Intel (INTC 1.60%) was once considered a stable tech stock that generated reliable returns. Its position as the world's largest manufacturer of x86 CPUs for PCs and data centers gave it a wide moat, and its robust cash flow growth comfortably funded its big buybacks and dividends.
However, Intel's stock declined about 15% this year as the Philadelphia Semiconductor Index surged 50%. Let's look back at why Intel underperformed the broader industry, and see if it might fare better next year.
How Intel lost its mojo
For decades, Intel led the CPU market by following "Moore's Law," a prediction set by its co-founder Gordon Moore to double the number of transistors in the same area of silicon every two years. This was the foundation of the two-year "tick-tock" cycle that drove Intel's business for half a century.
In each "tick" launch, Intel shrank a chip down to a smaller size. In each "tock" update, the chip was upgraded but the size remained the same. However, it became increasingly difficult for Intel to manufacture smaller chips at its own foundry, and Moore's Law faded away in recent years.
As Moore's Law ended, Intel's foundry fell behind TSMC (TSM 0.92%), the world's largest contract chipmaker, in the "process race" to create smaller and more power-efficient chips. As a result, "fabless" chipmakers that outsourced their chips to TSMC -- including AMD (AMD 1.51%) and Apple, which recently replaced Intel's Mac chips with its own silicon -- started producing more advanced chips than Intel.
Throughout late 2018 and all of 2019, Intel struggled to ramp up its production of newer 10nm chips. Those difficulties disrupted the production of its 14nm chips, which caused a massive CPU shortage for PC makers. At the same time, a resurgent AMD launched its new generation of Zen CPUs, which rectified the poor single-threaded performance of its previous Bulldozer-class CPUs. AMD didn't suffer any shortages, since it outsourced its chips, and many PC makers started using its cost-efficient CPUs again.
Between the fourth quarters of 2016 and 2020, AMD more than doubled its share of the CPU market from 17.8% to 38.4%, according to PassMark Software, as Intel's share plunged from 82.2% to 61.6%.
But the pain doesn't end there. In July, Intel stunned investors by admitting its new 7nm process was "trending approximately twelve months" behind its internal target -- which meant those next-gen chips wouldn't arrive until 2022 and 2023. By comparison, TSMC is already manufacturing 7nm chips for AMD and other chipmakers, and it will likely be producing 3nm chips by 2022.
But how bad was the damage?
Intel's situation seems dire, but the company is still growing. Its revenue and earnings rose 2% and 6%, respectively, last year, as the growth of its data center chips offset its supply issues in the PC market.
In the first nine months of 2020, Intel's revenue and earnings both rose 12% year-over-year. Its data center revenue grew as the elevated usage of cloud services during the pandemic boosted demand for its chips, while surging shipments of PCs for remote work and online learning temporarily boosted its client computing revenue.
For the full year, analysts expect Intel's revenue to rise 5%, and for its earnings to stay flat as the pandemic-induced tailwinds wane. Next year, they expect its revenue and earnings to decline 6% and 7%, respectively, as it falls further behind AMD and TSMC.
We should always be skeptical of analysts' forecasts, but it's tough to ignore Intel's recent missteps. Intel's CEO Bob Swan, the former CFO who took the helm after Brian Krzanich's abrupt resignation in 2018, has mainly focused on reducing the company's capex, boosting its buybacks, and divesting its profitable NAND memory chipmaking business instead of resolving its pressing R&D issues.
Those moves suggest Swan is more interested in financial engineering than chip engineering -- and those myopic strategies could cause Intel to fall further behind AMD. Meanwhile, Swan's two leading R&D plays -- discrete GPUs and automotive chips -- probably won't generate enough revenue to offset its ongoing challenges in the CPU market.
But can we consider Intel an undervalued income stock?
Intel trades at just 11 times forward earnings and pays a forward dividend yield of 2.6%. It spent just 27% of its free cash flow on its dividends over the past 12 months, which gives it plenty of room for future hikes.
That low valuation and respectable yield should limit Intel's downside potential, but there are plenty of other cheap dividend stocks to buy in this growth-oriented market. Intel probably won't attract the bulls again until it resolves its R&D problems, decides to go "fabless" like AMD, or hires new leaders to breathe fresh life into its aging, insular business -- so it's better to avoid this stock and stick with better-run chipmakers for now.