Intel (INTC 0.64%), the world's largest manufacturer of x86 CPUs for PCs and data centers, was once considered an ironclad investment.
But over the past three years, Intel's stock slipped about 5% as the Philadelphia Semiconductor Index advanced nearly 80%. Let's identify the three key problems that weighed down Intel's stock, and see if the 52-year-old chipmaker can resolve these issues in a timely manner.
1. Resolving its manufacturing issues
Back in the 1965, Intel co-founder Gordon Moore predicted the number of transistors in the same area of silicon would double every two years. That prediction, known as "Moore's Law", supported Intel's two-year "tick-tock" business model for about half a century.
Image source: Getty Images.
In each "tick" launch, a smaller architecture (currently measured in nanometers) is introduced. In each "tock" launch, a new microarchitecture is launched, but the size remains the same. But the tick-tock cycle gradually lengthened due to the difficulty of manufacturing smaller chips, and Moore's Law fizzled out in recent years.
Over the past two years, Intel has struggled to simultaneously develop and manufacture 14nm, 10nm, and 7nm chips -- which resulted in CPU shortages and disappointing delays for newer chips.
Its main rival in the "process race" to create smaller chips, TSMC (TSM -0.34%), didn't suffer the same problems. As a result, "fabless" chipmakers like AMD (AMD -0.35%) and NVIDIA (NVDA -3.33%), which didn't manufacture their own chips, made smaller chips than Intel by outsourcing their production to TSMC.
Therefore, Intel can shut down its foundry and go fabless, or it can double down on its foundry investments to catch up to TSMC. Going fabless would cut costs and resolve its production issues, but could also cripple its ability to surpass AMD in the future. Ramping up its foundry investments would be costly, but it could help it reclaim the process lead and produce "best in breed" chips again.
Either of those decisions might get investors interested in Intel again. Unfortunately, the company still plans to manufacture its own CPUs, while only partly outsourcing the production of its new discrete GPUs to TSMC.
2. Identifying AMD and NVIDIA as long-term threats
Intel's development and manufacturing issues allowed AMD to expand its share of the global CPU market from 17.8% to 36.5% between the fourth quarter of 2016 and 2020, according to PassMark Software. Intel's share plunged from 82.2% to 63.5% during the same period, with losses across the desktop, laptop, and server markets.
Image source: Getty Images.
AMD is challenging Intel with smaller and more power-efficient chips across all three markets, and its planned purchase of Xilinx, the leading producer of programmable FPGAs (field-programmable gate arrays), will allow it to counter Intel's takeover of FPGA maker Altera five years ago.
NVIDIA's planned purchase of SoftBank's Arm Holdings, which designs CPUs that compete against Intel and AMD's x86 CPUs, is another red flag. Arm's chips are primarily used in mobile devices, but they're gradually gaining ground in PCs and data centers with higher-end designs.
The combination of NVIDIA's market-leading discrete GPUs and more powerful Arm CPUs could loosen the x86 architecture's iron grip on those markets, and erode Intel's position as the world's top CPU maker.
AMD and NVIDIA are both getting ready to disrupt Intel's core business, but Intel's management didn't mention either chipmaker in the company's latest conference call. If Intel wants to win back the bulls, it needs to call out both chipmakers as major threats and reveal some plans to counter their aggressive moves.
3. Hiring an outsider CEO with an engineering background
CEO Bob Swan, who succeeded Brian Krzanich after his abrupt resignation in mid-2018, was Intel's CFO before he took the helm. Unfortunately, Swan's background in finance arguably leaves him at a major disadvantage against AMD's Lisa Su and NVIDIA's Jensen Huang, who are both electrical engineers.
Swan hasn't been able to resolve Intel's pressing R&D and manufacturing issues over the past two years. He also reduced Intel's capex this year while ramping up its multi-billion dollar buybacks and divesting its profitable NAND memory chip business.
Those moves all suggest Swan is more interested in financial engineering than electrical engineering, and that shortsightedness could cause Intel to fall further behind TSMC and its fabless clients. If Intel wants to excite investors again, it should buck its trend of promoting insiders and hire an outsider CEO who can bring fresh ideas to the table.
The key takeaways
Intel's business isn't headed off a cliff yet, but it's falling into the same traps of myopic buybacks and divestments that have ensnared other mature tech companies like IBM. It's not too late to turn things around, but Intel's dismal third-quarter report indicates it needs to take drastic measures soon.
