This best-in-class semiconductor company is down more than 40% from its peak. But has it fallen far enough?
It has been a trying time for investors across the technology space, even those holding shares of best-in-class, profitable companies like Nvidia ( NVDA 0.35%). Nvidia pioneered the use of graphics processing units (GPUs), and not just for high-end gaming and visualization. It also adapted the parallel-processing capabilities of GPUs to help accelerate artificial intelligence applications, which require huge amounts of extremely fast processing power.
While macroeconomic headwinds have taken down the stock from its peak in November, there are still a lot of positive things going on at Nvidia, and it has an exciting pipeline of innovation. So after a steep decline that has it trading more than 40% below its high — and with a bit of bounce underway in recent weeks — is now the time to buy this all-star chip stock? Recent results came in strong, with some external headwinds
During a challenging quarter for the economy, Nvidia continued to shine. Its first quarter revenue rose 46% year over year to $8.29 billion, and adjusted ( non-GAAP ) earnings gained 49% to $1.36 per share. Both figures beat analysts’ consensus expectations. However, Nvidia did guide for a sequential revenue decline due to a $500 million headwind caused by the latest COVID-19 lockdowns in China (which it expects to sap $400 million in gaming revenue) and sales not being made in Russia ($100 million in data center revenue). While the lost revenue from Russia may never be recovered, the missing sales to customers in China should reappear as the country lifts restrictions in Shanghai and Beijing.
But the most important story continues to be Nvidia’s data center segment, which surpassed the gaming segment in revenue once again. This happened for the first time back in fiscal Q2 2020, but that period fell during the early chaos of the pandemic. Last quarter, Nvidia’s data center revenue was up a staggering 83% year over year, better than the gaming segment at 31%, with increased momentum and visibility for the rest of the year.
While gaming sales should remain solid over the long term, thanks to the growth of video games broadly and the expected rise of the metaverse , Nvidia’s data center segment should continue to outperform and become its most important segment by a significant margin.
This is because artificial intelligence is just now taking off in earnest due to a couple of key breakthroughs. On the recent conference call with analysts, CEO Jensen Huang pointed to the new innovation of transformers ushering in a sea change for the AI industry. Previously, in order to use and benefit from AI, a business would have to organize and label all of its data — a hugely time-consuming, expensive, and sometimes impossible process. However, with transformers, a machine can train itself without the need for human-labeled data.
This innovation is opening up AI insights to a much broader range of industries, where the technology’s use is a key enabler and competitive advantage. And the more easily and affordably businesses can access and use AI, the better it will be for Nvidia’s data center segment. The transformer innovation — which has taken place over just the past couple of years — is a big reason Nvidia’s data center revenue has tripled in just two years. Image source: Getty Images. Product launches this year should power data center sales
In addition to these industrywide breakthroughs, Nvidia has also been innovating at a fast clip. Its new H100 chip, which contains over 80 billion transistors and promises up to 30 times the performance of the A100, is set to launch later this year. Keep in mind, the A100 is the chip delivering all of the company’s current outstanding data center performance, so the segment is likely in for continued growth after the H100 launches.
Nvidia will also be introducing its first central processing unit (CPU), dubbed Grace, later this year. CPUs have been the domain of Intel and Advanced Micro Devices , but Nvidia is coming out with an ARM-based 144-core chip that’s built specifically for AI applications in data centers. Now, Nvidia will have a full stack of chips for a complete data center, including GPUs, CPUs, DPUs (networking processors), systems-on-chips (SOCs), switches, and interconnects. A vertically integrated full ecosystem for data centers could enable lots of growth with increasing margins for Nvidia.
Aside from its data center chips, Nvidia will also refresh its RTX gaming chips later this year. […]
Protected by Patchstack