“Jensen, aren’t you a genius?” “Am I a genius? Well, I guess so.
There are many people in the world who are smarter than me.
But what is clear is that there is absolutely no one who works harder than me.” This book is a compilation of vivid voices and NVIDIA history through direct interviews with Jensen Huang and over 100 NVIDIA executives.
This is a documentary story that summarizes the 31-year story of NVIDIA in its most complete form, spanning 448 pages, after more than a year of in-depth research, and a management book that captures Jensen Huang's corporate philosophy.


Barron's senior reporter Tae Kim interviewed over 100 people, including Jensen Huang, co-founders Malachowski and Priem, early venture investors, and representatives from competitors.
As a result, the incidents and accidents of the time, the behind-the-scenes development, and the horrific and thrilling anecdotes that went back and forth between the extremes of 'bankruptcy crisis' and 'market dominance' were realistically depicted down to every single conversation.
From the day in 1993 when three entrepreneurs dreamed of starting a business while drinking unlimited refill coffee at Denny's restaurant to today, the story of how they opened the 'GPU era' after repeated failures and successes, successes and failures, is incredibly fascinating.
You can get a glimpse into various episodes that have never been exposed in any book or article, Jensen Huang's characteristically straightforward speech that combines humor and decisiveness, whiteboards and top 5 emails, public criticism of internal politics, and a unique and fierce organizational culture represented by the "speed of light."

However, it has now grown into a major company that provides high-performance processors for the era of AI.
NVIDIA's processor architecture is optimized for AI tasks because it excels at performing massive, concurrent computations essential for training and executing large language models.
NVIDIA recognized the importance of AI early on and has been investing proactively for over a decade, including hardware enhancements, AI software tool development, and network performance optimization.
Thanks to this foresight, NVIDIA's technology platform is perfectly positioned to capitalize on today's AI era.
--- p.15, from “Meeting Jensen, June 14, 2024”

I introduced myself as a PC gaming enthusiast who had been building my own computers since the 1990s.
He said that he first learned about NVIDIA when he was looking for a graphics card for his PC, and that he always chose NVIDIA products.
And he said that his first success was his choice to invest in Nvidia early in his career while working for a fund on Wall Street.
“Well done,” Jensen joked, his expression unwavering.
“NVIDIA was my first success too.”
--- pp.18-19

But sometimes communication devolved into outright conflict.
“Chris and I used to have very rough and ruthless arguments.
“We didn’t have any physical conversations, but we fought, yelling and screaming at each other,” Priem recalled.
“Chris was trying to get something out of me about the chip decision.
At some point, I couldn't calm down even after he gave me the answer he wanted, so I tried to continue the fight.
Then Chris says, 'Stop, stop, it's over now.
“I used to say, ‘I got the answer I wanted.’”
Then, Priem stormed out of the office, leaving the other team members to look at Malachovsky with worried eyes.
Whenever one of them hesitated and asked if the team was now disbanding, Malachovsky would always reply:
“This team is fine.”
--- p.64

“I used to tell people at AMD, Intel, and other companies that if they wanted to see how Nvidia was doing, they should come to the Nvidia office parking lot on the weekend.
“It was always packed with people’s cars going to work,” said Riva.
Even the marketing department came in every Saturday, and working 60 to 80 hours a week was common.
Andrew Logan, a marketing director at Nvidia, remembers leaving the office with his wife at 9:30 p.m. to go see the movie Titanic.
On the way out the door, a colleague shouted.
“Hey, Andy, are you taking your half day off today?”
--- p.147

But on the other hand, the saying, “We have 30 days left until we go under” was also true.
In the tech industry, a single wrong decision or product launch can be devastating.
NVIDIA was lucky enough to survive the disasters of NV1 and NV2 twice, and succeeded with RIVA 128 with only a few months left in its lifespan.
But this luck couldn't last forever.
A good corporate culture can protect your company from the serious consequences of most mistakes.
However, risks such as mistakes and market downturns were unavoidable.

--- p.150

As Steve Jobs prepared to leave, he told the NVIDIA attendees:
"You definitely need to do more in the notebook market. ATI is giving Nvidia a hard time in that market," he said, referring to ATI, which has become Nvidia's main rival since the demise of 3dfx.

Chris Diskin answered without a moment's hesitation.
“Steve, to be honest, I think you’re wrong.”
The conference room was plunged into silence.
Jobs looked at Diskin with his characteristically intense gaze and said:
“Can you tell me why?” Diskin could tell that not many people had dared to contradict Steve Jobs at that moment.
Jobs was clearly expecting a compelling answer, and Diskin had one.

--- pp.226-227

NVIDIA called this chip programming model "Compute Unified Device Architecture," or CUDA for short. CUDA allowed not only graphics programming experts but also scientists and engineers to harness the GPU's computational power.
With the help of CUDA, they were able to manage the complex combination of technical instructions to execute parallel calculations across the hundreds (and later thousands) of compute cores of a GPU.
Jensen believed that CUDA would allow Nvidia to expand its market reach into every corner of the tech industry.
It wasn't new hardware, but new software that was going to change Nvidia.

--- p.247

It seemed to Ernst that Jensen was growing weary of the investors' aggressive questioning and would soon leave.
So he decided to ask another question.
“Jensen, I have a two-year-old daughter.
So I just bought a new Sony A100 DSLR camera, and I periodically download photos to my Mac and lightly edit them in Photoshop.
But every time I open a high-resolution image, my Mac computer immediately slows down.
“It’s worse on my ThinkPad laptop. Can the GPU help with this?” Jensen’s eyes lit up.
“Please don’t write about this story, it hasn’t been released yet.
Adobe is also our partner. With CUDA support, Adobe Photoshop can instruct the CPU to offload work to the GPU, resulting in significantly faster performance.
This is what I call the ‘age of GPUs.’”
--- p.254

Nvidia has been aggressively expanding into new market segments to diversify its revenue streams so that its overall business is not swayed by weak demand in certain areas.
One of them was console graphics, and even though Microsoft had already signed a deal with another company to build the graphics chips for its Xbox console, Nvidia pushed for the contract again.
And despite having virtually no experience with Apple's Macintosh architecture, Nvidia managed to break into the Mac lineup.
Additionally, NVIDIA was able to enter the professional workstation market with the Quadro line of products aimed at the CAD market that it had originally intentionally avoided.
Now, Jensen has created an entirely new computing technology based on GPUs.

--- p.263

Catanzaro said.
“The AlexNet paper by Alex Krizhevsky and Ilya Suzkever really shook the world.
One thing people often forget is that this paper was fundamentally a systems paper.
It wasn't about some cool new mathematical concept of how to implement artificial intelligence.
What they've done is leverage accelerated computing to dramatically scale the datasets and models that can be applied to specific tasks.
that
And the results have been remarkable.” This research has ignited Jensen’s interest in artificial intelligence.
he is
I started talking to Bill Daly frequently and focused on how deep learning, especially GPU-based deep learning, could be a huge opportunity for NVIDIA.
There was considerable debate within management over this issue.
Even among Jensen's key executives, some opposed further investment, saying deep learning was just a passing fad.
But CEO Jensen decided to ignore these opinions.
“Deep learning is going to be huge,” Jensen said at a 2013 executive meeting.
“We have to put everything on the line here.”
--- pp.344-345

Nvidia's second, and lesser-known, strength is its pricing power.
Nvidia doesn't have high hopes for a mass-market product.
This is because popular products face pressure to lower their prices as competition intensifies.
Rather, Nvidia has been pursuing the opposite strategy of raising prices since its inception.
“Jensen always said, ‘We have to do what other people can’t do.
We have been saying, 'We have to provide special value to the market.'
And he thought he could attract top talent to the company by doing revolutionary work in cutting-edge areas,” said Nvidia executive Jay Puri.
“Our corporate culture isn’t just about chasing market share.
Rather, it is about creating a market.”
--- p.392

Dierks was overcome with excitement.
He resigned the next day and told Pellucid's top executives that he was going to Nvidia.
Then the executive got mad as hell.
“I knew you could go! I’m going to sue you and Nvidia.
“I won’t let you ever set foot in Silicon Valley again.”
He told Dierks that Nvidia would be scared away by these legal threats.
Nvidia was only a one-year-old company at the time and its financial situation was limited.
But when Dierks told Jensen about this threat, the Nvidia CEO was completely unfazed.
“Try it,” Jensen replied.
At that moment, Dierks realized what kind of boss he wanted to work with.
He accepted NVIDIA's offer and has been working for NVIDIA for over 30 years.

--- pp.414-415

Jensen rewards performance with stock grants based on how valuable the employee is to the company.
Former HR chief John McSorley said.
“Jensen treated Nvidia stock like his own blood.
He examines stock allocation reports meticulously, as if looking at them under a microscope.”
Stock compensation is in the form of restricted stock.
When an employee joins a company, they are provided with a securities account.
After the first year, the employee receives a lump sum of stock equivalent to one-fourth of the initial stock grant.
For example, if the total promised amount was 1,000 shares, the employee would receive 250 shares at this time.
After that, employees will receive one-fourth of their assigned annual stock grant regularly every quarter.

--- p.415