphics.
- 16:08SPEAKER_00: Today graphics are still extremely important to NVIDIA, but the company's technology is being used for way more than that.
- 16:15SPEAKER_00: Another lesson is that failure doesn't automatically mean your entire idea is bad.
- 16:21SPEAKER_00: NVIDIA, they they made their mistakes early on, but the company changed direction.
- 16:26SPEAKER_00: Jensen didn't just say, well, our first plan didn't work, guess we're done.
- 16:30SPEAKER_00: Which sounds obvious, but it's probably a lot harder when millions of dollars are involved.
- 16:35SPEAKER_00: I mean, yeah, absolutely.
- 16:36SPEAKER_00: I mean, and probably the biggest lesson is thinking long term.
- 16:39SPEAKER_00: You have to pay attention to where technology is going, not just where it is today.
- 16:44SPEAKER_00: And his personal story makes all of this even crazier.
- 16:47SPEAKER_00: It makes it wild when you when you look, I mean, when you look at it, when you put everything together, you've got this kid born in Taiwan, he spent some time in Thailand, then he comes to America when he's nine years old.
- 16:58SPEAKER_00: Nine years old.
- 16:59SPEAKER_00: Nine.
- 17:00SPEAKER_00: And then he goes to school in Kentucky.
- 17:02SPEAKER_00: I mean, what are they known for?
- 17:04SPEAKER_00: Chicken?
- 17:05SPEAKER_00: Yeah.
- 17:06SPEAKER_00: And then he goes to school in Oregon as well, works at a Denny's, and studies electrical engineering, and works for semiconductor companies.
- 17:14SPEAKER_00: I mean, it's just it's crazy.
- 17:16SPEAKER_00: And then he gets his master's from Stanford, then around 30 years old, he starts the giant, the giant company, dude, NVIDIA.