NVIDIA (NASDAQ: NVDA) The company’s shares have soared 145% over the past year due to the semiconductor company’s key position in the artificial intelligence (AI) economy. In particular, Nvidia’s graphics processing units (GPUs) have become the industry standard for accelerating complex data center workloads such as training machine learning models and running artificial intelligence applications.
Earlier this year, the Wall Street Journal reported that “NVIDIA’s chips power all the most cutting-edge AI systems, and the company’s market share is estimated at more than 80%. Some analysts believe that NVIDIA controls 95% of the AI chip market. But Amazon is (Nasdaq: AMZN) He recently issued a warning to Nvidia shareholders.
Amazon CEO Andy Jassy issues warning to NVIDIA shareholders
Amazon CEO Andy Jassy recently told analysts that no single artificial intelligence chip will fit every use case, simply because businesses and developers want options. “You saw a few years ago when some companies tried to claim that TensorFlow was going to be the only important machine learning framework, and then PyTorch and others overtook it,” Jassy explained.
Nvidia consistently breaks performance records in MLPerf benchmarks, an objective test that measures how well AI systems perform training and inference tasks. But Nvidia’s graphics processing units (GPUs) are also very expensive: the company’s H200 GPUs can cost up to $40,000 each, and entire server racks can cost millions. Jassy sees robust demand for products that fall between those two extremes, and Amazon is developing custom AI silicon.
“We have a close partnership with Nvidia and have the largest selection of Nvidia instances, but we have heard loud and clear from our customers that they want better price/performance,” Jassy told analysts. “That’s why we’ve invested in our own custom silicon: Trainium for training and Inferentia for inference.”
Importantly, Amazon has previous experience developing custom chips: For several years the company has been designing general-purpose central processing units (CPUs) called Graviton processors. These chips are based on the Arm architecture and, according to Jassy, offer 30% to 40% better price-performance than flagship x86 processors from AMD and Intel.
Amazon has the same goal with its custom AI silicon. Trainium and Inferentia aim to offer customers a cost-effective alternative rather than competing with Nvidia GPUs on performance alone. Some companies are willing to pay a premium for the fastest AI chips, while others are happy to use slower chips to save costs.
The story continues
Morgan Stanley expects custom silicon to become more popular
Amazon isn’t the only company designing custom silicon. Broadcom has developed custom AI chips for Alphabet’s Google and Meta Platforms, and earlier this year landed its third major customer. The company didn’t disclose the customer’s name, but Reuters reported it was TikTok’s parent company, ByteDance. Additionally, analysts at JP Morgan recently estimated that Broadcom also won a contract to make custom AI chips for OpenAI, bringing its total number of major customers to four.
Morgan Stanley’s Joseph Moore expects this trend to continue: “Hyperscalers like Amazon, Google, and Meta are using custom silicon for their data centers due to the cost and energy constraints of AI computing. As hyperscalers continue to invest in AI infrastructure, we expect demand for energy-efficient, cost-optimized custom chips to outpace growth in commodity GPUs,” he wrote in a client note.
Here’s the big picture: Nvidia currently dominates the AI processor market, but according to information from Amazon and Morgan Stanley, the company will almost certainly lose market share over the next few years. This doesn’t mean that Nvidia will lose its leadership position in the AI chip market — I don’t think that will happen — but as more cost-effective alternatives become available, some spending will inevitably be diverted from expensive Nvidia GPUs.
Additionally, Nvidia will likely lose some pricing power as it fights to maintain its market share. In fact, the company’s gross margins fell 330 basis points sequentially in the most recent quarter, despite strong demand for AI chips and a 122% increase in revenue. Investors should expect this trend to continue in the coming quarters.
But Nvidia isn’t necessarily a bad investment. In fact, Wall Street expects the company’s earnings to grow 37% annually over the next three years. At that consensus estimate, its current valuation of 56 times earnings looks reasonable. These figures work out to a PEG ratio of 1.5, well below its three-year average of 3.1.
Personally, I believe that investors who want to profit from the artificial intelligence boom should bet on Nvidia, and I’m not alone in that belief: Argus’ Jim Kelleher recently said, “In the era of deep learning, AI and GPU-driven application acceleration, we believe[Nvidia]is a stock most tech investors should own.”
Should I invest $1,000 in Nvidia right now?
Before you buy Nvidia stock, consider the following:
The analyst team at Motley Fool Stock Advisor has identified 10 stocks that investors should buy right now, and Nvidia is not among them. The 10 selected stocks have the potential to generate big gains over the next few years.
Let’s look back at April 15, 2005, when Nvidia made this list… if you had invested $1,000 at the time of our recommendation, you would have $731,449.!*
Stock Advisor offers investors an easy-to-follow blueprint for success, with portfolio construction guidance, regular updates from analysts and two new stock picks every month. Stock Advisor The service is More than 4 times First S&P 500 recovery since 2002*.
View 10 stocks »
*Stock Advisor returns as of August 26, 2024
John Mackey, former CEO of Amazon subsidiary Whole Foods Market, is a member of The Motley Fool’s board of directors. Suzanne Frey, an executive at Alphabet, is a member of The Motley Fool’s board of directors. JPMorgan Chase is an advertising partner of The Ascent, a Motley Fool company. Randi Zuckerberg, former director of market development and public relations at Facebook and sister of Meta Platforms CEO Mark Zuckerberg, is a member of The Motley Fool’s board of directors. Trevor Jennewein has invested in Amazon and NVIDIA. The Motley Fool has invested in and recommends Advanced Micro Devices, Alphabet, Amazon, JP Morgan Chase, Meta Platforms, and NVIDIA. The Motley Fool recommends Broadcom and Intel and recommends Intel’s November 2024 $24 call options. The Motley Fool has a disclosure policy.
The post Amazon Issues Serious Warning to Nvidia Shareholders: Are AI Stocks in Peril? was originally published by The Motley Fool.
