Nvidia (NVDA 4.35%) stock has rewarded investors handsomely in 2023 thanks to the hype around artificial intelligence (AI) and the role that the chipmaker is playing in the proliferation of this hot tech trend, as evident from the 80%-plus gains recorded by the semiconductor giant so far this year.
Nvidia's hot rally has made the stock prohibitively expensive. It is now trading at a whopping 154 times trailing earnings and nearly 25 times sales. However, investors who have been holding this high-flying chipmaker in their portfolios should consider holding on to it as it could deliver more upside.
Let's look at two reasons why Nvidia stock could still head higher.
Its Grace chips could give the data center business a massive boost
Nvidia announced its entry into the server CPU (central processing unit) market a couple of years ago, aiming at a market that has been dominated by Intel and AMD. Those chips are all set to hit the market this year, and leaks suggest that they could give established rivals a run for their money.
Nvidia's Grace server processors are based on the Arm architecture as compared to the x86 architecture of the AMD and Intel chips. Last week, Nvidia claimed that the Grace CPU Superchip -- which is specifically designed for accelerating data center workloads such as AI, high-performance computing (HPC), digital twins, cloud applications, and data analytics -- can deliver "2x performance gains over x86 processors at the same power envelope across major data center CPU applications."
Of course, investors should take this claim with a pinch of salt as it is based on Nvidia's testing and not on independent third-party reviews. But if there is a degree of truth to these claims, then it won't be surprising to see data center operators lining up to buy Nvidia's chips. That's because the Grace CPU Superchip could allow data centers to handle "twice as much peak traffic" and reduce power consumption in half, based on Nvidia's claims.
The chipmaker says the Grace CPU Superchip will allow data centers to enhance computing capacity while keeping power consumption under check, something that has become necessary amid the threat of global warming. The good news for Nvidia is that the Grace chips have already started gaining acceptance thanks to the advantages they could bring over x86 processors from AMD and Intel.
The Los Alamos National Laboratory will deploy Grace chips in the Venado AI supercomputer for renewable energy and materials science workloads, while European and Asian data centers are also considering using Grace for their workloads. Nvidia also points out that the Grace chips are now sampling with multiple OEMs (original equipment manufacturers) such as Asus, QCT, Atos, Gigabyte, Hewlett Packard Enterprise, Wistron, Supermicro, and others, and they are set to go into production in the second half of the year.
If Nvidia indeed delivers the gains that it is promising over AMD and Intel's chips, its data center business could get a nice shot in the arm given the huge opportunity available in this space. The Arm server CPU market was worth just $12 billion last year, but it is expected to generate $82 billion in revenue by 2030. That won't be surprising as the adoption of Arm chips in servers is set to grow rapidly thanks to the mix of power and efficiency they offer.
As such, the rapid growth of Nvidia's data center business -- which is its largest in terms of revenue -- could send this tech stock higher.
Generative AI inferencing could be another nice catalyst
Generative AI has hogged the limelight of late thanks to the booming popularity of ChatGPT, a chatbot that can write poems, essays, and code based on user prompts. However, generative AI is expected to do a lot more and disrupt industries such as advertising, entertainment, art and design, healthcare, and manufacturing, among others.
Not surprisingly, the generative AI market is expected to clock annual growth of 36% over the next decade. Nvidia has already taken steps to benefit from this fast-growing niche. The chipmaker recently launched four inference platforms for generative AI applications aimed at accelerating different services such as AI-enabled video, image generation, deployment of large language AI models, and recommendation models.
Each of these platforms could help Nvidia tap lucrative generative AI niches. For instance, Nvidia claims that the large language model (LLM) deployment platform is ideal for "deploying massive LLMs like ChatGPT at scale." It is worth noting that Nvidia GPUs (graphics processing units) have played a critical role in the deployment of ChatGPT. The company is now using its expertise to deliver up to 12 times faster inference performance with the GPT-3 LLM as compared to the prior hardware platform.
All this indicates that Nvidia is in a nice position to make the most of the AI chip market that's expected to grow at almost 30% annually through 2032 to $227 billion at the end of the forecast period.
Ultimately, the Grace CPUs and the AI chip platforms would be tailwinds for Nvidia's data center business that generated $15 billion in revenue in fiscal 2023, a 41% jump over the prior year. The huge revenue opportunity in server and AI chips that we saw above indicates that Nvidia's data center business could be at the beginning of a massive growth curve and help this tech stock sustain its hot rally in the market.
