Nvidia (NVDA -3.33%) stock has been on a tear since the beginning of 2023 -- it's up 80% year to date. One key driver of that surge is that it's a leader in artificial intelligence (AI) and machine learning, and since OpenAI raised the public's awareness of how advanced AI has become with its release of ChatGPT in November, investors bid up the prices of most top AI stocks.
Investors have become so excited about Nvidia's AI-related prospects that its share price doesn't appear to have been much affected by the recent spate of issues in the banking sector, which included the second-largest bank failure in U.S. history. As a result of those bank failures, many economic experts now see a recession this year as being more likely. So considering that its stock price has run up so fast and that the economy is likely to worsen, should you buy Nvidia today or wait for macro headwinds to drive its price back down?
ChatGPT opens new opportunities
ChatGPT is just one of many large language models (LLMs) in AI. Nvidia defines an LLM as a deep learning algorithm that recognizes text, and can generate text, video, audio, and other content based on knowledge gained from datasets.
When OpenAI made ChatGPT publicly available, it opened people's eyes worldwide about how fast AI is advancing. As a result, it became a no-brainer for most companies to begin developing AI strategies to compete in their markets. However, it's an onerous set of tasks for most companies to build AI supercomputer infrastructure, create machine learning models, process big data, and train an AI. Thus there is a considerable need for an AI factory that organizations can hire to help custom-make LLM and other AI models for specific use cases.
During its fiscal 2023 fourth-quarter earnings call, CEO Jensen Huang said Nvidia is introducing Nvidia AI cloud services, which will provide everything a company needs to use LLM models, from the infrastructure to the operating system to the pre-trained models. It will provide the service through its network of partners and use many of the biggest cloud providers as hosts. Huang believes this new service will accelerate the adoption of LLM AI. Of course, should this new service take off, it would significantly expand Nvidia's data center business.
Shareholders are speculating about AI and gaming growth
On Feb. 22, Nvidia reported rather lackluster results for its fiscal 2023 Q4. Revenue sank 21% year over year to $6.05 billion. Moreover, management forecast fiscal 2024 first-quarter revenue of only $6.5 billion. If it hits that target, it would equal a 22% decline from the prior-year period, and give the company five straight quarters of declining year-over-year revenues.
NVDA Revenue (Quarterly YoY Growth) data by YCharts.
The data center segment produced $3.6 billion in revenue in fiscal Q4, up 11% year over year. In gaming, quarterly revenue was down 46% year over year, and for the full year, gaming revenue was down 27%. Since the data center and gaming segments together provided 90% of its Q4 top line, these weak results are likely not why the stock rose 14% the day after the earnings release.
Why did the market bid the stock price up? Likely because investors are speculating that demand for sophisticated AI products will increase rapidly this year, which will have a knock-on effect of reaccelerating Nvidia's data center revenue growth.
It's in the midst of introducing a set of new data center hardware products, which are arriving just in time to manage that expected increase in AI computational loads and networking needs.
Nvidia's new AI H100 center Tensor Core GPU has gotten off to a strong start two quarters into its ramp-up. The company reported that H100 revenue is already much higher than the revenue for the previous generation GPU, the A100, which declined sequentially in the fourth quarter. According to management, the H100 chip is up to 9 times faster than the A100 for training AI systems and up to 30 times faster in the inferencing of transformer-based large language models. Since the company optimized the H100 for developing, training, and deploying sophisticated AI systems like LLMs, shareholders should monitor future management commentary concerning the chip to get a handle on how much LLMs are impacting the company's revenues.
As for its second-largest segment, gaming, revenue growth was up 16% sequentially. And its inventory correction efforts, which were the cause of its recent revenue declines, are moving into the rearview mirror. If the inventory issues that have been plaguing Nvidia's gaming business are over, that segment's revenue could be on the verge of a turnaround.
If you are ready to speculate that Nvidia's two largest revenue generators have bottomed cyclically and are poised to recover, you might view the stock as a buy. First, however, you should weigh the chipmaker's valuation.
Nvidia's trading at a premium
The market currently values Nvidia at a price/earnings-to-growth (PEG) ratio of 4.6 compared to the general semiconductor industry's PEG ratio of 3. Most growth investors consider a stock trading at a PEG ratio above 2 to be overvalued.
Considering the likelihood that the economy could worsen this year, and the possibility that its revenues have yet to entirely bottom, prospective investors should be wary. If Nvidia misses analysts' revenue or profitability expectations in the next few quarters, the stock's drop could be steep. Despite its high upside potential over the next several years, cautious investors might want to wait for a significant pullback in price before buying shares.
