At first glance, Intel (INTC -0.76%) might seem like a reliable long-term investment. It's the world's largest producer of x86 CPUs for PCs and servers, it's firmly profitable, and it pays consistent dividends. But if you had invested $3,000 in Intel on the first trading day of 2015, your investment would only be worth about $2,300 today.
If you had invested $3,000 into Intel's rival Advanced Micro Devices instead, your investment would be worth roughly $78,000. A $3,000 investment in Taiwan Semiconductor Manufacturing (TSM -2.45%), Intel's top competitor in the foundry market, would also have blossomed into nearly $11,000. So why did Intel underperform its industry peers by such a wide margin?
Image source: Intel.
Losing the process lead to TSMC
Intel is an integrated device manufacturer (IDM) that manufactures most of its own chips at its first-party foundries. By comparison, "fabless" foundries like AMD outsource their production to third-party contract chipmakers like TSMC. For decades, Intel manufactured the world's smallest and densest chips by adhering to Moore's Law, a prediction from its co-founder Gordon Moore that the number of transistors within the same area of silicon could double every two years.
Intel formalized that two-year cadence as its "tick-tock" model in 2007. With every "tick," it shrank its chips to a smaller node (which is currently measured in nanometers). With each "tock" the following year, it upgraded the chip's microarchitecture.
But starting in 2012, the gaps between Intel's tick and tock launches became longer as it struggled to manufacture smaller chips. In 2014 TSMC started to install ASML's extreme ultraviolet (EUV) systems -- which are used to etch circuit patterns onto silicon wafers -- to produce more advanced chips for Apple and other chipmakers.
Instead of following TSMC's example, Intel didn't aggressively expand its EUV fleet even thought it had been one of ASML's top investors back in 2012. As a result, TSMC eventually overtook Intel in the "process race" to manufacture smaller, denser, and more power efficient chips in 2020. That victory also enabled TSMC's client AMD to launch more advanced chips than Intel. It also maintained a more stable supply of chips as Intel grappled with product delays and chip shortages.
Faced with all those challenges, Intel's share of the PC market plunged from 82.2% to 63.3% between the fourth quarters of 2016 and 2022, according to PassMark Software, as AMD's market share nearly doubled from 17.8% to 34.6%.
Losing the mobile market to Arm
Intel's other glaring failure was its loss of the mobile CPU market to Arm-based chipmakers like Qualcomm and MediaTek. Instead of licensing the power-efficient Arm architecture, Intel stuck with the same power-hungry x86 architecture it used for its PC and server chips.
Intel was so confident in its x86 architecture that it actually sold its own Arm-based chipmaker, Xscale, to Marvell in 2006. It also refused to produce mobile CPUs for Apple's first iPhone, which launched in 2007, because it believed Apple could never sell enough iPhones to justify its own development and production costs.
Intel introduced its own Atom chips for mobile devices in 2008, but they were less power efficient than Arm chips and suffered some compatibility issues with apps developed for Arm devices. In 2014, Intel started to aggressively subsidize its Atom shipments with billions of dollars in "contra revenues," which included big rebates for OEMs, but it eventually abandoned that futile and loss-leading effort in 2019. As a result, Intel largely missed the big shift from PCs toward mobile devices.
An inability to maintain stable leadership
As Intel grappled with all those challenges, it kept shifting gears under three different CEOs within just four years. In 2018, Brian Krzanich -- who was widely blamed for Intel's loss of its process lead to TSMC, its costly mobile mistakes, and the "diworsification" of its business -- resigned due to an inappropriate relationship with an employee.
Krzanich's successor, his former CFO Bob Swan, was ousted in early 2021 after being criticized for prioritizing cost-cutting measures, buybacks, and divestments over reclaiming the process lead from TSMC. Swan also briefly considered going fabless and outsourcing Intel's production to TSMC. He was replaced by Pat Gelsinger, who steered Intel in the opposite direction and doubled down on upgrading the company's plants, buying new EUV systems from ASML, and catching up to TSMC in the process race by 2024 or 2025. It's trying to spin all those plates as the growth of the PC market stalls out in a post-pandemic market.
Therefore, Gelsinger needs Intel's investors to brace for slower revenue growth and rising expenses for the foreseeable future as it tries to rectify its previous mistakes. That's why analysts expect Intel's revenue and earnings to decline in both 2022 and 2023 -- and why its stock still can't be considered a bargain at 16 times forward earnings.
