HomeArticle

Apple chips: Complete a revolution?

半导体行业观察2025-07-01 10:17
Apple's self-developed ARM chips are impacting the x86 market. Intel is laying off employees and undergoing transformation, and domestic manufacturers are entering the market.

Recently, the news that Intel plans to shut down its automotive business and lay off employees has attracted attention, adding another heavy blow to its frequent adjustments in recent years. This semiconductor giant, once a leader in the industry, is now deeply mired in the "darkest moment" of transformation and struggling to move forward.

Meanwhile, on the other end of the tech world: As Apple marks the fifth anniversary of the launch of its self-developed PC chips, it announced that macOS 26 Tahoe will be the last macOS version to support Intel chips. This means that starting next year, the major new versions of Apple's desktop operating system will only run on Macs equipped with Apple Silicon chips.

It's not hard to tell that this not only marks the success of Apple's self-developed chips but also quietly accelerates the wave of change in the entire PC processor industry.

Apple's in-depth exploration and breakthroughs in the PC chip field are impacting the market landscape of PC processors centered around the x86 architecture with unprecedented force. This force is also inevitably shaking the foundation of Intel's x86 PC processor business, adding a new and powerful variable to its severe challenges.

The Glory and Dilemma of the x86 PC Dominator

Looking back at history, in the long journey of the development of the PC processor chip industry, a series of ups and downs in technological competitions have quietly taken place.

As early as 1971, Intel launched its first microprocessor, the 4004, and then continuously evolved. By 1978, it released the 16-bit processor, the 8086.

In 1981, the "Blue Giant" IBM chose the 8088, a derivative of the 8086, as the processor for its products. This was later called "Intel's greatest victory ever." This decision allowed the x86 architecture to catch the high-speed train of PC popularization.

In this process, with the rise of the Windows operating system, the x86 architecture, with the "Wintel Alliance" compatible with the Windows system as its weapon, swept aside a host of strong competitors such as Motorola 68K, DEC Alpha, IBM PowerPC, Sun SPARC, and MIPS. From the performance revolution of the Pentium series to the multi-core era of the Core architecture, Intel continuously pushed the x86 architecture to the extreme with its "Tick-Tock" strategy, almost monopolizing the global PC market share and creating Intel's hegemony in PC chips for more than four decades.

Meanwhile, within the x86 camp, AMD has always played the important roles of a "disruptor" and a "pursuer." In 1999, AMD launched the Athlon processor, surpassing Intel's products of the same level in performance for the first time. In 2003, AMD released the world's first 64-bit desktop processor, the Athlon 64, leading the PC into the 64-bit era.

In 2016, relying on the launch of the Zen architecture, AMD significantly improved the performance of its chips. With the help of TSMC's advanced process, the performance of AMD's PC processors quickly caught up with and even surpassed Intel's, driving AMD's market share in the PC processor market to rise continuously in recent years, showing a trend of keeping pace with Intel.

It can be seen that from the early stage of a hundred flowers blooming to the later oligopoly competition, the x86 camp represented by Intel and AMD, with its powerful performance, rich software support, and high compatibility, continuous technological innovation, and the construction of a huge ecosystem, finally stood out in the cruel competition in the PC processor market and ushered in the decades-long x86 era of PC processors.

However, the wheels of history are always turning. In recent years, with the surging wave of technology, the once indestructible x86 processors and the Intel Empire are now facing unprecedented challenges.

Apple's Self-Developed ARM PC Chips: From Experimentation to Disruption

Apple's sudden rise has created cracks in the once unbreakable "Wintel Camp."

First, the emergence of "macOS + x86" made "Windows + x86" no longer the only option. Then, in 2020, the release of Apple's M1 chip sparked a counterattack for the ARM architecture. Apple resolutely announced that its Mac series products would transition from Intel chips to its own ARM-based chips. With a unified memory architecture and software-hardware collaborative optimization, it began to break the performance myth of the x86 in the PC field.

A single stone can stir up a thousand ripples.

At that time, many people had doubts about the prospects of Apple's self-developed PC chips. However, just five years later, with continuous innovation and breakthroughs, Apple has not only successfully disrupted the market landscape of PC processor chips but also ushered in one of the most successful periods in Mac history.

Looking back at Apple's foray into the chip design field, as early as 2008 when it acquired P.A. Semiconductor, it quietly planted the seed for self-developed chips.

Since then, Apple has been expanding its territory in the mobile chip field. From the launch of the first-generation A4 chip to the subsequent upgrades brought by the A5 with a dual-core architecture, the revolutionary 64-bit architecture introduced by the A7, the integration of the M9 motion coprocessor in the A9, and the neural network engine in the A11 Bionic chip, each iteration has achieved a leap in performance and energy efficiency, building strong competitiveness for Apple in the smartphone and tablet markets.

After accumulating profound technological knowledge in the mobile chip field, Apple began to turn its attention to the PC market.

Actually, the Mac was initially equipped with the Motorola 68000 processor, which was used not only in the Mac but also in various video games and some Atari computers. However, in the early 1990s, Apple was frustrated with the slow improvement speed of its chip manufacturer and realized that the fate of its platform depended on the success or failure of other manufacturers.

At that time, Apple planned to develop its own chips. Unexpectedly, its arch-rival IBM suddenly reached out to discuss cooperation in designing the next-generation chips. After the almost abandoned Motorola joined, the AIM Alliance began to work on creating a new set of chips, which later became the PowerPC chips.

As the new-generation chips at that time, the PowerPC chips were very different from the then-dominant Intel processors. The first Mac computers equipped with PowerPC chips (naturally called Power Macs) were launched in March 1994. To achieve this, Apple not only needed to port its software to the new chip design but also had to ensure compatibility with the old Mac software.

However, Apple's biggest mistake at that time was not having control over its own development tools. Metrowerks, a software company eventually acquired by Motorola, created the authoritative PowerPC development environment, CodeWarrior. Apple learned an important lesson from this, and to this day, almost all of its development work is done with its own Xcode.

Time came to the summer of 2003, and the PowerPC era was in full swing. The new G5 (fifth-generation) processor had been released, and Steve Jobs promised that it would eventually reach a record speed of 3GHz. The industry was excited that Mac laptops would also have such powerful performance.

But the result was not as Apple had hoped.

IBM was never able to provide Apple with a 3GHz chip, and the G5 chip was not suitable for laptops. Meanwhile, inside Apple, a secret project was ensuring that the new Mac OS X could run on Intel processors.

Twenty years ago, Jobs announced this change at WWDC: the AIM project was terminated, and Apple would switch from PowerPC to Intel.

At that time, Intel processors were constantly being upgraded and iterated. The speed of the Mac also experienced unprecedented breakthroughs.

In many ways, this period was the most important decade in Mac history.

The increasing success of the iPod (and later the iPhone) gave consumers who had never considered buying a Mac a chance to own one. The new-generation Windows emulator could run at full speed on Intel hardware, providing an alternative for PC users who might need to run a small number of Windows programs. The Mac began to develop rapidly.

However, as Intel's technological innovation slowed down, Apple was once again frustrated with the slow speed of chip development and the lack of control over its own platform.

But this time, there was a key difference, as mentioned at the beginning of this section: Over the past decade, Apple had been designing chips for the iPhone and iPad, and the applications developed by developers with Xcode could be compiled and run on Apple's processors. Years of experience gave Apple the confidence to apply the experience of iPhone and iPad chips to creating powerful Mac chips.

Based on this, Apple decided to turn a new page again.

In November 2020, Apple's first milestone M1 chip based on the ARM architecture was launched, instantly becoming the focus of attention in the chip industry, especially in the PC CPU chip industry.

In various evaluations, the newly launched M1 chip's benchmark scores were comparable to those of high-end x86 processors, and its performance was close to that of the 11th-generation Core i7. It broke the shackles of the ARM architecture's low power consumption and low performance. While the M-series processors' performance was close to that of x86 processors, they still maintained low power consumption, about half that of Intel processors. As a result, Macs equipped with M-series processors could be designed to be thinner and lighter and have longer battery life. Once launched, they were very popular in the market, and the M1 Mac became the most highly rated Mac at that time.

This was the first time that ARM architecture processors had caught up with Intel in terms of performance, representing the victory of the ARM's low-power, high-performance licensing model.

Meanwhile, the launch of the M1 chip not only meant the end of Apple's 15-year cooperation with Intel but also allowed people to truly see the huge potential of the ARM architecture on the PC side, driving up the ARM penetration rate in the PC market.

Based on the success of the M1 chip, Apple didn't stop innovating.

In the following years, Apple launched the M2 and M3 series of processors one after another, helping to upgrade the entire Mac series. In this process, Intel was not only under pressure from both sides in the PC chip market but also saw its manufacturing hegemony gradually collapsing. For example, TSMC's 5nm process technology helped boost the performance of Apple's M2 chip, while Intel's 7nm process was delayed by a full three years. Due to the gap in process technology, the performance and power consumption of Apple's M-series chips improved rapidly.

At the same time, with the maturing of the Chiplet small-chip design and the smooth progress of TSMC's advanced packaging technology, the performance of Apple's PC processors continued to break through, making it the leader in the ARM architecture processor market.

In the past five years, many Apple fans have thought that the Mac has experienced a revival.

Thanks to the excellent "leading role" of Apple's M-series chips, the entire consumer electronics industry's view of ARM architecture PC products has changed significantly.

Since the birth of the M-series chips, Apple has changed users' perception of the performance of ARM chips and even restructured its business lines to fully support the launch of the M-series chips. So far, Apple's M-series chips represent the "upper limit" of the performance and ecosystem of consumer-grade ARM architecture PC solutions.

Not long ago, Apple announced that macOS 26 Tahoe will be the last Mac OS version to support Intel chips. This means that starting next year, the major new versions of Apple's desktop operating system will only run on Macs equipped with Apple Silicon chips (i.e., the 2020 M1 models and newer models).

According to relevant data statistics, before the release of the ARM-compatible M1 Mac, the market share of ARM in PC chips in the third quarter of 2020 was only 2%. More than a year later, according to the data from Mercury Research for the first quarter of 2022, the estimated share of ARM PC clients was 11.3%, almost doubling from 5.9% a year ago.

It can be seen that as Apple's Mac product line fully switches to ARM architecture chips, the market share of ARM chips, which once accounted for only a small part of the personal computer chip market, has increased rapidly.

Canalys data shows that in Q4 2024, Apple had a 10.2% market share in the entire PC market and a 45% market share in the AI-enabled PC market, far exceeding Intel (less than 10%). Although the overall market share of ARM architecture PCs is only 12%, Apple's success has shaken the foundation of the x86 camp.

As macOS Tahoe will be the last Mac OS version to support Intel Macs, Apple is preparing to end the third chip transformation in Mac history.

A Single Stone Stirs Up a Thousand Ripples: The ARM Camp Takes Action

The success of Apple's M-series processors has inspired other leading companies in the chip market, especially big players like Qualcomm, NVIDIA, and MediaTek, to enter the market.

Actually, in the field of ARM-based notebook chips, Qualcomm has more experience than Apple. Qualcomm launched the first and second generations of the Snapdragon 8CX processors for the PC in 2018 and 2020 respectively and has now developed them to the third generation. Unfortunately, these processors didn't make much of a splash in the market.

It wasn't until 2023 that Qualcomm finally shed the label of having weak PC chip performance. The Snapdragon X Elite PC processor it released had a huge leap in performance, with the Oryon CPU outperforming similar products from Intel, AMD, and Apple at that time, demonstrating Qualcomm's leading position in PC chips.

The Snapdragon X series, based on the ARM architecture, has become Qualcomm's main tool for expanding into the PC field.

In terms of commercialization progress, Qualcomm has deeply cooperated with global leading OEMs such as Microsoft, Lenovo, Dell, and HP to jointly create future-oriented AI PCs. Currently, more than 85 PC product designs based on the Snapdragon X series have entered the mass production or development stage, covering the entire price range from entry-level to flagship. It is expected that by 2026, this number will exceed 100, further accelerating the popularization of AI PCs.

Actually, Qualcomm's rapid progress in PC CPU performance is due to its acquisition of the chip design startup Nuvia for $1.4 billion in 2021. This company was founded by former Apple chip chief architects. After acquiring Nuvia, Qualcomm followed in the footsteps of Apple's M-series chips, completely abandoned the ARM public architecture, and developed a high-performance PC processor from scratch to challenge Intel and AMD.

It is reported that Qualcomm will launch a new generation of ARM architecture PC chips at the 2025 Snapdragon Summit, further breaking the performance limit and enhancing the market competitiveness of ARM chips in the PC field.

On the other hand, NVIDIA also chose to enter the PC CPU market.

Looking back at its development history, similar to Qualcomm, this is not NVIDIA's first foray into the PC processor market. Although NVIDIA's GPUs have long been the standard configuration for high-end gaming PCs, its previous attempts in the PC processor field have all ended in failure.

As early as 2011, NVIDIA participated in Microsoft's Windows on ARM project. In 2012, it provided the Tegra3 processor for the Surface RT. Due to problems such as poor x86 application compatibility and high prices, it suffered a major defeat. The subsequent Surface RT2 with the Tegra4 chip also failed to attract much attention in the market.

In 2014, NVIDIA announced the launch of "Project Denver," planning to develop an ARM architecture CPU to be combined with its iconic GPUs to provide processing power for PCs and other devices. However, this project ultimately came to nothing, and NVIDIA used some of the cores from Project Denver in its Tegra processors.

After several setbacks, NVIDIA temporarily withdrew from the consumer-grade PC market and turned to fields such as automobiles and game consoles.

This series of failures also made Microsoft put its ARM architecture notebook plan on hold for a while. Later, Microsoft commissioned Qualcomm to lead the porting of the Windows operating system to ARM architecture processors, and Qualcomm thus obtained an exclusive agreement to develop Windows-compatible chips. That's all in the past.