On November 10, 2020, Apple presented new versions of its all three Macs - the Macbook Air, Macbook Pro, and the Mac Mini — all powered by its new proprietary chip, the M1. The new chip is similar to the iPhone 12's A14 chip, which is found in Apple's newest iPad Pro. For those three devices, this marks a break between Apple and its chip maker, Intel Corp.
On Friday, January 22, 2021, Intel's stock price dove almost 10 percent following an earnings call that took place the previous day between Intel analysts and shareholders, Intel's current CEO, Bob Swain, and its incoming CEO, Pat Gelsinger. Gelsinger takes the helm on February 15, 2021.
One of America's preeminent companies
Gelsinger is an electrical engineer who previously spent 30 years at Intel, working under such legends as Andy Grove, Robert Noyce, and the creator of Moore's Law, Gordon Moore. Moore's law is the observation that the number of transistors in a dense integrated circuit (IC) doubles about every two years.
Intel was founded in 1968 in Mountain View, California by Gordon Moore, a chemist, and Robert Noyce, a physicist who was the co-inventor of the integrated circuit. Both Moore and Noyce had left Fairchild Semiconductor to found Intel. Intel's third employee was chemical engineer Andy Grove, who ran Intel during most of the 1980s and the 1990s, a time when that company was experiencing explosive growth.
On the earnings call, Gelsinger said, "We're interested in resuming that position of the unquestioned leader in process technology." But resuming that position might be difficult for Intel which, over the last 30 years has been the world's largest chipmaker, in what is a $400 billion industry.
Intel's chip designs were considered the best, and it manufactured those chips in its own state-of-the-art foundries. However, over the last several years, Intel has begun lagging behind Taiwan Semiconductor Manufacturing Company (TSM) and Korea's Samsung Electronics Co. (SSNLF), who make chips both for Intel's customers, including Amazon and Apple, and for Intel's competitors such as Advanced Micro Devices Inc. (AMD).
The 7-nanometer problem
Intel has fallen behind its competitors in producing the next-generation 7-nanometer chips. The company announced in July 2020 that those chips would be delayed by six months, and that manufacturing wouldn't begin until late 2022 or early 2023. A nanometer is a billionth of a meter, and the smaller the chip, the more transistors can be packed into devices.
Both TSM and Samsung are already producing 7-nanometer chips, and by the time Intel's 7-nanometer chips come out, TSMC and Samsung might be making far more advanced processors.
At issue is Intel's decision to continue to manufacture its own chips, while its chief competitor, Advanced Micro Devices (AMD), outsources its chip production to TSMC. Gelsinger told those on the earnings call, "I am confident that the majority of our 2023 products will be manufactured internally. At the same time, given the breadth of our portfolio, it’s likely that we will expand our use of external foundries for certain technologies and products."
Recently, Intel's revenue from cloud service providers fell 15% from a year earlier, showing a possible weakness in Intel’s data center business, with many data centers apparently sitting on stockpiles of unused Intel chips.
Both Intel and AMD make the x86 family of chips for PCs, and AMD is making inroads with its high-end desktop processors and in servers. AMD's new TSM-manufactured Zen 3 chip provides a speed boost over Intel.
Overall, Intel's fourth-quarter profit was $1.52 a share on $20 billion of revenue, beating analysts' predictions of $1.11 a share on revenue of $17.5 billion.
The M1 problem
Now, Intel's biggest problem might be named the M1.
The M1 utilizes the same Arm architecture that is used in every smartphone today. An ARM processor is one of a family of CPUs based on the RISC (reduced instruction set computer) architecture developed by Advanced RISC Machines (Arm). Qualcomm, which makes chips for Android phones, is currently touting more-powerful versions of its Snapdragon chips for use in PCs, and several makers of Windows laptops are already using them.
The M1 utilizes an 8-core CPU with four cores dedicated to performance and only used for demanding tasks, and four cores dedicated to efficiency and used for simpler tasks. The M1 also has a 7- or 8-core General Processing Unit (GPU), and a 16-core "Neural Engine" for applications that use machine learning algorithms.
The M1 is also cool, that is, it creates less heat than other chips, which means that devices containing the new chip don't have to include noisy cooling fans. For years, Apple users have complained that Intel-powered Macs run too hot, are too loud, and suffer from degraded performance under demanding workloads.
The M1 also supports 20 hours of video playback, and 17 hours of web browsing under battery power. During its November 2020 announcement, Apple executives described the new chip as the "world's fastest CPU core," and so far, the M1 is living up to its hype.
Because Intel processors and the M1 use two different instruction sets, a software that runs on Intel’s chips requires a "translator" in order to run on Apple's M1 devices. Apple has created such a translator, called Rosetta2, and here's what some normally staid tech reviewers are saying about Rosetta2: "It’s brilliant. I’ve been covering technology professionally for 13 years ... and I can’t recall any form of software emulation or translation that works as well at launch as Rosetta2. I can’t even recall one that comes close."
In fact, since the M1 chip is faster than Intel's chips, even with translation, apps created for Intel chips often run faster on M1-powered Macs.
Only time is going to tell whether Pat Gelsinger and the technologists he is bringing with him to Intel will be able to pull the company out of its doldrums. If not, the long, slow slide of one of America's most iconic tech companies will be painful to watch.