Why Intel's Competitive Edge Is Crumbling

Intel (INTC -3.67%) is the largest chipmaker in the world, with products that range from personal computing to data centers. Currently, the company is also the market-leading provider of central processing units (CPUs) for laptops, desktops, and servers. But competition from chipmakers like ARM, Advanced Micro Devices (AMD +0.04%), and NVIDIA (NVDA -1.46%) has put pressure on Intel in recent years, and it's starting to take a toll on the iconic company. Here's why investors should be worried.

Image source: Getty Images.

Competition with customers

Intel's chips are built on the x86 instruction set, a fundamentally different architecture than ARM's chips. The difference between the two is simple: Intel's chips are made for performance, and ARM's are designed for efficiency.

However, ARM's chips also offer a level of customization that Intel cannot match. That's because Intel sells its chips directly, while ARM licenses its CPU architecture to partners. This has allowed various tech companies to build on ARM's energy-efficient architecture, adding their own customizations to improve performance. In fact, as of November 2020, the most powerful supercomputer in the world runs on custom-built ARM-based processors.

Two years ago, Amazon started using its own ARM-based Graviton processors to replace Intel chips in Amazon Web Services data centers. For the sake of comparison, Amazon's custom-built chips offer up to 40% better price performance than Intel's. Apple recently made a similar decision, replacing Intel chips in its MacBook Air and MacBook Pro  laptops with its own M1 chips. These custom-built ARM-based processors  actually surpass the performance of Intel's latest Tiger Lake CPUs for laptops. And just a few weeks ago, Microsoft followed suit, announcing that it would design its own ARM-based chips for its data centers and Surface PCs.

This is bad for Intel. The company is losing customers in markets it has historically controlled, like data center and PC, as those buyers' preferences shift. Companies want custom solutions, and Intel's current products cannot provide them.

Competition with NVIDIA

For years, Intel's Xeon CPUs have dominated the data center market. However, NVIDIA has also become a critical player in this space, since its graphics processing units (GPUs) are much better at accelerating heavy workloads like artificial intelligence. In fact, NVIDIA's recently launched Ampere GPU is up to 237 times faster than Intel's CPUs.

In 2020, NVIDIA doubled down on its data center business and acquired networking solutions provider Mellanox for roughly $7 billion . Compared to Intel's products, Mellanox's networking solutions offer better performance and, as a result, have captured a larger market share. For NVIDIA, this acquisition has already allowed the company to combine its GPUs with Mellanox technology to create products like the BlueField-2X DPU, which enhances a Mellanox SmartNIC (network interface card) with ARM CPU cores and the AI capabilities of NVIDIA's GPUs. In other words, NVIDIA can bundle its accelerators and networking products together, enhancing data center security and efficiency. This not only fits the trend toward high-performance networking in data centers, but also further differentiates NVIDIA's offering, as Intel does not have a comparable product.

Even worse for Intel, NVIDIA announced its intention to acquire ARM. If approved, this would combine NVIDIA's best-in-class accelerators, Mellanox's high-performance networking, and ARM's energy-efficient CPUs. And while ARM processors have only a tiny portion of the data center market today, NVIDIA's focus on R&D could help ARM produce a server CPU that further displaces Intel from data centers.

Competition with AMD

In recent years, AMD's Ryzen chips have helped the company take significant market share away from Intel in both the laptop and desktop CPU markets. For desktops, AMD released its Ryzen 5000 series chips in November 2020. Compared to Intel's latest desktop chips, the 10th-generation Core CPUs codenamed Comet Lake, these Ryzen chips offer better performance across the board -- gaming, single-threaded (light) computing, and multi-threaded (heavy) computing. However, Intel plans to launch Rocket Lake, its 11th-generation Core CPUs for desktop, in the first quarter of fiscal 2021. This new technology may help the company regain some lost ground.

Additionally, AMD has also taken share in the server market (data centers) with its second-generation EPYC CPUs, codenamed Rome. And Intel has not helped itself here -- the company has delayed the launch of Ice Lake, a third-generation Xeon scalable processor for servers. This chip was originally set for production in early 2020, but management now says Ice Lake won't ramp until the first quarter of fiscal 2021. This could delay the launch of its next server chip, Sapphire Rapids, which is currently set for late 2021. If that happens, AMD may have an easy time taking more of this market with the upcoming release of its third-generation EPYC CPU, codenamed Milan.

Data source: Mercury Research.

Intel's financial performance

All of this competition has put pressure on Intel, and it has weighed heavily on the company's financial performance. Overall revenue growth has been slow, up only 26% over the last three years, compared to NVIDIA's 65% and AMD's 71% growth. Additionally, NVIDIA's and AMD's gross profit margins have expanded, while Intel's have dropped from 62% in 2017 to 53% in the most recent quarter, indicating a loss of pricing power. Investors should watch these metrics going forward.

On the positive side, Intel still generates much more revenue than its competition, which means the company has more cash to spend on operating and capital expenses. If Intel can put that advantage to work, it may be able to repair its crumbling competitive edge. However, given the strong performance of Intel's competitors, and the resultant deterioration in Intel's financial performance, I think investors should be cautious about buying this stock.