By Joseph Moss, International Banker
On May 30, Nvidia joined an elite group of firms when its market capitalisation touched $1 trillion. This achievement is a testament to the incredible year enjoyed by the tech behemoth thus far, one that has consistently crushed analysts’ expectations, thanks largely to the advances made within the booming GenAI (generative artificial intelligence) sector. And having strengthened further since then—its market capitalisation now sits at around $1.12 trillion—it seems the Nvidia rally is far from over. But with a modest sell-off in September following news that ChatGPT’s creator, OpenAI, may look elsewhere for its AI-chip supply, has Nvidia’s stock price already peaked?
Nvidia—the world’s largest maker of chips used in computers, smartphones and many other applications, the inventor of the graphics processing unit (GPU), a key global producer of the application-specific integrated circuits (ASICs) vital in powering the computing behind AI applications and a leading name in data centres and gaming—enjoyed stock-price gains of almost 220 percent this year through October 6, only further furnishing its already glowing reputation as one of the best US stock-market performers over the last 10 years. Such buoyant gains also mark an astonishing turnaround from last year, when the Santa Clara, California-based company lost more than 50 percent of its value amid a broad sell-off across the technology sector.
Indeed, Nvidia has been front and centre of this year’s booming tech theme on Wall Street—generative AI—especially since last November’s launch of ChatGPT. May 25, for instance, saw Nvidia’s stock soar by a mighty 25 percent after the firm delivered its strong first-quarter results, along with guidance for the second quarter that smashed all analysts’ expectations. Specifically, Nvidia showcased its AI-powered GPU chips as being primed to play a dominant role in the GenAI investing frenzy that large language models (LLM), such as ChatGPT, BERT (Bidirectional Encoder Representations from Transformers) and Cohere, have ushered in. “While the natural reaction is to question the sustainability of this strength in [the] Data Center, we believe the proliferation of Generative AI across various verticals coupled with a fully committed (and currently under-utilized) supply chain will support sequential growth in the October and January quarters,” Goldman Sachs analyst Toshiya Hari noted.
It should come as no surprise, then, that Nvidia’s earnings have comfortably beaten expectations throughout the year. Its first-quarter revenues came in at $7.19 billion, 19 percent higher than the previous period. At the time, the company provided a forecast of its second-quarter (Q2) revenue hitting $11 billion. But when Nvidia announced its Q2 results on August 23, the actual figure for the April-June period was posted at a whopping $13.5 billion. This was an 88-percent surge from the previous quarter and a staggering 101 percent higher than a year earlier.
“The world has something along the lines of about $1 trillion worth of data centres installed, in the cloud, in enterprise and otherwise. And that $1 trillion of data centres is in the process of transitioning into accelerated computing and generative AI. We’re seeing two simultaneous platform shifts at the same time,” Jensen Huang, co-founder and chief executive officer of Nvidia, noted during the company’s post-earnings conference call. He subsequently stated that the high global demand for Nvidia products could be explained by the world “transitioning from general-purpose computing to accelerated computing, [and] the best way for companies to increase their throughput, improve their energy efficiency, improve their cost efficiency, is to divert their capital budget to accelerated computing and generative AI”.
With Nvidia estimated to control more than 80 percent of the global market for chips most suitable to run such AI applications, it is easy to see why investors have been so bullish on the tech giant this year. Released last year, Nvidia’s H100 chip has been particularly singled out for praise and is widely regarded as the best chip available to power GenAI and LLM applications. “During the quarter, major cloud service providers announced massive NVIDIA H100 AI infrastructures,” Huang stated in the Q2-earnings press release. “Leading enterprise IT system[s] and software providers announced partnerships to bring NVIDIA AI to every industry. The race is on to adopt generative AI.”
According to Ted Mortonson, tech strategist at US financial-services firm Baird, speaking with Business Insider in May, the H100 has enabled “a leapfrog in training, inference, basically generative AI” and was crucial in enabling ChatGPT to make its debut last November. “They have the entire AI silicon stack. And those are basically three components. They have the most advanced GPU, they have advanced networking embedded in the silicon, advanced memory embedded, and they’re now developing a new CPU.”
And by controlling the entire ecosystem on both the hardware and software sides, Mortonson also observed that Nvidia has effectively become a one-stop shop for companies looking to advance their AI ambitions, akin to how Apple controls both the iPhone and the iOS mobile operating system. “When you cobble all these things together, it is an integrated, immensely powerful AI engine. And they are years ahead of anyone else,” he noted, whilst highlighting that the development of CUDA (Compute Unified Device Architecture), Nvidia’s parallel computing platform and programming model for general computing on GPUs, is well ahead of the competition.
Moreover, Nvidia is set to launch several hugely anticipated next-generation AI superchips next year, including the GH200, the first 100-terabyte GPU memory system, and the Blackwell B100 GPU. “Per our estimations, B100 GPU could be an even bigger AI game changer than H100 at its launch from a technology standpoint and set for a rapid adoption which subsequently will drive up Nvidia’s [average selling price], sales, and margins,” Citi stock analyst Atif Malik wrote in a recent note.
But it has not always been sunny for Nvidia’s stock performance this year. Indeed, the share price experienced a moderate sell-off during the first three weeks of September, losing around 11 percent, partly on concerns that the Federal Reserve (the Fed) will keep interest rates at higher levels longer to crush inflation. What’s more, concerns over Nvidia’s long-term prospects still linger after an October 5 report from Reuters disclosed that OpenAI may exclude the chipmaker as a supplier, instead manufacturing its custom AI chips, which some believe would be a significant blow to Nvidia’s AI-chipmaking aspirations. The report also stated, however, that it remains unclear whether this plan is definite. “Doing so would be a major strategic initiative and a heavy investment that could amount to hundreds of millions of dollars a year in costs, according to industry veterans. Even if OpenAI committed resources to the task, it would not guarantee success.”
Despite such concerns, analysts remain largely unperturbed and committed to their bullish projections for Nvidia. “We continue to expect Nvidia to maintain ~90% share in the AI GPU market for the next 2-3 years,” Citi’s Atif Malik wrote in a note on October 6. Malik added that if the Reuters report is true, it would be unsurprising and consistent with the view that more GPUs and ASICs will be needed to ramp up the necessary AI infrastructure. “We foresee ASICs being primarily used for smaller and more specialized models and GPUs for both training and inference of larger/more complex models,” Citi noted.
Citi has raised its target price for Nvidia from $420 to $520 per share whilst maintaining a “Buy” rating. According to calculations by the Nasdaq Stock Market, moreover, Nvidia remains a “Strong Buy” based on 37 analysts’ recommendations offering 12-month price targets, with the average price target at $647.04, a high estimate of $1,100 and a low estimate of $560. A recent CNBC report found that Wall Street remains strongly bullish on Nvidia, with about 85 percent of analysts holding a buy rating on the stock and the average price target implying a 43-percent upside from its August 6 close of $457.62 per share.
Goldman Sachs also added Nvidia to its “conviction list” of companies on October 2, chosen from its shortlist of 20 to 25 buy-rated stocks across its Americas coverage. The company is likely to maintain its status as the accelerated computing-industry standard for the foreseeable future, Goldman Sachs stock analyst Toshiya Hari suggested, adding that this opinion is based on Nvidia’s “competitive moat and the urgency with which customers are developing and deploying increasingly complex AI models”.
Hari additionally noted that demand for the company’s GPUs in data centres should remain high for some time and that Nvidia’s supply of chips from contract manufacturers is steadily improving. “A strong and broadening demand profile in the data centre, plus an improving supply backdrop, should support sustained revenue growth through [the] calendar year 2024.” Goldman has placed a price target of $605 for Nvidia, which, if achieved, would represent an upside of around one-third from current levels.