**: Xingyue said financial report.
Last night was a thrilling night.
The global market is waiting with bated breath for Nvidia's earnings report, and Bloomberg even issued a message on social ** "If Nvidia does not exceed expectations today, the earth will **!" exaggerated remarks to express the anxiousness of waiting. While waiting, before the market on Wednesday when the earnings report was announced, Nvidia's stock price fell 8% at one point, and U.S. stocks even fell first, and the Nasdaq index fell more than 1% at one point.
However, after the U.S. stock market in the early morning of February 22, Beijing time, after Nvidia released its Q4 earnings report, everything reversed.
According to the financial report, Nvidia's Q4 financial report revenue was 22.1 billion US dollars, a year-on-year increase of 265%, gross profit margin (GAAP) was 76%, net profit was 12.3 billion US dollars, a year-on-year increase of 769%, and the annual fiscal year revenue reached 60.9 billion US dollars, an increase of 126%, and the main data reached a record high, exceeding market expectations.
After the release of the dazzling financial report, Nvidia's stock price rose by more than 12% after hours, and its market value soared by $277 billion in one day, setting the highest single-day market value growth and breaking the previous record of single-day market capitalization** set by Meta. In the words of Chinese netizens, Nvidia has risen the total market value of "Ali + Jingdong +" in just one day.
Nvidia won the battle to defend the earth and guarded "the hope of the whole village". So, why is Nvidia "remembered" by so many people? What is the importance of Nvidia today?
The AI that is on fire will send Nvidia to the sky as well?
The explosion of AI large models is an important boost to Nvidia's soaring market capitalization.
In 2022, OpenAI released ChatGPT, which added the first fire to the soaring market value of Nvidia.
Before ChatGPT was born at the end of October 2022, Nvidia's stock price fell to $108, giving it a market capitalization of less than $300 billion. After the chatbot ChatGPT was born and ignited the AI wave, generative AI became the outlet, and NVIDIA's stock price rose 6 times in more than a year. On February 15, 2023, OpenAI released the Wensheng ** large model Sora, and the popularity of AI has risen again, and NVIDIA has once again become the focus of the industry.
Behind the above phenomenon is inseparable from the word "computing power".
The era of artificial intelligence has led to a sharp increase in the demand for computing power.
According to Ping An's research report, large models require large computing power, and the iterative upgrading of large model algorithms will provide a strong impetus for the growth of the global and Chinese AI computing power market. According to the research report of Huatai**, the high computing difficulty and high data requirements of AI generation will support the continued strong demand for computing power, and it is recommended to pay attention to investment opportunities in the computing power sector.
At the same time, some industry insiders said that the inference computing power of the newly launched SORA model is much larger than that of large language models. Some data have shown that the inference computing power consumption of stable diffusion is similar to that of the LLAMA 70B (70 billion) parameter model. In other words, in terms of inference computing power, a 1 billion Wensheng graph model is about the same as a 100 billion large language model. The inference computing power of SORA, the first generative model, is definitely much larger than that of the first generative model.
AI has made computing resources be robbed like crazy, and NVIDIA is the first to benefit by virtue of its chip advantages.
It is understood that previously, NVIDIA's revenue pillar was still the game business. According to the analysis of industry insiders, previously, NVIDIA's revenue came from five major sectors - games, data centers, professional vision, automobiles, and OEMs. The core of these are games and data centers, and AI computing-related revenue is attributed to data centers.
The turning point is in 2022. It is reported that in 2017, Nvidia transformed from traditional computers to artificial intelligence, and the proportion of data center revenue increased year by year. Five years later, it was not until ChatGPT came out that the transformation effect was fully reflected. In 2022, NVIDIA's data center revenue will be $15 billion, accounting for 56%, surpassing the gaming business for the first time.
According to the latest financial report, Nvidia's data center business grew by 409% year-on-year in Q4, which was driven by strong demand for large models, recommendation engines and generative AI.
The reason is that NVIDIA is the provider of underlying computing power, and technology giants and startups need NVIDIA's AI chips to develop large models.
According to industry analyst firm OMEDIA, Nvidia currently accounts for more than 70% of the market share of the AI chip market. Google, Amazon, Microsoft and Meta, these companies are inseparable from Nvidia's ** behind the AI arms race, especially the H100 GPU chip.
At the same time, Musk also said that the competition in the AI field will be a big battle, and billions of dollars will need to be invested to remain competitive. He said Tesla will spend more than $500 million on Nvidia's GPU chips this year, and that billions of dollars may be needed every year in the future to stay competitive.
In an analyst conference following the earnings release, Nvidia presented stronger earnings expectations to investors.
Nvidia expects the data center infrastructure market to double in five years. Gartner's forecast is more optimistic, with the market research firm predicting that the market for AI chips could double to $140 billion within three years.
From the perspective of financial indicators, the company's operating expense ratio continued to decline to 144% low. This is mainly due to the sharp increase in revenue, which has significantly reduced the proportion of expenses. The current proportion of the inventory side is at a historical low, which also shows that the company's current product situation is still in short supply.
However, Nvidia, which is in the limelight, also has its own "troubles".
First, the issue of production capacity. It is reported that NVIDIA's biggest problem at present is that its weak output cannot meet the huge demand. Nvidia CFO Colette Kres said that the current market demand far exceeds Nvidia's best capabilities, especially the new generation of chips B100 that will be launched later this year.
Second, the "siege" crisis. On the one hand, old rivals AMD and Intel are "grabbing food". It is reported that AMD's MI300X series stands out in the AI GPU market, and its AI chip sales can reach $3.5 billion in 2024, and it has always been lower than its competitors
At the same time, Intel's fifth-generation Xeon processors are also accelerating, and the United States intends to provide more than $10 billion in subsidies to Intel. Therefore, some industry insiders said that in the more cost-sensitive market segment, the two companies may have a certain impact on Nvidia's market share and pricing power in high-end GPUs.
On the other hand, the rookies are playing catch-up. For example, the recent hit GroQ. It is reported that GROQ is an American AI chip company that has launched the fastest large-scale model inference chip LPU. Judging from the data, the inference speed of GroQ's self-developed LPU is 10 times that of Nvidia's GPU, and even the cost is only 1 10.
Startups like Groq, known as the "Nvidia Challenger", have made a name for themselves in 12 in the United States alone, and they have only been established for an average of five years.
More importantly, at present, some of NVIDIA's major customers, such as Microsoft, Amazon, Meta and Google, have begun to deploy their own AI chips, hoping to gradually get rid of their dependence on NVIDIA GPUs and form their own AI ecological platforms. Data shows that Nvidia sold 2.5 million chips last year, with Microsoft and Meta being the two largest buyers. In the past two quarters, the two AI giants have accounted for a quarter of Nvidia's AI chip sales.
In short, as global demand for accelerated computing and generative AI surges, the company's share price has more than quadrupled since the beginning of last year, but it has been supported by performance along the way. Even if the current market capitalization reaches 1More than 6 trillion, the company's current PE is still less than 40 times, and it is far lower than Intel's 106. Xingyue said that the financial report believes that sustained high performance growth will be able to support the company's stock price and bring confidence to the market in the short term. But if you want to go to the next level, the hope of the whole village may need to work harder.