Jensen Hang has worn the same leather

jacket for 30 years. Might give the

impression that he's laidback, just a

normal dude, but he said publicly it's a

brand decision. That tells you

everything you need to know about the

guy. Nothing Jensen does is accidental.

[music]

Before Nvidia, Jensen was a middling

engineer at a few companies that you've

probably heard of. He wasn't a taste

maker. He wasn't a titan of industry. He

was watching the whole thing go by from

the middle with relatively little power.

He was born in Taiwan and moved to the

US at the age of nine. He was raised

partly in Kentucky, partly in Oregon, so

he doesn't have one of these coastal

elite origin stories like most of our

other tech overlords. He studied

electrical engineering at Oregon State

and then Stanford for his masters. Solid

credentials there. No flags on the play.

Pretty straightforward. And then he

worked at AMD and LSI Logic as a

microprocessor designer. And I want us

to keep that in mind as we go through

the rest of his biography here. This is

a very unique thing about Jensen

compared to most other people that we

see in the AI space. If you've watched

my expose on Sam Alman, Sam Alman, since

he's been an adult at least, has never

had a real job. He's never really been

hands-on keyboard. He wasn't a prolific

coder. And you see this story repeat

itself over and over again in our

industry. And it's always been very

confusing to me because you essentially

have people that can talk well,

understand some business dynamics, and

can con folks at the helm of these

massive companies, but they've never

actually spent time doing the work that

they expect their workforce to do. It's

very peculiar. Jensen was actually

working on microprocessors. So, at a

fundamental level, he understands and

has lived the life of the people that

work for him. His ascent to success

started very late and it was very rocky.

He didn't actually found Nvidia until he

was 30 years old. Now might sound young

but for reference again contrast with

Sam Alman who by that point had already

started to make a name for himself in

the valley. And he co-founded Nvidia

with two other dudes Chris Malikowski

and Curtis Prium. And this was at a

Denny's in San Jose, which is a great

California tech origin story. Honestly,

Nvidia by no means was an instant

success. It actually nearly died two

times along the way. And you need to

understand these failure points to

understand Jensen's mentality and the

way that he runs the company at present

day. This story is full of so many

flops. It's difficult for me even to

tell. It makes me uncomfortable as

somebody who works in tech to even think

about having a company that is going

through this. But their first failure

point is with Sega. They go under

contract to develop a chip for the Sega

Saturn, which if you are younger than me

is a very old gaming console and they

botch it. They don't make something

that's up to spec that Sega wants and

they lose the contract. Sega pulls out.

This is like a major contract for them.

Financially, they have at this point

about a year of runway left that they

can float and make payroll. And instead

of continuing to work on that chip that

they were developing, you know,

repurpose it, maybe sell it to Nintendo,

maybe sell it to another gaming console

manufacturer or even build their own,

Jensen makes the baffling decision to

say, "Cancel it. We're not even going to

work on that anymore. We're going to

start work on this completely new chip

called the NV1." And guess what? The NV1

flopped commercially also. So, there's

another failure. But they managed to

just scrimp and save and barely keep the

company alive through this period until

they get to the next thing. This brings

us to a point that anyone who's spent

time working in Tech in the Valley will

understand and will tell you. The

companies that survive are the ones that

ship something. The companies that

survive are the ones that ship

something. Most of the companies that

fail, and this is an interesting point

because before I worked in tech, this is

how I thought, you know, you got to do

it. You see Apple, you see all of these

titans perfecting something. They

release a perfect version of something

and it's unassailable. It's awesome.

That's not how it works. That's not how

it works in real life. The company that

ships something has a chance at

surviving. A slim chance. They have a

slim chance. The company that is working

on perfecting something before they have

shipped going to fail. Going to fail 10

times out of 10. And this is the lesson

Jensen takes away for his leadership and

business acumen from this near failure.

And we know this because he talks about

it publicly. The second time Nvidia

comes close to being wiped out is in

1997. I wonder, are you familiar with

the graphics card company 3DF? Yeah,

probably not, unless you grew up very

nerdy, uh, like myself and are my age or

older. And I'm about to tell you why.

It's because of Nvidia. So, 3DFX is like

the dominant GPU company. They have

market share. They have products

bringing in a bunch of cash. They are

just this paragon of GPU development for

the era. Nvidia is still not nearly a

successful company. They're hemorrhaging

cash. They don't have enough money

coming in the door. And Jensen decides,

"Let's drop everything and work on this

new Revo 128 chip. Let's put all our

eggs in this basket and just go for it."

They ship it in just 4 months. And to

make a long story short, you probably

don't know the name 3dfx if you're

watching this, and Nvidia is worth three

trillion. So, the play worked out for

them. The thing Jensen doubles down on

and takes away from that is he's the

type of businessman that optimizes for

speed over perfection. If he sees an

opportunity, he is willing to take the

bet and drop everything and go after

that opportunity, even if it seems crazy

to other people. Then later on in 2006,

Jensen makes another very risky

trade-off. He decides to invest a ton of

money into a technology that has no

clear payoff, no clear market need, and

it costs hundreds of millions of dollars

of the company money to work on. His

shareholders hated it, and it's the

reason Nvidia is synonymous with AI

today. The technology was CUDA, launched

in 2006. It was a way for engineers to

be able to directly leverage the compute

resources on a GPU chip for different

tasks, not dealing with graphics. This

may seem obvious today in hindsight, but

in 2006, if you had a bunch of money

invested in Nvidia and they're like,

"Hey, we're going to blow tens of

millions of dollars on, I don't know,

letting developers access direct compute

resources on the GPU. How's that sound?"

You'd be like, "It sounds bad." Because

the reason you develop a GPU in 2006 is

for gaming. It's for 3D graphics. That's

what a GPU is. Graphics processing unit,

bro. It's literally in the name.

Graphics processing unit. Not developer

playground. Not hack around and see what

happens with our GPU. Graphics

processing unit. Okay. So Jensen's

obviously grilled about this by people

that have a lot of money in Nvidia and

by the press like why are you

hemorrhaging cash on this thing that

doesn't have a clear payoff and he's not

even able to really articulate why he's

doing it. He publicly stated that he was

taking a bet that the GPU would be

useful for something other than graphics

at some point. But by his own admission

in interviews and in public statements,

he he didn't know exactly what that was

going to look like. He just believed

that it would be useful at some point

and it's a crazy bet to make but there's

crazy upside. It took a long time for

CUDA to pick up steam but researchers

started using it, papers started being

written about it, tooling starts being

built on CUDA and in the 2010s when deep

learning starts kicking off. This

becomes the de facto platform that

researchers, engineers, scientists use

to develop machine learning algorithms

and Nvidia is the only one selling it.

So 2006 we have nothing maybe some

hobbyist and strange researcher

interest. Around 2010 2011 things pick

up a little bit and then 2012 is the

defining moment. A paper comes out on

Alexnet. This is a big research paper.

It is one of the things that ostensibly

kicks off the modern deep learning era

and everything in the paper is built on

CUDA. So it sets the standard. So that's

a bet Jensen made 6 years prior with no

obvious or immense payoff burning a

bunch of money and finally they become

dominant in this totally new and

emergent technology space and they're

uncontested because nobody else has

burned as much money developing this.

The ripple effects of this are still

obvious today. You have competing chip

manufacturer AMD. They make chips with

stats on par with Nvidia's chips, but

all of that compute is locked up because

they h they don't have decades of

developing tooling and ways to interface

with their primitives on the GPU for

machine learning. Even if AMD dropped

everything today and said, "We are not

doing anything with gaming, graphics,

anything like that anymore. we are just

all in on machine learning. We're going

to build something way better than CUDA.

It's going to be way better, way easier

for developers to use. It wouldn't

matter because at this point for over a

decade, the de facto standard for

engineers working on machine learning on

AI has been to develop on CUDA. And it's

not just the opportunity cost of those

engineers switching over, learning a new

framework, something like that. Once a

primitive comes out like this, like

CUDA, there's tech and frameworks built

on top of it that make it easier and

easier for engineers to develop on that

framework. So, you have over a decade of

customuilt tooling, fixtures, things

that just make your life easier when

you're working on this stuff built on

CUDA that AMD doesn't have. Even if they

threw everything at it, they can't buy

developer market share like that. If you

want a metaphor, it's like you move into

a new furnished house. You buy it and

it's furnished. I know it's kind of

weird, but walk with me here. You buy a

new furnished house and you live in it

for 10 years. You're going to swap out

the furniture. I don't like that couch.

Let's put some art up over there. The

arrangement of this room is all wrong.

And in 10 years, that house is going to

be like very easy and comfortable for

you to come home to, to cook in your

kitchen,

easy for entertaining, easy for your

hobbies because you've customized it

over 10 years. You you have 10 years in

the game. AMD is giving you that nicely

furnished house if you're a developer,

but you just moved in. You just moved

in. The configuration isn't optimal. You

don't like some of the stuff in there.

Uh even if the house is nicer, even if

the house is nicer, it's not set up for

you yet. Fast forward to today. How is

Jensen running the company? And he's one

of these case studies that I've I've

looked into. I've studied leaders in the

tech space. As a tech leader myself, I'm

always looking to learn from other folks

that have made their way, especially

self-made people, not so much coastal

elite, easy come up networking conmen,

but these self-made people. And so I

looked into Jensen, and his management

style is very, very strange. I I'm not

sure I agree with it. It's very strange.

He has 60 direct reports. And so when I

say direct reports, these are people

that he manages directly. So those

people may manage other people, but I'm

talking like direct reports. They would

say Jensen is my is my boss. He's my

manager. If you don't come from a tech

background or you're not that familiar

with tech leadership, this is a lot. I

mean, it is a crazy amount. The most

I've ever had was 20 and that was absurd

and unsustainable. probably the perfect

size is I mean I don't even know if it's

a perfect size. You probably max out

around eight or 12. It takes a lot of

emotional energy to deal with that many

people if you're going to be a good

manager. And ultimately you only have so

much emotional energy every day and

every week. So, if you're going to do a

really good job uh as as a manager, kind

of the general path of manager, I would

say top out around like 8 or 12 because

not only it's not like you're managing

uh CVS or something where it's just like

who's here, everyone has clear roles and

responsibilities. The company defines a

lot of these things. It's like you got

to do technical architecture. You got to

get in the weeds with people. There's

the personal issues that come up within

the team, outside of the team. So it

it's uh it's really it's a it's a huge

gig. It's a huge gig to manage

technicals. And so 60 is I mean it is

baffling. Nobody nobody in this industry

manages 60 people directly. Nobody other

than Jensen. But the way he's able to do

it, he's not some exceptional human

being because of this. That's not the

case I'm making. I I'm saying the way

his entire management philosophy is set

up is completely different than anybody

else I've seen in tech. Like I said, I

don't not sure I really agree with it.

With these 60 direct reports, he has no

one-on-one meetings. His reported

philosophy for not doing this is that

private information creates political

advantage that can be unfair. So, he

does everything in a group setting where

context is shared between everyone. But

he's not some utopian manager. He said

he wants all of his employees to feel

quote a bit of fear. He defends this by

saying the fear he wants them to feel is

the fear of failing and the fear of not

shipping something. His critics say that

this creates a pressure culture that

burns people out and forces them out of

the company. His proponents say this is

why Nvidia ships and is dominant in the

industry. I want to take us back to the

leather jacket. As I mentioned, he's

worn the same leather jacket for 30

years. This is an intentional choice. He

treats his personality as a product, as

a brand. And I see this kind of

hard-nosed personality first, carefully

manicured personality that the public

sees. Personality first kind of deal

come through in his management style as

well. It's very strange. Very strange in

tech. Can't name another person that has

60 direct reports. And certainly most

CEOs don't say that they want their

employees to feel fear. And whether

right or wrong, whether you agree or

disagree, on a very pragmatic level,

this makes Jensen really, really hard to

replace at Nvidia, his stakeholders have

to know and be aware that this is a

massive succession risk for one day when

he steps down or decides to leave the

company. How do you replace a guy like

this that has carefully crafted this

whole persona and management style

that's very different? Nobody else has

this skill set. And one of the things I

I mean by that is Jensen's org is also

very flat. And what I mean when I say

flat is, you know, usually you have like

a CTO,

senior VP, VP, director, yada yada, down

and down and down through the ranks,

manager, manager, manager, manager. And

their organization is very flat. So

there's he has you know 60 people under

him and then presumably those 60 people

have wide swaths of direct reports under

them too. So this telegraphs down and

you don't have a massive reporting chain

like you'd find in any other company of

that size. And to add to this, Jensen

involves himself at product decision

levels that are very strange for a CEO.

Like he might comment on chip design. He

might comment on exact marketing

campaigns. At a company of this size, a

CEO almost becomes more like a

figurehead. And Jensen is very much

involved in some of the day-to-day,

which again makes him very difficult to

replace. and you wonder if it's

intentional. By the numbers, Nvidia's

transformation is one of the fastest and

most profitable in tech history. But

their financial filings also tell you

where the risk is. Remember 5 years ago,

gaming was the primary profit source for

Nvidia. Now, as we've discussed before,

when they've decided to not release new

graphics cards for private consumer

gamers, it's like a rounding error for

them. It's kind of like this token thing

that they do on the side. It's no longer

a big revenue driver compared to the AI

space. What and what the company will do

5 or 10 years from now is anybody's bet.

It's a total wild wild card. And I'll

tell you why. Because Jensen has made

this crazy pivot from the the company

going from a gaming company to a machine

learning AI data center company. And

tomorrow it could it could be anything

else. Win or lose, who knows? But as

evidence of this, we have in 2020 Jensen

in Nvidia is thinking we really want to

vertically integrate. We want to control

the whole stack. Whole stack. I mean,

you get the sense that this guy wants to

be in charge of the mines that are

getting the raw materials for the chip,

be in charge of the chip creation,

everything, everything, nose totail. And

as evidence of this, in 2020, Nvidia

tried to acquire ARM, the chip

manufacturer. They were owned by

SoftBank at the time. and the FTC tied

up the whole thing in a bunch of

litigation. The regulators were not

happy with this. Is it going to create a

monopoly? Does that give Nvidia an

unfair control of this market segment?

So, SoftBank withdrew from the deal a

couple years later and the whole thing

fell through. But we can see clearly the

ambition that Jensen has that Nvidia as

a company has to expand their remit to

be fully vertically integrated and to

maybe build something completely

different tomorrow than what they're

building today. But will Nvidia even be

around? And if they are, will Jensen be

in charge of it at that point? We have

documentation on form for filings from

Jansen and other key exeacts at Nvidia

executing quite large sales of their

shares in the company. Legal, somewhat

standard, but worth noting when the

company is on a generational bull run.

And you got to be wondering, are they at

an all-time high? Do they see the

writing on the wall for this? And

they're starting to cash out. So, here's

what most pieces on Jensen fail to

assess. In summary, did he make a an

incredibly good bet? He just knew

something was coming down the pipeline.

This is business intuition. Call it what

you will. Did he make a great bet or is

he very reckless? And I hate to say it

because all of my exposees are like, you

know, revealing all this dirty laundry

on people, but I think it was good

business intuition. And I'll tell you

why. I'll tell you my case. And I want

to hear your counterpoints in the

comments as well because I go back and

forth on this. But today, what I think

and how I feel about it is that when he

got the CUDA tooling going, it was a

visionary move. It was a visionary move.

It wasn't just luck. Predicting the

future is luck. But he didn't only

predict the future. He doubled down on

this for 18 years. So if you've never

been in leadership, it is constantly a

struggle to get money to do the things

that you want to do. Money is the

constraining resource a lot of the time.

Unless you're at like a early stage like

high VC funded startup, uh, which is

super fun and and a totally different

vibe. resource scarcity. It's not even

scarcity. It's just like there's only so

much to go around, right? And it's hard

to get a big bag of money to go and do

what you want to do. You got to convince

uh if you're at the CEO level, you got

to convince your board, you got to

convince your peers, you got to convince

a lot of people that this is okay what

you're doing and it's good to burn the

money on this. And so it wasn't Jensen

was just like, I think we'll make some

developer tooling. Maybe that's how it

started. But then every quarterly

earnings, he's probably getting

questions about this. His board members

are like, "What the hell are you doing,

Jensen? We're just pissing money away on

this thing that has no ROI. It has no

market share. It's not helping get

people on the platform. What are you

doing? You got to shut it down." And he

defends this for 16 years. For 16 years

before it has a massive payoff and now

he's the one laughing all the way to the

bank. And so my hunch is that this is

not just like a narcissist's whim that

all my idea is good and I'm going to

keep with it. Like he's getting badgered

for this for over a decade to cut it out

and he defends it. So for for my money,

I do think that that's I think that

that's correct intuition about where the

market is headed. That's not to say

there are not very real risks to

Nvidia's company right now. Every major

hyperscaler, Microsoft, Google, Amazon,

Meta, they are all working on their own

chips and most of them are even working

with external chip manufacturers to try

to replace Nvidia because they control

so much market share and so much of the

AI spend. They're looking to liberate

themselves from that to give more market

competition to themselves to drive

Nvidia's price down. And again, this is

not like tinfoil hack conspiracy. All of

those hyperscalers, they mention this

publicly in press releases, in their

earnings documents that like, hey, we

are trying to do this to cut our GPU

spend. This is their stated purpose. You

even have companies like Chat GBT

working with Cerebrris, which is a new

chip manufacturer. They make a new type

of chip that is very fast at AI

inference. You can watch my video on

them if you want to learn more about

that. And even a new company that spun

up very recently in Canada that's making

headlines right now, Talos, which is

making these chips that are just

lightning fast for inference, like

hundreds of times faster than Nvidia

GPUs on inference. So Nvidia's moat is

shrinking and a lot of these other

manufacturers are one or two

breakthroughs away from being able to

replace a massive amount of Nvidia GPUs

that they need to do their business.

Suddenly Jensen and the other leadership

stock sales make a little bit more sense

in that context. Yeah. As I mentioned

before, the succession problem is very

real too. How do you replace a guy who

has 60 direct reports? Nobody has that

skill in the industry that is involved

very much in the day-to-day of product

decisions and isn't a total con man.

Like he understands how chips are built

because he built chips unlike a lot of

other CEOs of these companies. I was

actually surprised while doing my

research on this piece to find out that

Jensen is 61. So, he's getting up there.

He cannot be CEO of the company for a

very long time. The board and the key

shareholders need to start thinking

about who who are we going to replace

Jansen with? How how do we find a

replacement that doesn't crater the

company? Let's end by coming back to the

leather jacket. The leather jacket

brings everything together. Same leather

jacket 30 years. That's discipline.

discipline and fidelity to a cause that

he believes in. Making his own brand

probably was goofy when he started doing

it 30 years ago. People are like, "Dude,

come on. That's tryh hard." Now it's

iconic. Nobody questions it. I'll leave

it up to you in the comments to discuss.

I want to hear your points on both

sides. Is Jensen a visionary or is he an

extremely disciplined opportunist

that executed faster than anyone else

when he saw Elaine? Thank you for

watching. If you are not subscribed,

let's fix that. Hit the button below.

Hit the bell to be notified when new

posts come out. And if you want the

facts and figures on all of this, if

you're watching this on February 24th,

the day this is coming out, Nvidia's

earnings are tomorrow. We're going to

have the video here within an hour or

two of earnings, you want to be

subscribed and notified for that when it

comes out. And also be sure to sign up

for the newsletter to get the facts and

figures behind the AI bubble.