Nvidia (NASDAQ:NVDA) is facing important threats on two fronts. As a result of these issues, I’ve become cautious about NVDA stock at this point.
In a previous column, I wrote that artificial intelligence (AI) was the key to the huge jump of the company’s data center revenue in the fourth quarter of last year, and the strong growth of its data center revenue, in turn, was a key reason that its results came in ahead of analysts’ average outlook. Revenue from data centers accounts for about one-third of Nvidia’s overall sales and is important to its strategy and to NVDA stock going forward.
But a recent innovation by Rice University, achieved in collaboration with Intel (NASDAQ:INTC), looks poised to threaten Nvidia’s competitive advantage when it comes to AI. Specifically, the researchers reportedly have been able to radically increase the speed of Intel’s Xeon chips which are used for AI training. As a result, Xeon is now “3.5 times faster than Nvidia’s Tesla (chips) in AI deep learning,” according to wccftech.
Further, the Mar. 2 press release announcing the news points out that the researchers were able to achieve the speed increase on computer processing units made by Intel, instead of the much more expensive graphics processing units in which Nvidia specializes.
According to Anshumali Shrivastava, the assistant professor in Rice’s Brown School of Engineering who led the team which invented the new technique, “a top-of-the-line GPU platform like the ones Amazon (NASDAQ:AMZN), Google and others offer for cloud-based deep learning services has eight Tesla V100s and costs about $100,000.”
Although I could not find the price of the 44-core Intel chip which the Rice/Intel team used, other types of Xeon chips can be obtained for only a few hundred dollars each.
It will likely take time for all the kinks of the new technique to be worked out and to convince data centers to try it. But Nvidia’s sales could start to be meaningfully negatively impacted by the discovery six months to a year from now.
According to Jon Peddie Research, AMD’s shipments of GPUs jumped 22.6% in Q4 versus Q3, while Nvidia’s fell 1.9% quarter-over-quarter. Nvidia’s shipments fell in Q4 even though it’s usually a stronger period for GPUs than Q3. AMD’s share of the GPU market increased by 3% in Q4 versus Q3, while Nvidia’s share fell by almost 1%.
NVDA Stock Still Has Strong, Positive Catalysts
Barron’s reported that ” ADAS (advance driver-assistance systems) is coming on strong,” while Nvidia ” launched a new ADAS offering in January 2019, and NVDA is already partnering with 370 companies in the auto sector.”
I also predicted that “by the end of this year, many companies — including multiple Nvidia partners — will be offering services with truly autonomous vehicles.” But in 2019, the company’s auto business accounted for less than 10% of its total sales.
Meanwhile, Nvidia’s pending acquisition of Mellanox (NASDAQ:MLNX) could enable it to continue benefiting from the explosive growth of AI, even if its GPUs lose market share to Intel’s CPUs. That’s because Mellanox’s technology allows servers to essentially use the memory of other servers without any of them utilizing their operating systems. That, in turn, helps enable the creation of huge “neural networks” which are used to facilitate AI.
Finally, as InvestorPlace contributor Louis Navellier pointed out recently, the company’s products are facilitating the deployment of 5G technology.
The Bottom Line on Nvidia Stock
Without a doubt, Nvidia still has impressive products and powerful, positive drivers. However, it’s also facing very important, strong threats from Intel and AMD. And, with a trailing price-earnings ratio of 56, NVDA stock is far from cheap. Given all of these points, along with the current weakness of the market, I’d avoid Nvidia at this point.
As of this writing, Larry Ramer did not own shares of any of the aforementioned companies. Larry Ramer has conducted research and written articles on U.S. stocks for 13 years. He has been employed by The Fly and Israel’s largest business newspaper, Globes. Larry began writing columns for InvestorPlace in 2015. Among his highly successful, contrarian picks have been GE, solar stocks, and Snap. You can reach him on StockTwits at @larryramer.