For decades, Intel Corporation (NASDAQ:INTC) was at the heart of the technology industry. As the company whose CPUs dominated the PC era, it was second to none. But Intel missed out on the mobile revolution and with PCs losing relevance every year, its leadership position has weakened. Nudging itself into Intel’s place? Nvidia Corporation (NASDAQ:NVDA).
Click to Enlarge
Nvidia started as a PC graphics card manufacturer, but now, NVDA stock is riding a wave of virtual reality gaming, artificial intelligence, deep data analysis, cloud computing and self-driving cars.
As its GPU-accelerated products take over processor-intensive computing, is Nvidia set to dethrone Intel?
GPUs Make the Leap From Graphics to Raw Processing Power
In the old days of computing, graphics processing units (GPUs) played second fiddle to central processing units (CPUs).
Intel’s processors were literally at the heart of almost every computer sold. The CPU was responsible for running the computer, including its operating system. When a GPU was installed, the CPU was still primary, but it farmed off graphics-intensive tasks to the GPU. Most PCs used integrated graphics that were built into their Intel CPU. Graphics cards from companies like Nvidia — with a standalone GPU — were primarily used in specialized gaming PCs.
It helps to think of the CPU as a generalist, while the GPU does one thing over and over. This difference began really turning to the advantage of NVDA about a decade ago.
Desktop PC CPUs from Intel might have four cores, letting the CPU manage a handful of different tasks simultaneously. GPUs were equipped with hundreds of cores, and software developers realized they could take advantage of this raw parallel processing power for much more than just graphics. The GPU could be used to greatly accelerate software performance when repetitive tasks are being performed — encoding video, for example, or data analysis.
GPU acceleration can make some tasks 100 times faster than a CPU alone.
Because the GPU is designed for one primary purpose — everything including computer graphics boils down to crunching numbers — NVDA is able to produce them at far lower cost than a CPU at an equivalent performance level, while producing less heat and using less power.
Why GPUs Are in Higher Demand Than CPUs
Why the demand for GPUs?
Think of how the demand for processor-intensive computing power has increased over the past decade. One early example was Bitcoin. The virtual currency could be purchased, but it could also be “mined” by using a PC to run a complex algorithm in search of Bitcoins. Miners soon discovered that a GPU — with its parallel number-crunching capability — was far more effective than a CPU at mining Bitcoins. Anywhere from 50 to 100 times faster. Soon miners were running rigs consisting of multiple GPUs connected together.
Meantime, GPUs have rapidly become more powerful, with more cores and faster clock rates, even as CPU development has become relatively incremental.
Intel’s new seventh-generation Kaby Lake Core i7 CPU for desktop computers still has four cores. In comparison, Nvidia’s latest Titan X graphics card has 3,584 processing cores on tap. The Intel CPU costs less than a third of the Titan X’s $1200 MSRP, but when used in number-crunching applications, the Nvidia card’s price is a steal.
The technology that has the big potential for future growth involves extreme number crunching. Think virtual reality, the hottest trend in gaming. Cloud services — which are often a single, processor-intensive task like image recognition offloaded to the web. Artificial intelligence, machine learning, data mining and deep data — these all require incredible processing power. Computers running INTC CPUs are too expensive and too general purpose to be running these tasks. Instead, you’re more likely to find server farms powered by GPUs.
NVDA has also integrated its powerful GPUs into the forefront of one of the technology holy grails: self-driving cars.
Nvidia Drive PX2 is the company’s AI car computing platform. With two GPUs providing the horsepower, NVDA says it is capable of 34 trillion deep learning operations per second. Additional units can be run in parallel for fully autonomous driving. Nvidia has already partnered with multiple high-profile automakers to equip their self-driving cars with its technology, including BMW, Volvo and Tesla Motors Inc (NASDAQ:TSLA)
Growing Importance of the GPU Reflected in NVDA Stock
You can track the relative trajectories in INTC and NVDA by comparing their stock prices. Over the past five years, Intel has grown in value by a little more than 40%. That’s not bad, but doesn’t look quite so good when you compare it to the height of PC popularity. Since 2000, when PC sales were notching 14.5% annual sales growth, Intel stock has actually dropped by 10%.
In comparison, Nvidia stock is up by a whopping 650% in the past five years, and nearly 2,600% since 2000.
Intel Isn’t Going Anywhere, But the Age of NVIDIA Has Arrived
Computer sales continue their slow slide, down another 5.7% in 2016. Much of Intel’s revenue is tied to the PC market. Nvidia is not immune to the effect of declining PC sales, either.
But while INTC is trying to tie its future to technology like the Internet of Things (which is struggling with security issues and competing standards), NVDA has leveraged the massive parallel processing power of its GPUs to be a leader in multiple cutting-edge areas like AI, deep data, virtual reality and autonomous vehicles.
Increasingly, it’s looking like we’re entering the age of Nvidia. NVDA stock reflects that reality.
As of this writing, Brad Moon did not hold a position in any of the aforementioned securities.
