He made two Intel overnight, surpassing Musk in popularity

We have reinvented computing

Article | Chinese Business Strategy Zhang Jingbo

Taipei night market, two beautiful women live streaming.

Suddenly, a man in black poked his head and rushed to the camera. Under the invitation of the two, the man even sang to the audience.

At that time, they didn't know that the man in black was actuallyThe most powerful person in the global technology industry at present.

[Two intels worked overnight]

Yes, you read it right!

Not long ago, Wall Street performed miracles. A company called Nvidia's market value soared 24% overnight, approaching the trillion dollar mark.

In contrast, the market value of its competitors, Qualcomm, Intel and AMD, the world's three major chip giants, is less than half that of Nvidia.

In fact, on the evening of May 25th alone,The market value of Nvidia is equivalent to two Intel.The latter has been fighting for a full 55 years for today.

That's it, it's not over yet.

In the next two days, Nvidia's market value continued to soar, breaking the trillion dollar mark in the session,Becoming the first chip company in human history with a market value exceeding one trillion yuan.

A few years ago, Samsung and Intel were still fighting fiercely over who was the leader of the chip industry. Now, there's no need to argue, because the answer has become uncontrollable.

The strangest thing is that at a time when the global Semiconductor industry is full of sorrows, most of its peers are struggling, and Samsung Electronics has achieved the worst performance in 14 years.

What made Nvidia ignore the recession of the whole industry and rise against the trend?

The answer is the current hot ChatGPT.

Behind ChatGPT is a large model, and training large models requires super computing power.

According to Guosheng Securities' estimation, training the ChatGPT large model once would cost over $2 million.Most of this money is invested in Nvidia's GPU chips.

In fact, OpenAI has bought 25000 Nvidia GPUs to train ChatGPT.

And this is just the tip of the iceberg.

Since the beginning of this year, with the ChatGPT craze, global technology companies have been frantically chasing big models. The demand for data center chips from Microsoft, Google, and Amazon has exceeded everyone's expectations.

In China, there are currently over 30 companies that have released large models.

No matter how different the big models developed by these companies are, they are all competing for a product, that is, the GPU of Nvidia.

With orders coming like snowflakes, Nvidia's chips are hard to find.

It was described that Nvidia's chip has become the most scarce resource on the planet. Musk complained:It's even harder to deal with than drugsreach!

AIH100reach4

A few months ago, when ChatGPT first broke out, everyone thought OpenAI was the big winner. At present,The biggest winner is actually Nvidia.

Overnight work on Intel has also made Jensen Huang, the founder of Nvidia, a Silicon Valley, and evenThe most powerful person in the global technology industry at present.

Its popularity even surpassed Musk's.

This man, who comes from Qingtian, Zhejiang and always wears a black leather suit, is known as the most aggressive man in Silicon Valley due to punching Intel and kicking AMD.

But Lao Huang, who appears aggressive on the outside, hides a childlike curiosity deep inside.

In Taipei, he strolled the night market alone with a bag of big Fried Dough Twists in his hand.

reach

It is somewhat contradictory for a person with a trillion dollar market value and the most powerful global technology industry to have such a great curiosity about the streets and alleys.

But it's precisely thiscuriosityLet Lao Huang rewrite the history of the chip industry.

Do something different for the sake of humanity

On September 30, 2012, it was destined to be recorded in history.

On that day, a deep learning model called AlexNet won the global visual recognition challenge with a second place finish and became famous in the first place.

A new round of AI revolution has been triggered.

Unlike the traditional CPU, AlexNet uses two Nvidia GPUs to provide computing power for model training.

GPU is designed specifically for parallel computing and can handle a large number of identical computing tasks simultaneously. In contrast, the CPU needs to handle more complex and ever-changing calculations.

Someone used a metaphor, saying that the CPU is like an old professor who can calculate calculus and differentiation. GPUs, on the other hand, are like elementary school students who can only calculate simple addition, subtraction, multiplication, and division.

But if there is a job that requires calculating addition, subtraction, multiplication, and division hundreds of millions of times within 100, then dozens of elementary school students are obviously beating old professors. And AI model training precisely requires this kind of calculation.

In fact, as early as the beginning of this century, some people in American universities began using GPUs to train AI models. After all, compared to CPUs, it has lower costs and higher efficiency.

AlexNetcuriosity

reachAIreachAlexNetAI

In order to meet the revolution, Jensen Huang made a bold decision:

AllinAI enables every chip manufactured by Nvidia to focus on AI.

In 2012, it was an era of mobile internet and smartphones. Jensen Huang's decision to allinAI is more or less solitary.

But this is not the first time he has taken a different path.

In 2006, in order to provide GPUs with universal computing power similar to CPUs, NVIDIA introduced the CUDA parallel computing architecture, allowing programmers to call on GPU's computing power.

Prior to this, GPUs, like their name - graphics processing chips, could only be used for graphics rendering.

This decision made Nvidia stand out from other video card manufacturers at that time. It has thus rewritten the history of GPUs.

In this regard, Jensen Huang explained that the reason why Nvidia dared to take the lead in the field of parallel computing is that:

Our goal is to contribute to humanity. This means that we will not follow the same old path as others, but rather create products that others have never built before

curiosity

In 2012, after deciding to gamble heavily on AI, Jensen Huang continued to improve the performance and ease of use of Nvidia GPUs in terms of software and hardware. Especially in 2016, the first P100 chip was launched for AI computing scenarios.

On this basis, Nvidia has builtThe world's first AI supercomputer DGX-1.

However, Jensen Huang did not sell it, but donated it to a non-profit organization in San Francisco, which is today known as OpenAI in the global technology industry.

At the donation ceremony, Lao Huang took out a marker and excitedly wrote a sentence on the chassis:

For the sake of computing and the future of humanity, I donated the world's first DGX-1

At that time, apart from OpenAI, few people truly understood this sentence. In the eyes of most people, Lao Huang made a loss making business.

However, Musk, who was then the co-founder of OpenAI, specially tweeted thanks to Jensen Huang, because DGX-1 shortened the training time of OpenAI from one year to one month.

With DGX-1's OpenAI, confidence has exploded and we are determined to build a more powerful AI.

reachBuilding an AI system will forever change the way humans interact with computers.

That is, ChatGPT, which is popular worldwide today.

In the mobile The Internet Age, Microsoft, which is under the pressure of Google and Apple, will not miss such an opportunity and immediately decide to invest 1 billion dollars in OpenAI.

Later on, everyone knew about it.

Jensen Huang exchanged a DGX-1 worth 129000 dollars for the explosion of Nvidia.

Not only were A100, H100 and other GPU chips looted, but more than half of Fortune 100 companies installed Nvidia DGXAI supercomputing.

We have reinvented computing

curiosity

20reachcuriosity

A normal adult may not necessarily think about time and space issues

curiosity

1993!

curiosityGraphics chips will shine brightly in the future.

2006GPUcuriosity

curiosityCUDA2012AI

curiosity

During the 10 years since the launch of CUDA architecture, the whole Wall Street has been asking Nvidia a question: Why do you make this investment? For us, its market valuation is 0.

Jensen Huang remained unmoved. He always insisted, even at the cost of tens of billions of dollars.

For this reason, Nvidia's profits suffered a huge blow. The company's market value hovered around 10 billion dollars for a long time, and it was beaten by its old rival Intel.

AIreachGPUCUDA

During this period, Apple, Google, AMD and many other technology giants tried to copy Nvidia's successful experience, but it was always difficult to challenge the latter's monopoly position.

curiosityIn the eyes of many people, it is an important reason for Jensen Huang to lead Nvidia to constantly create miracles.

This quality first appeared in his youth.

!

Ordinary children cannot do such things.

The stubborn old Huang eventually achieved a great success.Today, Nvidia has a 60% market share in the AI chip market, which is an absolute leader.

One fifth of this comes from the Chinese market.

Although he doesn't speak Chinese very well, Lao Huang describes himself as having Chinese blood in his bones.

Not long ago, he reminded Washington that Nvidia can produce chips in other places, but the market in Chinese Mainland is unique and cannot be replaced.

In the future, as GPU replaces CPU, Nvidia may usher in a greater explosion.

reachNvidia has made contributions to computing and the future of mankind.

In the past decade, with the explosion of AI and Moore's law approaching the limit, there has been a huge gap between the supply and demand of computing power.

And GPU largely alleviates people's computational anxiety.

60IBM 360!Lao Huang excitedly said.

This man, who has been eager to become the most influential technology company in the world since his first day of entrepreneurship, and make significant contributions to the computer industry and even humanity as a whole, once stated in an interview:

The most important thing in life is not to have regrets

reach!

Welcome to follow 'Chinese Business Strategy', learn about influential figures, and read the legends of strategy.

All rights reserved, unauthorized reproduction prohibited

Part of the images are sourced from the internet

If infringement is involved, please contact for deletion