After the market close on Monday, Nvidia (NVDA 6.18%) kicked off its annual GPU Technology Conference (GTC), which some have dubbed the "Woodstock of Artificial Intelligence (AI)." Thousands of tech glitterati showed up in person to watch CEO Jensen Huang's keynote address, while millions more tuned in online.
During the two-hour presentation, Huang expounded on the company's latest creations, collaborations, and partnerships, which will have a ripple effect across much of the tech world. This morning, investors are still digesting the impact of these announcements, which is causing something of a hangover in AI stocks -- likely following the old adage "buy the rumor, sell the news."
With that as a backdrop, AI-centric server maker Super Micro Computer (SMCI 8.89%) tumbled 10.3%, chipmaker Advanced Micro Devices (AMD 2.37%) slumped 5%, and chip designer Arm Holdings (ARM 4.11%) declined 2.8% as of 1:26 p.m. ET on Tuesday.
A check of all the usual suspects -- regulatory filings, financial reports, and changes to analysts' price targets -- was mixed in terms of company-specific news (more on that in a bit), suggesting much of the move was tied to Nvidia's day in the spotlight.
Image source: Getty Images.
The unveiling of Blackwell
Nvidia announced its long-awaited Blackwell architecture, the successor to the company's highly sought-after Hopper processors, which have become the gold standard for AI processing. Huang called the updated B200 graphics processing unit (GPU) the "world's most powerful chip" for AI.
The GB200 superchip combines two B200 processors and a Grace CPU on a single board. Nvidia also developed a new network switch chip -- the NVLink -- that facilitates speedier communication while reducing the burden on the main processors.
These next-generation processors are being packaged into full systems, including the GB200 NVL72, which contains 36 Grace CPUs and 72 Blackwell GPUs in a complete rack-scale design. The liquid-cooled system also contains a 72-GPU NVLink, which "acts as a single massive GPU," delivering 30 times faster performance for the large language models (LLMs) that underlie generative AI.
As Nvidia's chief rival in the AI chip space, AMD's stock price reaction isn't unexpected, as Nvidia raised the bar for AI performance expectations. Analysts from JPMorgan posit that Nvidia "continues to be one to two steps ahead of its competitors," while analysts at Bank of America suggest Nvidia continues to "widen its competitive moat."
Furthermore, the Grace CPU in Nvidia's GB200 was designed by Arm Holdings, which means the company gets a cut of each sale, either in the form of royalties or licensing revenue.
Super Micro Computer, commonly referred to as Supermicro, partners with Nvidia on its own rack-scale, AI-centric servers to ensure optimal integration and performance from Nvidia's AI workhorse processors. Supermicro announced the expansion of its product portfolio to include these latest powerhouse AI chips.
It's easy to see how the ripples from Nvidia's announcements will impact each of these companies in both positive and negative ways.
There was one piece of company-specific news that may have fueled its downturn Tuesday. In a regulatory filing, Supermicro announced a secondary stock offering, with plans to sell 2 million shares of common stock, bringing the total share count to 58.55 million.
Current shareholders are no doubt unhappy about the dilution, which amounts to about 3.5% -- far less than today's stock price drubbing. However, since the stock was recently sitting near an all-time high, it maximizes the amount of capital Supermicro will raise. At its current price of roughly $900 (as of this writing), the company could raise roughly $1.8 billion.
Are any of these stocks a buy?
Most market commentators believe that it's still very early days for the adoption of AI. If they're right -- and I believe they are -- each of these stocks has a vast opportunity ahead. For example:
- Supermicro has become the server of choice for many businesses looking to adopt AI. Furthermore, the company's building-block strategy allows companies to start small and easily expand their existing systems as the need arises.
- Arm Holdings controls a vast library of intellectual property (IP) and will continue to see its licensing and royalty revenue grow from the accelerating adoption of AI.
- AMD offers alternatives to Nvidia's AI processors, as well as CPUs of its own design to accompany them, so it will also benefit from the secular tailwind represented by AI.
In terms of valuation, Supermicro is the least expensive of the group, with a forward price-to-sales ratio of 2, while AMD and Arm Holdings are selling for 9 times and 33 times forward sales, respectively.
However, for those with a long-term outlook and stomach for some increased volatility, each of these stocks represents an intriguing opportunity, especially since it's still early days for AI.
