Skip to Content


Can an Old-School Chipmaker Like Intel Be Better Than Nvidia?

November 21, 2018

Silicon is the seventh most abundant element found on earth. Yet, it is one of the most valuable of elements when it comes to the modern economy. Not much can work without it. And trillions of dollars have been generated thanks to taking this common stuff and using modern alchemy to make it into commercial “gold.”

By “gold,” I mean computer chips, found in nearly everything from toasters to cars—and of course the modern umbilical cord known as the smartphone. Among the leaders in silicon alchemy are Intel (INTC) and Nvidia (NVDA). Nvidia came to the market in October of 2000 and since then has generated a return of 1,059.25%—however most of that was earned since 2016. Intel came to the market in October of 1971 and since 1989 has generated a return of 7,818.30%.

So, Nvidia is the upstart in the chip world, but over the past three years has seen its stock mop the floor with Intel’s shares with a return of 368.03% to 47.74%. And how and why did this happen?

One word: Bitcoin.

Bitcoin and its myriad of copycat crypto coins rely on heavy computing power to solve for complex algorithms to create (or “mine”) the coins. And these are heavily dependent on graphics processing units, or GPUs. Nvidia, headquartered in Santa Clara, California, carved out a big chunk of the GPU market, dominating it along with Advanced Micro Devices (AMD).

This surge in demand for GPUs, particularly over the past few years, actually resulted in GPU shortages, providing Nvidia with windfalls. However, like with most manias from tulips to dot-coms, Bitcoin mania eventually passed. This is one of the reasons that Nvidia’s shares have been particularly punished during the Red October and Negative November markets.

GPUs have other purposes in life beyond coin mining. They plug into gaming consoles and high-end PCs and are vital to artificial intelligence, supercomputers and autonomous cars. So, the company has a future in these products, which are 85% or more of its revenues.

While Nvidia is the recent fallen star, Intel still is the champ for all microprocessors, controlling 90% of the globe’s market. It makes central processing units (CPUs) that control the operations of devices and GPUs. And the company also makes GPUs as well as other electronic bits from flash memory to digital imaging gear and parts for networks and communications setups.

So, it’s not surprising that Intel has held on a bit better since October 1. For the trailing year it is positive for investors, with a return of 8.22% compared to Nvidia’s loss of 32.15%.

One of the biggest differences between Intel and Nvidia is that Intel actually makes products, while Nvidia merely outsources. Nvidia relies on two companies for the bulk of its GPUs and related bits—Taiwan Semiconductor (TSM) and Samsung Electronics (SSNLF). And a handful of other companies do some additional packaging and assembly for Nvidia, including Hon Hai—also known as Foxconn (HNHPD)—Advanced Semiconductor (ASXCF) and BYD Auto (BYDDY), which is partially owned by Berkshire Hathaway (BRK-A).

This brings up a pending problem. The US government has many rules restricting sales of technology goods to non-US customers. And with trade negotiations or trade tirades underway, there are threats that off some countries could be cut off from more US technology goods, particularly China. Nvidia, with 70% of its customers in Asia and a significant portion of that base in China, is significantly at risk.

In particular, China’s supercomputer Tianhe-1A (which is the world’s fastest computer) relies on 7,168 of Nvidia’s GPUs. And despite those being made in Taiwan and South Korea, because Nvidia is a U.S. company, that might be restricted.

This could play out very poorly for Nvidia’s shareholders. If China gets cut off, Nvidia’s outsource companies could simply be re-contracted by a re-configured Nvidia that changed its domicile of ownership. And if we play out some chess plays, the US Patent Office could simply cancel out the companies’ patents in retaliation.

Intel is also at risk, with 25% of its sales in China and 40% of its revenues coming from three customers: Dell Technologies, Lenovo (LNVGY) and HP (HPQ)—all with China exposure.

But Intel could more easily shift operations, given its own bases of production and assemblies.

Then there are the divergent focuses on shareholders. Intel pays a dividend that’s ahead of the average of the S&P 500, at 2.52%, and has been rising in distribution by an annual average of 5.92% over the past five years. Nvidia pays barely anything, with a yield of 0.44%, with a payment that was upped by a penny over the past year to 16 cents.

I like the diversity of Intel’s product mix and production capabilities, which is why I have it in one of my Profitable Investing model portfolios. I’d continue to avoid Nvidia, given the changes in the market for GPUs, the competition for GPU market share (including from Intel) and the lack of production capabilities. And of course, the much higher levels of political risk in Nvidia aren’t a plus, either.