Jen-Hsun Huang's Fortune Surpasses $100 Billion for the First Time: From Night Market Boy to "Silicon Valley God of War"

In the surging artificial intelligence circle, NVIDIA, which specializes in training large models, has become the focus of the industry's close attention. Generations of updated chips, keep the leather godfather Jen-Hsun Huang's "turnaround rate" and Nvidia's trillion-dollar market value.

The basic hardware is polished to the extreme, so as to enhance the arithmetic power at high intensity. Jen-Hsun Huang, leading NVIDIA, contributed the industry's most thirsty chips to support the continuation of AI training, and almost became the biggest winner, continuing to guard the ring for the hard-won trillion-dollar market capitalization.

However, NVIDIA became "AI Apple" road is not all the way smooth, Huang Jen-Hsun's NVIDIA entrepreneurial history, not only among the trillion-dollar market dream and success, but also in the crisis of the continuous strategic revision, retreat for progress, predicament to survive.

Groups of skeptics, facing bankruptcy, no one asked for, NVIDIA's first-generation chips are subjected to cold eyes. Timely influx into the new track, the key moment to make the right choice, but again and again to save NVIDIA in the fire, send it to the altar today.

Looking back on Jen-Hsun Huang's entrepreneurial path, NVIDIA has never stopped running from a small company with dozens of people to the top of the industry. How does NVIDIA get through the folding and siege moments? In the process of running, how to maintain momentum and continue to seize new opportunities?

Compared with Intel and AMD, NVIDIA is more than twenty years young, and GPU products, all rely on TSMC and Samsung such as foundries, whether liquid or fixed assets, it is difficult to be compared with other semiconductor giants, why it is it took the trillion-dollar this crown.

Is Silicon Valley a myth again, or is Wall Street looking the other way?

There's a lot of flip-flopping about how NVIDIA succeeded, talking about all the decisions it made, but the heart of the matter isn't what NVIDIA got right, it's how it made sense of the mistakes before that, which is especially important for semiconductor companies.

Nowadays, we don't like to talk about sunk costs, no longer counting those costs that can't be recovered, the so-called king or loser, the successful ones do all the right things, and the losers are worthless, we only pay attention to the results and the present moment, as if we were driving a car with no rear-view mirror and no brakes.

The parties involved obviously don't think so. On the eve of NVIDIA's market capitalization about to break the trillion-dollar mark, Jen-Hsun Huang, as an honorary doctor at National Taiwan University, was invited to attend the university's graduation ceremony, and in his speech speech, he shared three stories of his failures, and while everyone was looking at his success and talking about his multiplication of his wealth, he was reminded of his own grizzled past.

1,Tainan to Silicon Valley

Jen-Hsun Huang is not a Silicon Valley legend like Moore or Noyce, when he joined the industry, Intel has been successful, Microsoft has gone to soar, even AMD has been a large semiconductor company.

Born in Tainan, he moved to the U.S. at the age of 9. When he first arrived, he and his older brother spent two years in a rural Kentucky boarding school, and in an interview with USA Today, Huang Jen-Hsun recalled that he and his brother were at a naughty age, often climbing on roofs and drilling through windows, stealing candy, etc., and that, following the gang of kids from the correctional facility, they picked up on all of the profanity. "Everybody smoked, so I smoked."

He learned to be independent, adaptable and strong over the course of two years. With his parents' official move to Oregon, he transferred to a regular school and, like many Chinese children of his age, became enamored with table tennis, a sport in which he was quite gifted, participating in the U.S. Table Tennis Championships at the age of 15, and at the age of 15, placing third in doubles at the U.S. Open Table Tennis Championships.

Had it not been for his academic aspirations, I'm afraid Jen-Hsun Huang would have become a professional table tennis player. After placing well in the Open Championships, he chose to hang up his racket and began to focus on his studies, and at the age of 16, he was admitted to Oregon State University, and from here, his personal history took on a touch of legend.

In Oregon State University, Jen-Hsun Huang chose to major in electrical engineering, which is not difficult to understand, the revolution in the U.S. semiconductor industry is on the rise, which is the hottest specialty of the year, and he ultimately got what he wanted, after graduation in 1983, ran to Silicon Valley, became an AMD chip engineer.

And during his four-year college career, Hwang Jen-Hsun gained not only knowledge, but also his first girlfriend and current wife, Lori, to whom he promised that by the age of 30, he would own a company of his own.

And the most interesting thing is that he then jumped to another chip maker LSI Logi, after serving as an engineer, director of the business department, the general director of the system integration chip, etc., halfway through the Stanford Masters of Science in electrical engineering, and finally in 1993 began to start a business, founded the company that is Nvidia, February 17, 1993, Nvidia began business, this day is Jen-Hsun Huang's 30th birthday, the perfect fulfillment of his promise made at the age of 20.

And Jen-Hsun Huang for NVIDIA to choose the direction of no other, is that we are familiar today to the graphics processor, according to Bay Area company display chip industry analyst Jon (Jon Peddie) recalled: "At that time, Jen-Hsun Huang also called specifically, consulting about the display chip market and the future trend, I told him, this market has not yet started already a mess, there are now nearly 30 companies, you'd better not do this."

What were the changes in Silicon Valley in 1993? Intel has just launched the Pentium series of 80586, AMD is as before the 80486 processor, released Am486, the two processor giants are fighting each other, and Apple is and IBM to promote the development of PowerPC, the CPU of the three worlds is taking shape.

However, the future of graphics processors is not clear: SGI, the pioneer in the field of graphics, only provides graphics gas pedals for some workstations; LSI Logi, the old owner, has not yet released a dedicated graphics processor chip; ATI, also founded by Chinese, has not yet come up with any concrete products and is muddling along in the OEM market; the veteran hardware manufacturer Matrox has not fully committed to the industry; and 3Dfx, which started the 3D revolution, has not yet been established. 3Dfx has not yet been established, the market is chaotic, it is a good time to start a business.

Jen-Hsun Huang ultimately named the company Nvidia, and while it originated from Invidia, the Roman name for the ancient Greek Titan god Nemesis, the mythological embodiment of hatred and jealousy, he focused more on the first two letters of the word, N and V, or next version, representing his ambition to dominate the future trends in technology.

With six letters and a green eye, NVIDIA started a voyage into 3D graphics.

2,suffer two setbacks

However, as a rash NVIDIA did not have the expected success, and it nearly killed itself with its first mistake.

First of all, it should be noted that in 1993, there is nothing on the market for 3D graphics processing standards, we take different technical routes, NVIDIA's choice is quadrilateral texture mapping, which is relatively low-cost, based on this technology, it released its first display chip product NV1 in 1995, graphics processing, sound cards and gamepad interfaces, etc., is considered to be a relatively advanced market NV1 is one of the most advanced products on the market.

However, it took two years for the NV1 but quickly encountered the embarrassing situation of no one, the reason is very simple, Microsoft released Windows 95 Direct 3D standard in 1996, which is based on triangular texture mapping, NVIDIA's NV1 but can not be compatible with this soon-to-be-mainstream technology, which is equivalent to the automatic abandonment of the market for Windows computers, and in order to pursue profits, the NV1 The NV1 was priced at a premium and sales were way down.

The only good news at this point is that during the development of the NV1, NVIDIA's integrated solution successfully attracted Sega, a game console maker with high cost requirements, and also secured an order for the graphics processing portion of the Saturn's next-generation console - the Dreamcast - and the brash Sega paid $7 million once the Research and development funds.

But NVIDIA's CTO, Curtis Priem, had an unimaginable obsession with the doomed quad texture mapping, which was adopted not only for the NV1, but also for the NV2, which was being developed for Sega, in total disregard of the upcoming Direct 3D standard, which ultimately led to the parting of the ways between the two companies.

Huang Jen-Hsun frankly said in his speech, NVIDIA's architectural strategy at the time and 3D graphics technology trends have been the opposite of a big mistake is not too much, in addition to not being compatible with Windows, the actual game performance is also very general, if you want to complete the contract, you have to research and development of a graphic processor from scratch, and do not fulfill the contract, that $ 7 million in funds instantly become the debt of this small 30-person company! At this point, NVIDIA is only one step away from bankruptcy.

He chose to tell the truth, telling Sega's CEO, Akichiro Irumi, that NVIDIA's technology development was going in the wrong direction, and that Sega should hurry up and change partners, but that the company desperately needed the money, and that without Sega's investment, NVIDIA would soon go out of business.

Fortunately, probably for the reason that Japanese companies value their reputation, Akichiro Irisu actually agreed to his request and did not cancel the contract between the two companies, but was ready to use the NV2 for the next generation of children's toys on top of the Sega Pico, and NVIDIA temporarily got out of the dilemma of closure and gained six months of valuable time.

In order to save NVIDIA in the crisis, Jen-Hsun Huang also carried out some drastic reforms to the company, replacing the CTO and technical director David Kirk (David Kirk) from the host game manufacturer Crystal Dynamics, who will lead NVIDIA's technical development team to face the mainstream standards, and to attack high-end graphics processors, targeting 3Dfx, which has already revolutionized 3D technology at this time.

In order to ensure production capacity, Huang Jen-Hsun also on the eve of the release of the new product, flew to Taiwan, where the manufacturing industry is well developed, and used his Chinese identity and relations to find the head of TSMC, Zhang Zhongmou, and graphics card motherboards and other manufacturers to reach a cooperation, for the Nvidia on the last line of insurance.

In April 1997, Nvidia's third-generation product, the NV3 Riva 128, which stood for "Real-time Interactive Video and Animation," was officially launched. Although it was still difficult to match the image quality of the then-hot 3dfx Voodoo, the Riva 128 quickly won the favor of consumers and OEMs by virtue of its 100M/s pixel fill rate and Open GL compatibility, and in less than a year, Riva 128 shipments exceeded one million units.

NVIDIA finally survived, and Jen-Hsun Huang that "we are only 30 days away from closure" is not an exaggeration, the early days of NVIDIA, has been struggling in the red line of closure, the key decision-making mistakes, almost buried the young company.

3,surrounded by three sides (idiom); surrounded and annihilated by three parties

After 1997, NVIDIA opened the mode of killing, has launched the Riva 128 ZX, Riva TNT, Riva TNT2, a set of strengthened combinations of all the display chip factory on the market to find the north, even the market leader 3dfx are losing ground, and in 1999 launched the GeForce 256, is known as the world's first GPU, which is the first 256-bit display chip, but also integrated 23 million transistors per second. Not only is the first 256-bit display chip, but also integrated 23 million transistors, at least 10 million polygons per second, creating an instant graphics display standard for the personal computer industry.

On August 31, 1999, the GeForce 256 was officially launched, and several giants of personal computers, including Dell, Gateway, Compaq and NEC, announced that they would pre-install NVIDIA GPUs, with IBM joining on the following day, and the six major graphics card producers also fell back on the NVIDIA camp, making it clear that they would be joining the NVIDIA camp.

In 1999, NVIDIA successfully listed on the NASDAQ at an IPO price of $12 per share, and then the stock price quickly soared to $70 per share, sales reached $4,40% in 1997, and the number of employees increased to 270, which made NVIDIA the Silicon Valley's newcomer that was reported by the newspapers in rounds.

Jen-Hsun Huang, who has been making a name for himself in display chips, also put forward a "Huang's Law", which states that "NVIDIA's core strategy is that products will be upgraded every six months, doubling their functionality." This is three times faster than Moore's Law, which was introduced in 1965.

Jen-Hsun Huang and NVIDIA's vision is not limited to the PC graphics card this corner, the previous cooperation with Sega, although ultimately aborted, but to another Sega partner to show their potential, this partner is no one else, is precisely for the Sega Dreamcast to provide Windows CE system Microsoft.

In addition, NVIDIA's GPUs began to sell well, it also participated in the development of the DirectX standard, although NV1 and NV2 did not use Microsoft's standard, but after NV3, it can be said that NVIDIA for the standard to stand up for the work, if not for it, it may be Microsoft will need to spend more time to popularize this standard.

And with these two main factors, NVIDIA managed to get the biggest order since its founding - the display chip for Microsoft's new game console, the Xbox.In March 2000, Microsoft announced that it was about to enter the game console market, which would be in direct competition with Nintendo and Sony, and that the initial console, the Xbox, would be equipped with a GPU from NVIDIA. NVIDIA's GPU.

The news was announced, NVIDIA's stock price on a roller coaster, to know that the game host is different from the PC market, a best-selling game hosts, such as Nintendo's N64 and Sony's PlayStation, is often tens of millions of sales, and a separate certain display chip, hold up to the death of the two or three million sales, and Microsoft as the world's largest technology company, to do hosts are not The company is the largest technology company in the world, so it's easy to make a console. People are talking, NVIDIA is really going to be developed.

Magnificent Microsoft, in order to let Nvidia go all out to develop the Xbox GPU, directly paid a deposit of $200 million, the total contract amount between the two reached $500 million, while Nvidia's annual sales of only $500 million, followed by Microsoft and the Xbox's media delivery processor (MCP) to Nvidia's R & D, is considered to be half of the little brother.

On November 15, 2001, Microsoft Xbox was the first to go on sale in the U.S. Its CPU is Intel's Pentium 3, and its GPU, NVIDIA's customized X-Chip, and even the system is a streamlined version of Windows 2000, so it can be said that the Xbox is a not-so-low-configuration x86 computer.

And in order to quickly break into the console market, Microsoft eventually set the North American price of the Xbox at $299. What is the concept of this price? NVIDIA's card king GeForce 3, launched in February 2001, starts at $329, and the customized X-Chip is equivalent to an improved GeForce 3, which means that consumers can use the price of a graphics card to buy a well-equipped, high-performance PC that even supports the recently launched Dolby 5.1.

Some analysts have estimated that Microsoft in the early days of the Xbox, each sold one, material costs may have to lose $ 125, such a play, even if the family is big Microsoft, burned a period of time after the money can not eat, and the Xbox host parts in the highest cost, is the GPU and MCP, its price even more than Intel's Pentium 3, want to save costs, can not bypass the Intel NVIDIA.

In April 2002, Microsoft formally proposed to NVIDIA, hoping to lower the supply price of GPUs, the results were rejected by the old yellow mouth, the two families in the price of the conflict quickly erupted, and eventually resorted to arbitration. The Xbox chip accounted for NVIDIA business 15% ~ 20%, the two broke up after the news came out, NVIDIA shares soon ushered in an avalanche.

Microsoft is that NVIDIA rely on their own DirectX standard to make a lot of money, this time also paid $200 million in advance, now the game machine business to carry out the difficult, you a little brother should have the good sense to slightly lower the price, and NVIDIA but feel that their own production of chips for the Xbox has led to the declining profit margins, and also the pressure on the price of that is the deception of the people too Sheng, direct to you to engage in the cut-off believe it or not?

The tension between the two companies continued into 2003, when not only did Microsoft stop treating NVIDIA as a close partner, killing the idea of letting it design the next-generation console Xbox 360 again, but even Intel noticed the fast-growing display chip maker and began to develop its own graphics processors, while the two giants began to vigorously support NVIDIA's biggest rival at the time -- fellow graphics card maker ATI -- in a siege against the three manufacturers. --At the same time, the two giants began to support NVIDIA's biggest rival at the time, ATI, which was also a graphics card maker, and these three makers launched a siege against NVIDIA.

For NVIDIA in terms of really can be described as a leaky house, the stock price from the beginning of 2002, $ 70 slipped to $ 7, the market value from $ 11 billion to $ 1 billion, and for Huang Jen-Hsun is even more painful, the year before he was just elected to the U.S. "Fortune" magazine's "40 under the age of 40 billionaires" to $ 507 million in the 12th place, after a year of his personal wealth on the The year after, his personal wealth shrank 10 times, and he was still a billionaire, but after some turbulence, he turned back into a multimillionaire.

In an interview with Wired magazine at the end of the year, he conceded, once again, that the"NVIDIA made a serious mistake, but if there's one thing all great companies have in common, it's the ability to realize they've made a mistake."

4,lit. lie on firewood and taste gall (idiom); fig. suffering patiently, but firmly resolved on revenge

In the face of encirclement, NVIDIA can only helplessly to Microsoft bow, it in February 2003 and Microsoft officially reached a settlement, and subsequently reached a cooperation, the use of Microsoft MCE media player software bundled with the sale of NVIDIA graphics chips, 2003 Xbox began to move towards the hot sales, GPU production costs are also gradually declining, and finally let NVIDIA's days a little better.

In order to recover the loss of Xbox 360 orders, NVIDIA also reached a cooperation with Sony at the end of 2002, for the next generation of game consoles PlayStation3 to provide display chips, as a counterattack against the old rival ATI to fight for the two new generation of game consoles Microsoft and Nintendo orders, but also a reprieve.

At the same time, NVIDIA released its own sincerity to Intel, at the end of 2003, there is news that NVIDIA is preparing to make the Nforce chipset quasi-compatible with Intel's Ben 4 Prescott, and in the January 2004 CES technology exhibition, NVIDIA is the first time to show the use of PCIE display chipset, and subsequently fully support the Intel PCIE standard, in November of that year, NVIdIA In November of that year, NVIDIA and Intel formally reached a mutual authorization agreement, the exchange of patents and chip authorization, the two tit-for-tat companies were able to reconcile.

Of course, NVIDIA knows that iron still needs its own hard this truth, starting in 2004, Jen-Hsun Huang summed up the lessons learned in product development, in order to break through faster in technology, NVIDIA will be each new product design team adjusted to thirty or forty people, and from different disciplines, and even intentionally arranged for engineers from different companies to match the group, in order to collide with the inspiration.

In April 2004, at the biannual product update conference, NVIDIA officially launched the GeForce 6800 series, with 220 million crystals, more than twice the performance of the previous generation, easily overpowering ATI's 9800 series, and quickly gaining the upper hand in the consumer electronics market.

And at this time its biggest rival ATI but broke the news of the imminent acquisition by AMD, heard the news Huang Jen-Hsun immediately said that this is good news for NVIDIA, it is simply pie in the sky.

"ATI This is a plan to admit defeat, let's be the only independent graphics chip company in the world!" He said.

As for the real idea of his heart, may be more of a mixed bag of flavors, AMD left the staff broke the news that AMD was not initially did not consider the market advantage more obvious Nvidia, but Huang Jen-Hsun insisted on his own as the CEO of the merged company, and ultimately the AMD backed off and chose Nvidia's rivals ATI.

Of course, his optimism also came a little too early, after AMD integrated ATI, launched several counterattacks, and even briefly surpassed Nvidia to become the market leader in the period from 2010 to 2012, this GPU battle can be difficult to distinguish.

The opportunity to change everything came in 2007, the year NVIDIA hosted the first CUDA (Compute Unified Device Architecture) technology conference in Santa Clara, California, USA, a parallel computing platform and programming model that enables developers to use programming languages such as C, C ++, Fortran, etc. to write GPU-accelerated applications.

CUDA initial development is not smooth, after all, the establishment of a model, you need a huge base of developers, in order to solve this first chicken or first egg problem, NVIDIA ultimately chose to use the game GPU, that is, GeForce to establish the hardware foundation, the use of a sufficient number of gamers market to promote the development of CUDA.

But CUDA also need more cost support, and even caused the third crisis in the development process of NVIDIA, after looking through the NVIDIA financial report, we can find that after 2008, NVIDIA's net profit and gross profit are significantly lower, and even a loss, and at that time, NVIDIA's market value is only slightly more than 1 billion U.S. dollars. Even shareholders are skeptical of the correctness of CUDA.

Good thing that NVIDIA's game GPU business at this time the development of the game can be quite, equipped with NVIDIA GPU PS3 in the price reduction also began to go to the hot sales, the two for NVIDIA R & D CUDA provides a valuable blood transfusion channel, so that it survived this difficult period, and at the same time began to run the GTC conference, to the world to promote CUDA this technology.

The real turnaround appeared in 2012, AI researchers discovered CUDA, they found that Tesla, a massively parallel computing chip based on the CUDA architecture, has a natural fit in artificial intelligence, deep neural network technology, greatly reducing the difficulty of training neural networks and other high-computing models, at this time to provide arithmetic infrastructure of Nvidia in the years of suffering, and finally ushered in the victory! The dawn of victory.

Today, 10 years later, NVIDIA has been equated with AI, and wherever AI development is talked about, it is destined to bypass NVIDIA, the chip company, which is already the biggest power engine for AI development.

Of course, NVIDIA is betting on more than just CUDA.

Before 2012, NVIDIA, seeking new growth points, targeting the emerging smartphone market, launched the Tegra series of mobile chips, the first product is the Zune HD media player launched by Microsoft in September 2009, when the market has Qualcomm, MediaTek, Texas Instruments and other rivals, and NVIDIA's idea is very simple: the Arm's public CPU architecture coupled with their own high-performance GPU, as long as people have a demand for mobile games, Tegra can occupy a place in the high-end market.

From the early Tegra APX 2500, to break out of the famous Tegra 3, NVIDIA once in the smartphone market to take a small share, but the problem also arises, NVIDIA does not have its own baseband, need to buy additional from other manufacturers, and self-developed mobile GPU seems to be powerful, but the power consumption of the heat is not small, equipped with the Tegra chip of the terminal product reputation is not really good. Performance is not good.

Within a few years, the Tegra chip's customers had shrunk from dozens to single digits, and by the time of 2015's Tegra X1, no other manufacturer was willing to carry Nvidia's mobile chips anymore, with the exception of Nintendo's Switch console, Google's Pixel C tablet, and Nvidia's Shield TV.

This is another NVIDIA fiasco, the good thing is that the Tegra series of products are not only suitable for mobile products such as cell phones, early also by the Tesla Model S and other cars as the main control chip of the car is equipped, NVIDIA see the mobile market did not break out of the Hall of Fame, simply completely transformed the Tegra for the car and self-driving chip, belongs to a fairly decent strategic shift.

Of course, NVIDIA has never given up on the mobile market this piece of cake desire, otherwise it would not have gone in September 2020 to announce the acquisition of Arm, this year, NVIDIA and MediaTek also reached a cooperation in MediaTek's flagship SoC integration of NVIDIA's GPUs, want to prove once again in the mobile market to prove their strength.

AI, automotive, and mobility may be the new chapter in NVIDIA's story for the next decade.

5,lit. dress embroidered at night (idiom); fig. xenophobia

Looking at dozens of large semiconductor companies in Silicon Valley, as NVIDIA CEO Jen-Hsun Huang, is also one of the most high-profile and pro-people.

He almost did not miss the NVIDIA product conference, always dressed in a black leather jacket, often have amazing remarks released, and dare to say and do is his consistent style, and colleagues bet NVIDIA stock price of more than 100 U.S. dollars, go to tattoo a NVIDIA Logo on the body, he also said that tattooing because of the too much pain, "screaming like a child! ".

While NVIDIA's market value was climbing late last year, Jen-Hsun Huang showed up in Taiwan, wearing a familiar black leather jacket, and entered the street booth of two young singers, with a curious look on his face as he played a Lady Gaga song on demand.

In his speech to NTU graduates, he summarized the lessons learned from three mistakes: facing mistakes and asking for help in the early days of the business, sticking to CUDA's inevitable pain in order to achieve the vision, and abandoning the mobile market by strategically retreating and deciding what not to do.

These three admonitions can be said to be throughout NVIDIA's 30-year history, and thus talk about NVIDIA's success today, it is destined to not be able to get around all the previous failures and setbacks, unlike other semiconductor companies, NVIDIA is like this part of the error as a scar on the body of the man, never ashamed to mention, Huang Jen-Hsun also said that he has experienced several very big failures, those journeys are full of humiliation and embarrassment, but It is these failures that define NVIDIA today.

Without the failures of NV1 and NV2, perhaps NVIDIA wouldn't have learned the importance of following more advanced standards; without the siege of Xbox, Intel and ATI, perhaps NVIDIA would still be sticking its neck out against the world today; and without the strategic decisions in CUDA and the mobile market, perhaps NVIDIA wouldn't have been able to make a dent in the field of autonomous driving and AI... ...

Error is to support the bones of NVIDIA, and the soul of which, naturally, is the CEO Jen-Hsun Huang, who is to NVIDIA, like Noyce to Intel, Saunders to AMD, from the birth of the company to date, he has long been in the company on the body of a deep personal branding.

Jen-Hsun Huang's final message to graduates is to run up to predators, or run hard to avoid becoming food for predators, a sentence that exposes the nature of the semiconductor industry's brutal competition, if you can't be a predator, then you will become food, and he who won in more than 30 graphic chip companies has a deep understanding of this.

And NVIDIA can finally stand on the top of the food chain, in addition to what was said earlier, in response to the error of the rapid adjustment, there is also the meaning of this sentence contains, never stop running, like the movie "Forrest Gump" constantly running Forrest Gump, stopping will be surpassed by the later, only to move forward in order to continue to seize new opportunities.