Why Intel’s competitive advantage is disintegrating

Intel (NASDAQ: INTC) is the world’s largest chip maker, with products ranging from personal computing to data centers. Today, the company is also a leading supplier of central processing units (CPUs) for laptops, desktops and servers. But competition from chip makers like ARM, Advanced micro devices (NASDAQ: AMD)and NVIDIA (NASDAQ: NVDA) has put pressure on Intel in recent years and is starting to hurt the iconic company. That is why investors should be concerned.

Competition with customers

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

However, ARM 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 allowed several technology companies to build on ARM’s energy efficiency architecture, adding their own customizations to improve performance. In fact, in November 2020, the world’s most powerful supercomputer runs on custom ARM-based processors.

Two years ago, Amazon began using its own ARM-based Graviton processors to replace Intel chips in Amazon Web Services data centers. For comparison, Amazon’s custom chips offer up to 40% better price performance than Intel’s. Apple recently made a similar decision, replacing the Intel chips in its MacBook Air and MacBook Pro laptops with its own M1 chips. These custom ARM-based processors really outperform Intel’s latest Tiger Lake CPUs for laptops. And just a few weeks ago, Microsoft did the same, 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, such as data centers and PCs, as buyers’ preferences change. Businesses want customized solutions and current Intel products cannot deliver them.

Competition with NVIDIA

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

In 2020, NVIDIA doubled its data center business and acquired the network solutions provider Mellanox for about $ 7 billion. Compared to Intel products, Mellanox network solutions offer better performance and, as a result, have captured a larger share of the market. For NVIDIA, this acquisition has already enabled 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 GPUs. In other words, NVIDIA can bundle its accelerators and network products, increasing the security and efficiency of the data center. This not only fits the trend for high performance data center networking, 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 accelerators, Mellanox’s high-performance network and ARM’s energy-efficient CPUs. And while ARM processors have only a small portion of the data center market today, NVIDIA’s focus on R&D could help ARM produce a server CPU that further drives Intel away from data centers.

Competition with AMD

In recent years, AMD’s Ryzen chips have helped the company gain a significant share of Intel’s market, both in the notebook and desktop markets. For desktops, AMD launched its Ryzen 5000 series chips in November 2020. Compared to Intel’s latest desktop chips, 10th generation Core CPUs codenamed Comet Lake, these Ryzen chips offer better performance across the board. board – games, single-threaded (light) computing and multithreaded (heavy) computing. However, Intel plans to launch Rocket Lake, its 11th generation of desktop Core CPUs, in the first fiscal quarter of 2021. This new technology could help the company recover some of the lost ground.

In addition, AMD also gained market share in servers (data centers) with its second generation EPYC CPUs, code-named Roma. And Intel did not help itself here – the company delayed the launch of Ice Lake, a scalable third-generation Xeon processor for servers. This chip was originally set for production in early 2020, but management now says that Ice Lake will not increase until the first fiscal quarter of 2021. This may delay the launch of its next server chip, Sapphire Rapids, which is set for the end of 2021 If this happens, AMD may find it easier to get more out of this market with the next launch of its third generation EPYC CPU, code-named Milan.

Data source: Mercury Research.

Intel’s financial performance

All of this competition put pressure on Intel and weighed heavily on the company’s financial performance. Overall revenue growth has been slow, just 26% over the past three years, compared to 65% growth for NVIDIA and 71% for AMD. In addition, NVIDIA and AMD’s gross profit margins increased, while Intel’s fell from 62% in 2017 to 53% in the most recent quarter, indicating a loss of price power. Investors should look at these metrics going forward.

On the positive side, Intel still generates much more revenue than its competitors, which means that the company has more money to spend on capital and operating expenses. If Intel succeeds in putting that advantage to work, it could fix its declining competitive advantage. However, given the strong performance of Intel’s competitors and the resulting deterioration in Intel’s financial performance, I think investors should be cautious when buying these shares.