Nvidia Stock Hits a Record High. Watch These 3 Things.

Customer Concern

Nvidia is a classic ‘picks and shovels’ play—a company that sells the tools for an industry trend and is therefore positioned to win no matter who comes out on top. 

However, the ‘picks and shovels’ theory only works so long as customers keep lining up to try their luck. Independent analyst Richard Windsor, who publishes Radio Free Mobile, has argued generative AI services such as ChatGPT have set an unsustainable benchmark price of around $20 a month for their services. 

“Freely available models from the open-source community, combined with start-ups who need to get volume for their newly launched services, are going to start eroding the price of the services,” Windsor wrote. He expects the price for AI services to settle at more like $20 a year. 

If that happens, customers might tire of paying Nvidia’s prices. Raymond James estimates it costs Nvidia $3,320 to make its cutting-edge H100 chip, which is sold to customers for $25,000 to $30,000.

Analysts at Deutsche Bank said after Nvidia’s earnings that they expect data-center customers to slow their rate of chip buying to ‘digest’ their purchases at some point. They kept a Hold rating on the stock after the earnings, and recommended waiting for a more favorable entry point. 

China Risks

Nvidia took care to dispel concerns that an unsustainable surge in Chinese demand was powering its growth. The company noted its proportion of revenue coming from China was within its historical average of 20%-25% and said even further restrictions wouldn’t have an immediate material effect on its results. 

But that doesn’t mean there aren’t longer term risks.  

“Over the long term, restrictions prohibiting the sale of our data center GPUs to China, if implemented, will result in a permanent loss of an opportunity for the U.S.,” Nvidia’s Chief Financial Officer Colette Kress told analysts on an earnings call. 

It’s not only geopolitics that poses a risk to Nvidia’s business in China. The company’s second-quarter automotive revenue fell 15% from the previous quarter, largely due to lower demand in China. 

With China facing mounting economic troubles, Nvidia customers such as
Baidu
(BIDU), ByteDance, Tencent, and
Alibaba
(BABA) might not be so keen to pay up for its chips and could look to local suppliers instead. Local media in China recently reported the head of AI company HKUST Xunfei said Huawei’s compute GPU capabilities are now on par with Nvidia’s A100—an advanced chip which Nvidia can’t sell to Chinese customers. 

Increasing Competition  

Nvidia has a dominant position in sales of graphics-processing units for AI related purposes, with around a 90% share of the market. 

However, that’s not guaranteed to be the case forever. Advanced Micro Devices (AMD) is launching its MI300 data-center GPU this year and could take advantage of constraints in Nvidia’s supply chain to grab some market share. 

Some of Nvidia’s own customers are also shaping up to be its competition.
Alphabet’s
(GOOGL) Google has its custom Tensor Processing Units, or TPUs, although it also partners with Nvidia for various AI applications.
Amazon.com
(AMZN) also has custom chips which it also offers as alternatives to Nvidia’s GPUs. 

None of this means that Nvidia is guaranteed to go down from here. Its rapid growth has kept worries about its high valuation in check and it has developed a formidable competitive position based on both its hardware and its software offering. So far, it has defied all of the concerns cited. 

However, it’s not necessarily one factor or another that could hit Nvidia stock. A mixture of issues and the increasingly high hurdles it faces to beat expectations could be enough to puncture confidence in the stock. The company’s outstanding performance so far is only moving the bar higher for the stock to keep pushing upward.

Write to Adam Clark at [email protected]