NVIDIA's market value soared by 500 billion overnight, and Huang Jenxun sat on the AI iron throne
Image source: @VisualChina
Titanium media note: This article comes from the WeChat public account China Entrepreneur Magazine (ID: iceo-com-cn), author | Zhao Dongshan, ed Li Wei, published by Titanium Media with permission.
Huang Jenxun, the 60-year-old godfather of AI who started his business at the age of 30, waited for his highlight this year - with the global explosion of AI big models, Nvidia, founded by Huang Jenxun, became the biggest winner because of the provision of powerful generative AI chips, and Huang Jenxun's net worth also rose.
In the early morning of August 24, Beijing time, NVIDIA announced its latest quarterly financial report as of July 31, 2023, showing that the company's revenue in the quarter reached a record high, reaching US$13.51 billion, a year-on-year increase of 101%; Net profit reached US$6.188 billion, a year-on-year increase of 843%; Its gross margin also reached a staggering 70.1%.
In fact, Nvidia's stock price has more than tripled this year. According to the real-time ranking of the Bloomberg billionaires list, as of July 4, Huang's personal net worth was about $38 billion, ranking 34th in the world, of which $24.2 billion was added this year, making him one of the most wealthy people in technology this year.
In May 2023, NVIDIA gave a revenue forecast of $11 billion, which shocked the entire Wall Street at that time, and its market value exceeded trillion dollars at that time. Huang Jenxun thus became the first Chinese entrepreneur in history with a market capitalization of trillions of dollars, and NVIDIA also became the first trillion-dollar chip company in history.
It turns out that Huang's revenue expectations are by no means big words, and NVIDIA's actual results are even much higher than expected. In addition, NVIDIA also created a milestone moment in the chip industry - NVIDIA's quarterly revenue exceeded Intel's $12.9 billion for the first time, which represented the fabless chip design company defeating the IDM (vertically integrated model) giant.
"A new era of computing has begun. Companies around the world are transitioning from general-purpose computing to accelerated computing and generative AI. Huang said.
Huang Jenxun's confidence lies in the fact that under the global AI big model arms race, the demand for underlying computing power has skyrocketed, and Huang Jenxun and NVIDIA, who control computing power, as computing power providers, have become the biggest winners in this wave of AI large models. And the above-expected performance of the new quarter also made Huang Jenxun ride the dust and sit on the iron throne of AI.
Kill crazy financial reports
Looking at the dismal global capital market, NVIDIA is one of the few to shine.
Among NVIDIA's $13.51 billion revenue in the second quarter, data center and gaming businesses were the most critical sources of revenue, accounting for about 95% of the combined revenue.
Because of the hot global demand for GPU chips used to process AI computing, NVIDIA's data center business revenue doubled, far exceeding the game revenue, and the proportion of revenue rose to 76%, far exceeding the previous quarter, setting a new quarterly record. The "doubling of NVIDIA's revenue" in the quarter was mainly due to strong demand from downstream data center business.
Computing power, algorithms and data constitute the three elements of the AI era. Among them, computing power is power-like infrastructure, how many GPU chips, how much computing power, has become the key to the outside world to evaluate the entry threshold of large model research and development and entrepreneurship. GPU chips have become the core components of ChatGPT training and deployment because of their performance advantages in graphics computing, and NVIDIA has almost no competitors in the field of GPU chips.
All this stems from Huang Jenxun's accurate judgment and past cultivation. As early as the establishment of OpenAI, Huang Jenxun personally handed over the world's first AI supercomputer DGX to OpenAI. He witnessed the launch of ChatGPT and the innovative iteration of GPT from 3.5 to 4.0. Since then, the competition of computing power of global AI companies has also become a competition of how many NVIDIA GPU chips are.
According to industry assessments, a ChatGPT needs 10,000 NVIDIA A100 chips to run. According to the estimation of market research agency Jibang Consulting, most of the more than 1.2 million AI servers shipped this year are equipped with NVIDIA's GPU chips, accounting for 60%~70% of the share.
According to NVIDIA CFO Colette Kress on the earnings call, the data center's revenue of $10.32 billion mainly comes from cloud service providers, as well as Internet technology companies that train and deploy generative AI, and HGX server computing platform with Hopper architecture is driven simultaneously, and data center computing products increased by 195% year-on-year and 157% month-on-month; The Network segment grew 94% year-over-year and 85% sequentially due to demand for InfiniBand network solutions that complement the HGX server computing platform.
In addition, Kress mentioned that as customers shift their investments to AI and accelerated computing, U.S. data center growth is the strongest, while demand from China accounts for 20% to 25% of Nvidia's data center revenue.
When asked by analysts about the U.S. government's rules requiring NVIDIA to restrict exports of high-performance AI chips to China, Kress said that given the strong global demand for NVIDIA products, if the U.S. imposes additional export restrictions on data center GPU chips sold to China, it is not expected to have a direct and substantial impact on NVIDIA's performance, but in the long run, such restrictions, once implemented, will cause the U.S. industry to permanently lose the opportunity to compete and lead in one of the world's largest markets.
In the second quarter, although the game business increased steadily in the quarter, its proportion was still compressed by the high-growth data center to less than 20%. Games revenue of $2.49 billion increased 22% year-over-year and 11% sequentially, 4.6% above analyst expectations of $2.38 billion.
NVIDIA's gaming business is mainly graphics cards and GPU chips in personal desktop computers, laptops and other devices. Demand in the home entertainment and consumer electronics markets has slowed, and destocking has become the main theme of NVIDIA's gaming business. NVIDIA's fiscal second quarter inventory totaled $4.32 billion, with inventory turnover days of 97 days, down from $4.61 billion and 165 days in the first quarter.
In key data center products, NVIDIA began to show a situation of "short supply", which also led to NVIDIA's gross margin reaching a staggering 70.1% in the quarter, and Intel's gross margin was 39.8% in the same quarter. Nvidia's gross margin for the third quarter is expected to return to 71.5%.
Legendary entrepreneurial story
Huang's success did not happen overnight. Behind the highlights, Huang Jenxun has worked silently for 30 years.
Huang was admitted to university at the age of 16, entered Stanford with a master's degree, and founded his own IT company at the age of 30. In the early days of its business, NVIDIA stumbled all the way and almost went bankrupt several times. When Huang was 36, Nvidia went public, but its share price fell to just $6 at one point. However, in the 30 years since its founding, Huang Jenxun has not relaxed for a moment, allowing the company to enter the trillion-dollar market capitalization club.
In Huang's life dictionary, there are only four things that matter: working, thinking, helping others, and creating. Because of the experience of being on the verge of bankruptcy during the entrepreneurial period, Huang Jenxun still often hangs the phrase "Remember, we are only 30 days away from bankruptcy" to stimulate the fighting power of employees.
As early as his college days, Huang Jensen aspired to become a global graphics emperor. At the age of 30, he co-founded NVIDIA with two engineers. Different from the chips made by Intel and AMD, Huang Jenxun paid more attention to the graphics chip that allowed games and images to run smoothly, because at that time, personal computers had just entered the home, but entertainment functions were still missing, unable to meet the needs of game running.
After starting his business, Huang launched the NV1 chip. But by betting on the wrong technology direction, NV1 depleted the company's early investment. To survive, the company had to shrink from more than 100 to more than 30. The emergence of Japanese game giant Sega made Huang see a turning point, and the other party took out $7 million to hope that Nvidia would develop a game console chip, later NV2.
"We are honest with our mistakes and humbly ask for help to save NVIDIA's survival. These traits are the hardest for the smartest and most successful people to develop. Mr. Huang said in a speech at National Taiwan University this year.
It wasn't until 1997, when NVIDIA developed the third-generation product RIVA128, that it really gained a foothold. This is the first high-performance 128-bit graphics chip that supports Microsoft's Direct3D graphics interface, and has received full support from Microsoft. With the development of the video game industry, the market demand for GPU chips is getting higher and higher, which can help the CPU (central processing unit) to share part of the graphics rendering, video encoding and decoding and other tasks, becoming an important part of the computer industry.
At the end of 2000, NVIDIA acquired its old rival 3DFX for $110 million. That same year, Huang challenged chip giant Intel. Huang Renxun benchmarked against Intel founder Gordon Moore's "Moore's Law" and proposed "Huang's Law", that is, NVIDIA's products are upgraded every 6 months and the functions are doubled. Such a technological update is 2 times faster than Moore's Law.
At the end of 2006, Huang realized that high-speed parallel computing could be used for users other than computer graphics, opened up GPUs to software developers, and launched the CUDA platform, allowing developers to use the computing power provided by NVIDIA for purposes other than graphics. Although it had little effect at first, with the advent of the era of artificial intelligence, developers soon realized that GPUs were excellent at supporting the complex calculations of modern AI systems.
But CUDA's costs were very high, and at a time when Nvidia's profits took a huge hit over the years, with a market capitalization of just around $1 billion. Years of sluggish performance have left shareholders skeptical of CUDA and looking to focus on improving profitability. But Nvidia persevered.
Later, NVIDIA GPU and CUDA programming languages have become the basis and standard for AI development and training. This surprised Huang. In a 2016 interview with Forbes, he mentioned that he had expected GPUs to be used outside of gaming, but never thought of turning to deep learning applications.
Reflecting on this episode, Huang said: "We invented CUDA for AI, and this journey forged our character, and to endure pain and suffering was a necessary pain on the way to pursue our vision. ”
Now, Huang Jenxun has waited for his own time. With his choice and persistence, he explained to the outside world how innovation and insight led him to create a new trillion-dollar market.