NVIDIA, which took the AI express train, once again handed over a dazzling "answer sheet".
According to the financial report, in fiscal year 2024, Nvidia's revenue will be $60.9 billion, a year-on-year increase of 126%, and its net profit will be 297US$600 million, a year-on-year increase of 581%. Among them, Q4 achieved revenue of 22.1 billion US dollars, a year-on-year increase of 265%; Net profit 122US$8.5 billion, a year-on-year increase of 769%.
Source: Nvidia.
Due to the outstanding performance, after the U.S. stock market on the day of the earnings report, Nvidia's stock price **908%, and drove the entire line of U.S. chip stocks**. As Michael Kramer, an independent market analyst, put it, "If Nvidia's stock price isn't substantial, there's no other factor that can push U.S. stocks higher."
However, it is worth noting that Nvidia's stock price has fallen for four consecutive days, from 739$75 shares** to 674$72 shares, a composite decline of as much as 879%。From this point of view, the good news brought by the brilliant performance only filled the stock price ** a few days ago.
Source:.
Admittedly, it is impossible for any company's stock price to remain elevated, and it is normal for Nvidia's stock price to fluctuate. However, it should be noted that in the past year, with the help of the AI concept, Nvidia's stock price has continued to "soar", with a cumulative increase of up to 40%. Almost every time the results are disclosed, Nvidia's stock price is advanced**.
At the beginning of 2024, as the stock price gradually rises, Nvidia's price-to-earnings ratio is also inflated, and many investors will choose to sell Nvidia** before the earnings report to lock in profits.
It's hard not to worry - in the future, will Nvidia's stock price soar or return to rationality?
H100 is in demand, and the data center business is still the key to revenue generation
At one time, Nvidia's identity was still a company mainly engaged in the game graphics card business. Taking the 2022 fiscal year financial report as an example, the sales of NVIDIA's gaming division reached $12.5 billion, accounting for 46% of the company's total revenue, which can be called "half of the country".
However, with the advent of ChatGPT, Nvidia's business model began to undergo a major shift. It is reported that as early as 2016, Nvidia delivered the world's first AI supercomputer to OpenAI, which eventually contributed to the birth of ChatGPT. After the advent of ChatGPT, many technology companies not only laid out AI models, but also spent huge sums of money to purchase AI chips. Nvidia's highlight moment also came.
It is understood that NVIDIA's market share in the global artificial intelligence chip market is expected to reach up to 90%, which can be said to be the "undisputed leader" in the field of artificial intelligence, which is also due to the wave of large models, its H100 GPU and other products are very rigid attributes, realizing a large-scale seizure of the market.
According to a report disclosed by Omdia, in 2023, Meta and Microsoft bought 150,000 H100 GPUs from Nvidia, and Google, Amazon, Oracle, and Tencent all bought 50,000 H100 GPUs from Nvidia.
Source: Omdia
With the sales of H100 GPUs rising, it is inevitable that NVIDIA's data center business will enter a period of explosive development. According to the financial report, in fiscal year 2024, NVIDIA's data center business revenue will be $47.5 billion, a year-on-year increase of 217%, accounting for 78% of total revenue, surpassing games and becoming the company's most important revenue pillar.
What strengthens the confidence of the capital market is that as more and more technology companies rush to land on the AI track, and the speed of technology iteration continues to accelerate, NVIDIA's data center business has the possibility of further growth.
For example, recently, OpenAI released a new generation model SORA, which will raise the effect of generative AI to a new level, which undoubtedly requires a larger scale of data training, which will also drive the growth of demand for AI chips.
Under such a trend, Guo Junli, research director of IDC Asia Pacific, has ** The scale of AI chips will exceed 65 billion US dollars in 2024. Combined with the market position, Nvidia will be a full beneficiary of the explosion of demand.
It is reported that in September 2023, Nvidia H100 GPUs will be sold out. At that time, I paid for the order, and I won't get the goods until at least mid-2024. In this regard, Nvidia CEO Jensen Huang also explained at the 2024 fiscal year financial report meeting, "The overall situation of AI chips is improving, but it is still facing a shortage." **Restrictions will continue throughout the year."
In addition, the financial report also shows that in Q4 of fiscal year 2024, NVIDIA's inventory-to-income ratio will be 24%, a year-on-year decrease of 61 percentage points; The ratio of accounts receivable to income was 45%, down 18 percentage points year-on-year. The decline in the proportion of inventory indicates that NVIDIA's products are in short supply, and the decline in the proportion of accounts receivable reflects that Nvidia is in a strong position in the ** chain.
Source: Dolphin Investment Research.
It's clear that AI pioneers such as large language model startups, consumer internet companies, and global cloud service providers continue to increase their AI business, accumulating strong growth momentum for NVIDIA. The recent emergence of SORA shows that AI development is still on an accelerating track, and the next wave is taking shape. These trends all imply that Nvidia still has a very high imagination.
However, it is worth noting that AI is an emerging industry after all, and while it is developing rapidly, it is also showing more and more uncertainties, including the layout of self-developed chips by downstream enterprises.
In this case, will Nvidia be able to achieve a more certain performance climb?
How does Nvidia respond to the layout of self-developed chips by technology companies?
Looking back at Nvidia's stock price in the past year, it is not difficult to find that its stock price trend has not been consistently high, but has been parabolic. Obviously, the scarcity of AI chips is the reason for investors to give Nvidia a high valuation, but in the process, many people are deeply worried about Nvidia's growth potential in the short to medium term.
For example, on February 18, 2024, Cathy Wood, the manager of Wall Street star ** and CEO of Ark Investment Management, announced that she had recently sold Nvidia** worth about $4.5 million.
Regarding why she sold Nvidia** before the release of the financial report, Sister Mu said that Nvidia is a leader in the field of artificial intelligence, but it will face more competition in the future, and the demand for chips may not be as hot as the market expects.
According to data compiled by S&P Global Market Intelligence, analysts generally believe that Nvidia's revenue in fiscal 2025 will be about $90 billion, with a growth rate of only 52%, which may not be enough to continue to help Nvidia maintain a high valuation.
In fact, judging from the trend of the AI industry, Nvidia does face a lot of challenges. As mentioned earlier, H100 GPUs are in short supply, and in order to grasp the initiative, many technology companies have begun to develop their own AI chips.
For example, at the beginning of 2024, there was news that OpenAI CEO Sam Altman plans to raise $5-7 trillion to build a huge network of chip manufacturing plants.
Coincidentally, in November 2023, Microsoft also launched two high-end custom chips for Azure, Maia 100 and Cobalt 100. Microsoft Vice President Rani Borkar revealed that Maia 100 has been tested on Bing and Office AI products, and OpenAI is also trialing.
Source: Microsoft.
In the face of the above challenges, NVIDIA chose to accelerate the increase of production capacity on the one hand, alleviate customer anxiety, and meet the demand to the greatest extent.
Recently, a report disclosed by UBS analysts showed that the delivery speed of Nvidia's AI chips is accelerating, with the delivery time as high as 8-11 months at the end of 2023, which has now been reduced to 3-4 months. Based on this, UBS raised the price target on Nvidia's stock price to $850 shares.
On the other hand, increase the customized chip business in order to "draw wages from the bottom of the kettle". According to the research company 650 Group, it is expected that the market size of data center custom chips will grow to $10 billion in 2024 and double in 2025.
Against this backdrop, Nvidia is building new business units focused on designing custom chips, including advanced AI processors, for cloud computing companies and other companies.
It is worth noting that NVIDIA's entry into the data center chip customization market is not only in line with the market trend, but also likely to make some technology companies, based on rational economy, give up self-developed chips.
In response to Altman Sam's "7 trillion" plan, Huang believes that "this money can buy all the GPUs in the world", which is not in line with the logic of industrial development and the economic benefits of enterprises.
In fact, from NVIDIA's point of view, the next focus of the semiconductor industry is not the number of chips, but the innovation of GPU architecture to increase computing speed, so that companies that want to use AI locally do not need to buy more equipment.
It is understood that in November 2023, Nvidia launched a new generation of AI chip H200, which can deliver a new generation of AI chips at 4With 8TB of speed and 141GB of memory, it delivers a 60% to 90% improvement in performance over the H100 when used for reasoning or generating answers to questions, and is expected to be available in Q2 2024.
Source: Nvidia.
In addition, NVIDIA will hold the GTC 2024 conference in late March 2024, when it will release the next generation of AI chip B100, which will improve performance by more than 100% compared to H200.
Considering that it takes huge financial, material and time costs to build self-developed chips from scratch, with the gradual release of production capacity and the rapid increase of computing power, it may still be the main path for technology companies to directly purchase Nvidia's products.
Therefore, in a considerable period of time, Nvidia is still expected to enjoy AI development dividends.
Author: wordless.
*: U.S. Stock Research Agency.