ARM Isn't as New as You May Think
ARM feels like the newest thing in PCs. Its history is older, stranger, and much more intertwined with today's competitors than you might expect.
In 1993, Apple shipped a battery-powered personal computer built around an ARM processor.
No, not a MacBook.
The Newton.
And here's the stranger part:
That wasn't even ARM's first personal computer.
If ARM seems like the newest challenger to Intel and AMD in the PC market, its actual history can be a little disorienting.
Apple helped establish the company behind ARM.
Intel made ARM processors.
AMD made ARM processors.
NVIDIA powered Windows ARM computers long before its latest attempt.
Microsoft has been trying to make Windows and ARM work together for years.
Some of those products succeeded.
Some failed spectacularly.
Some disappeared quietly.
And while all of that was happening, ARM became one of the most widely used processor architectures in the world.
So perhaps the interesting question isn't:
Why does ARM keep coming back?
Maybe it's:
Did it ever really leave?
Before ARM was ARM
The story starts at Acorn Computers in the United Kingdom.
In the early 1980s, Acorn wanted a new processor architecture for its computers. Sophie Wilson and Steve Furber led the work that became the , or ARM.
The first ARM silicon arrived in 1985.[1]
Two years later, Acorn introduced the Archimedes, the first family of personal computers built around ARM processors.[1]
So ARM wasn't originally a smartphone architecture.
Smartphones as we know them didn't exist.
It was a computer processor.
The characteristics that would eventually make ARM enormously attractive for mobile devices - useful performance with remarkably modest power requirements - were already part of its appeal.
But Acorn was a relatively small computer company.
The next chapter would give ARM a much larger future.
And Apple helped write it.
Yes, Apple
By the late 1980s, Apple was developing something unusual.
The Newton was intended to be a highly portable, battery-powered personal computing device. That created an obvious processor problem.
Performance mattered.
So did power.
Apple became interested in Acorn's ARM technology, and in 1990 Apple, Acorn, and VLSI Technology formed a new independent company:
Advanced RISC Machines Ltd.
ARM.[2]
Apple wasn't the inventor of ARM. That credit belongs to the team at Acorn.
But Apple was there extraordinarily early, helping turn Acorn's processor project into the independent ARM company whose technology would eventually spread across the computing industry.
The first Newton MessagePad arrived in 1993 using an ARM610 processor.[3]
The Newton itself never became the revolution Apple hoped it would be.
ARM had a rather different future.
ARM found a much bigger home
ARM's combination of performance, low power consumption, and licensable intellectual property turned out to be extremely useful as computing escaped the desk.
Phones.
PDAs.
Embedded systems.
Networking equipment.
Consumer electronics.
Eventually smartphones and tablets.
This is an important distinction in the history because ARM isn't a processor manufacturer in quite the same way Intel traditionally has been.
That's how Apple, Qualcomm, MediaTek, NVIDIA, and many others can all produce very different processors built around ARM technology.
And historically, that list included two names that may be more surprising.
Intel.
And AMD.
Intel used to make ARM processors
Yes.
That Intel.
In the late 1990s, Intel acquired technology and assets associated with Digital Equipment Corporation's StrongARM processors.
Intel developed that technology into XScale, an ARM-based processor architecture used in mobile and embedded devices.
The pitch sounds remarkably contemporary.
Intel itself describes XScale as designed to balance performance with the battery-life requirements of phones and PDAs.[4]
Then, in 2006, Intel sold much of that business to Marvell for $600 million.
Why?
Intel's own historical account says the company wanted to focus on x86 as the future of mobile, network-connected computing.[4]
It's tempting to look backward from today's smartphones, Macs, and Windows ARM PCs and laugh.
That wouldn't be particularly fair.
Companies make investment decisions based on the markets, technologies, opportunities, and information in front of them at the time. Intel continued using and licensing ARM technology in other areas, and x86 remained enormously successful.
But the historical irony is difficult to miss.
Intel didn't merely watch ARM become a competitor.
Intel once built ARM processors - and then sold that business while betting on x86 for mobile computing.
History has a sense of humor.
AMD wanted to be ambidextrous
AMD's ARM history may be even more surprising.
In 2014, AMD announced a strategy it literally called:
Ambidextrous Computing.
The idea was that AMD didn't necessarily need to choose between x86 and ARM.
Its Project SkyBridge initiative envisioned infrastructure that could support pin-compatible ARM and x86 processors, while AMD also began developing K12, its own custom high-performance ARM core.[5]
AMD later shipped the ARM-based Opteron A1100 server processor.[6]
This wasn't a single processor magically switching between ARM and x86. The architectures remained distinct.
The interesting idea was strategic:
Why shouldn't one processor company participate in both ecosystems?
AMD ultimately concentrated its resources elsewhere.
Zen happened.
That went rather well.
K12 never became the major product family AMD once envisioned, and AMD's ARM ambitions faded from the center of its strategy.
But they didn't disappear from history.
So when we look at today's PC market as a contest between ARM on one side and Intel/AMD on the other, it's worth remembering:
The teams haven't always been arranged that way.
Microsoft has been here before too
If today's Snapdragon PCs feel like Microsoft's first serious attempt at Windows on ARM, Windows RT would like a word.
Microsoft launched in 2012 alongside Windows 8.
It ran on ARM processors from multiple vendors, including NVIDIA.
Microsoft's own Surface RT used an NVIDIA Tegra processor.[7]
That's worth remembering now that NVIDIA is entering the Windows ARM market again.
Even processor competition isn't entirely new.
The problem was that Windows RT looked a lot more like conventional Windows than it behaved like conventional Windows.
It couldn't run the enormous library of existing x86 Windows desktop applications users expected Windows computers to run.
That distinction proved difficult to explain and harder to sell.
Windows RT faded.
But Microsoft didn't abandon the problem.
It changed the approach.
What if Windows came to your software?
In 2016, Microsoft and Qualcomm announced that full Windows 10 would run on Snapdragon processors.
This time, existing Win32 applications could .[8]
That was a fundamentally different proposition.
Instead of asking the Windows ecosystem to start over, Microsoft increasingly tried to bring the existing ecosystem along.
The first wave of Snapdragon-powered Always Connected PCs followed, with manufacturers including ASUS, HP, and Lenovo.[9]
And listen to the pitch.
Thin.
Quiet.
Fanless.
Always connected.
Long battery life.
Sound familiar?
Many of the ideas associated with today's Snapdragon X laptops were already there.
The hardware and software just weren't ready to deliver the proposition nearly as convincingly.
The compatibility boundary kept moving
This may be the most important part of Microsoft's ARM history.
Windows on ARM hasn't simply failed, disappeared, and restarted from scratch every few years.
The compatibility boundary has kept moving.
Windows RT largely required software built specifically for the platform.
Windows 10 on ARM introduced emulation for existing x86 applications.
Later Windows versions expanded x64 application support.
Microsoft introduced , allowing developers to combine native ARM64 and existing x64 components within the same application while transitioning software.[10]
Windows 11 improved ARM development support.
improved x86/x64 emulation again.
Microsoft later expanded Prism with AVX, AVX2, and other instruction-set support, allowing existing ARM PCs to run additional applications and games.[11]
Native software support expanded at the same time.
None of those steps individually solved every compatibility problem.
But they accumulated.
Today's Windows-on-ARM ecosystem didn't appear fully formed when Snapdragon X launched.
It inherited years of work from the attempts that came before it.
Then Apple brought ARM to the Mac
While Microsoft and Qualcomm were trying to make ARM work convincingly inside the sprawling Windows ecosystem, something very different happened across the aisle.
Apple began moving the Mac away from Intel processors.
In June 2020, Apple announced that the Mac would transition to its own Apple Silicon, creating what Apple described as a common architecture across its products.[12]
The first M1 Macs arrived later that year.[13]
The transition mattered far beyond Apple.
ARM no longer needed to prove that an ARM-based processor could power a premium mainstream laptop or desktop.
Apple was doing it.
At scale.
This doesn't mean Apple solved Windows on ARM.
Apple controls its operating system, much of its development ecosystem, its hardware designs, and the transition itself. Microsoft has to support an extraordinarily diverse Windows world containing decades of applications, peripherals, drivers, hardware manufacturers, corporate requirements, and user expectations.
Those are very different problems.
But Apple changed the surrounding context.
The question was no longer:
Can ARM power a serious personal computer?
Clearly it could.
The harder question became:
Can Windows make ARM ordinary enough that most users stop caring?
The world around the PC changed too
There's another difference between today's ARM PCs and the machines Microsoft was trying to sell in earlier eras.
The computer itself isn't always where the computing happens anymore.
Cloud storage is ordinary.
Software-as-a-service is ordinary.
Browser applications are ordinary.
Streaming is ordinary.
Remote desktops are ordinary.
Cloud development environments and hosted computing are increasingly ordinary.
Even many applications we think of as local experiences spend much of their time talking to services somewhere else.
That changes the importance of the processor architecture underneath the keyboard.
If you use Salesforce through a browser, the processor running Salesforce's servers isn't your compatibility problem.
The same is true for enormous categories of web applications, streaming services, cloud storage, and remotely hosted workloads.
This doesn't make local compatibility irrelevant.
If your job requires a particular x64 application, kernel driver, peripheral, GPU technology, virtual machine, or locally installed professional tool, the cloud doesn't magically fix that. That's exactly why ARM isn't always the answer.
But compared with 2012 - and certainly compared with 1993 - far more of an ordinary user's computing life can happen independently of the .
And ARM grew up in exactly that world.
Sometimes the cloud is ARM too
There's another fun wrinkle.
ARM didn't stop at phones and laptops.
It also moved into the datacenter.
Cloud providers now offer substantial ARM-based computing infrastructure, including Amazon's Graviton processors.[14]
So an ARM-powered Windows laptop might someday spend much of its day displaying a browser connected to software running on another ARM processor in a datacenter.
The user may have absolutely no idea.
Which may ultimately be the point.
Architectures matter enormously to engineers.
Most users would rather they didn't have to care.
Snapdragon X didn't start the story
Qualcomm's Snapdragon X generation deserves much of the attention it has received.
It made the Windows ARM proposition dramatically more convincing.
Performance became competitive.
Battery life became a headline feature rather than an apology for reduced performance.
PC manufacturers launched a broad range of machines.
Microsoft made ARM central to the first wave of Copilot+ PCs.
But Snapdragon X didn't suddenly invent Windows on ARM.
It arrived after decades of processor development, several generations of Microsoft compatibility work, enormous ARM growth outside the PC, Apple's highly visible architecture transition, increasingly mature ARM development tools, and a computing world increasingly comfortable with cloud services.
That's a very different starting line from Windows RT.
And now NVIDIA is back
In 2026, NVIDIA and Microsoft announced RTX Spark, a new Windows-on-ARM platform developed with MediaTek.[15]
NVIDIA is bringing Blackwell RTX graphics, CUDA, AI acceleration, and its enormous graphics ecosystem into ARM-powered Windows systems aimed at creators, developers, AI workloads, and gamers.
Qualcomm suddenly has serious company.
But NVIDIA isn't technically new to this race.
Remember Surface RT?
NVIDIA was there.
More than a decade later, it's returning to a Windows ARM ecosystem that looks dramatically different from the one it left.
That's perhaps more interesting than simply saying another chip company entered the market.
So why does ARM keep coming back?
Maybe that's the wrong question.
Acorn built ARM computers in the 1980s.
Apple helped establish ARM as an independent company and used it in Newton.
ARM became enormously successful in mobile and embedded computing.
Intel built ARM processors.
AMD built ARM processors.
Microsoft shipped ARM Windows machines.
NVIDIA powered them.
Apple eventually moved the Mac to its own ARM-based silicon.
Qualcomm kept pushing Windows ARM forward.
ARM expanded into cloud infrastructure.
Microsoft kept improving compatibility.
And now Qualcomm is expanding while NVIDIA and MediaTek enter the current generation of Windows ARM competition.
The common thread isn't that ARM repeatedly failed and somehow returned.
It's that different parts of the computing industry kept finding reasons to use it.
Power efficiency was one reason.
Mobility was another.
Licensability mattered.
Customization mattered.
Scale mattered.
Different companies found different answers at different times.
Some bets worked.
Others didn't.
Could Intel or AMD come back too?
Now we're firmly in speculation territory.
To be clear:
At publication time, we haven't identified confirmed evidence that Intel or AMD is preparing an ARM-powered Windows PC processor.
But history makes the question less ridiculous than it initially sounds.
Both companies have already participated in ARM.
Intel built XScale.
AMD built Opteron A1100 and once explicitly planned an "ambidextrous" future spanning ARM and x86.
If Windows ARM eventually becomes a large enough market, would either company decide it wants to participate again?
We don't know.
There are also much more speculative possibilities.
Modern processors already mix different kinds of CPU cores with GPUs, NPUs, media engines, security processors, and other specialized hardware.
Could some future system blur architectural boundaries even further?
Perhaps.
But combining ARM and x86 application-processing cores would be vastly more complicated than today's performance-core/efficiency-core designs. Operating systems, schedulers, drivers, applications, memory behavior, debugging, and security would all become part of the problem.
Interestingly, Microsoft is already solving a related, much less exotic interoperability problem in software. ARM64EC allows ARM-native and x64 code to coexist within an application.[10]
Sometimes the easiest way to combine two worlds isn't putting both processors on the same chip.
History can't tell us whether Windows on ARM will succeed this time, and it should probably make us suspicious of anyone who claims the answer is obvious.
What history does show is that today's attempt didn't begin from the same starting line. ARM is vastly more established, Apple has demonstrated it at mainstream PC scale, Windows emulation and native development are dramatically more mature, cloud-connected computing has reduced some dependence on local processor architecture, and additional hardware and software companies are investing in the ecosystem.
None of that guarantees success.
It does make today's Windows ARM effort another chapter in a much longer story - not ARM's sudden return from nowhere.
Maybe ARM never came back
There's an easy story to tell about today's PC market.
ARM disappeared.
Intel and AMD ruled the PC.
Apple brought ARM back.
Qualcomm brought it to Windows.
Now NVIDIA is joining the race.
It's neat.
It's also not really what happened.
ARM spent the intervening decades becoming enormously important while the traditional PC happened mostly somewhere else.
It was in the phones in our pockets.
The tablets on our couches.
Networking equipment.
Embedded devices.
Increasingly, servers in datacenters.
And throughout that history, many of the companies we now describe as ARM's competitors were occasionally participants themselves.
So maybe ARM isn't having another comeback.
Maybe Windows PCs are simply having another encounter with an architecture that never stopped evolving around them.
Whether this attempt finally makes ARM an ordinary part of the Windows PC market remains to be seen.
But ARM itself?
ARM isn't as new as you may think.
And it never really went away.
Sources & Further Reading
[1] Arm - The Official History of Arm. Arm's history of the architecture's origins at Acorn, the first working ARM silicon in 1985, and the Acorn Archimedes personal computers.
https://newsroom.arm.com/blog/arm-official-history
[2] Arm - Celebrating 40 Years of the Arm Architecture. Historical account of Acorn's ARM work and the creation of Advanced RISC Machines as a joint venture involving Acorn, Apple, and VLSI Technology.
https://newsroom.arm.com/blog/40-year-anniversary-of-arm-architecture
[3] Arm - 10 Devices That Define Arm. Arm's historical account identifies Apple's Newton MessagePad as using the ARM610 processor. https://newsroom.arm.com/blog/10-devices-define-arm
[4] Intel - XScale. Intel's corporate history describing its ARM-based XScale architecture, mobile/embedded goals, and the 2006 sale of the business to Marvell as Intel focused on x86 for mobile network-connected computing.
https://timeline.intel.com/2000/xscale
[5] AMD - AMD Unveils Ambidextrous Computing Roadmap. AMD's 2014 primary-source announcement of Project SkyBridge, planned pin-compatible ARM and x86 processors, and the K12 custom high-performance ARM core. https://ir.amd.com/news-events/press-releases/detail/524/amd-unveils-ambidextrous-computing-roadmap
[6] AMD - Opteron A1100. AMD's announcement and documentation of its ARM-based Opteron A1100 server processor.
https://www.amd.com/en/newsroom/press-releases/2016-1-14-amd-and-key-industry-partners-welcome-the-amd-opte.html
[7] Microsoft Support - Surface RT specifications. Microsoft's specifications for the original Surface RT identify its quad-core NVIDIA Tegra 3 processor. https://support.microsoft.com/en-US/surface/models/surface-rt-specifications
[8] Qualcomm - Windows 10 Powered by Snapdragon. 2016 Microsoft/Qualcomm announcement of full Windows 10 on Snapdragon with support for existing Win32 applications through emulation.
https://www.qualcomm.com/news/onq/2016/12/windows-10-powered-snapdragon
[9] Qualcomm - ASUS, HP, and Lenovo to Build Windows 10 PCs Powered by Snapdragon. 2017 announcement of Snapdragon 835 Always Connected PCs and their thin, fanless, connected, long-battery-life positioning.
https://www.qualcomm.com/news/releases/2017/05/qualcomm-announces-flagship-companies-build-windows-10-pcs-powered
[10] Microsoft Learn - ARM64EC for Windows 11 apps on Arm. Microsoft's hybrid application ABI allowing native ARM64 and x64 components to coexist as developers transition applications.
https://learn.microsoft.com/en-us/windows/arm/arm64ec
[11] Microsoft - Windows on Arm Runs More Apps and Games with New Prism Update. Documents Prism's later addition of AVX, AVX2, BMI, FMA, F16C, and related x86 instruction emulation.
https://techcommunity.microsoft.com/blog/windowsosplatform/windows-on-arm-runs-more-apps-and-games-with-new-prism-update/4475631
[12] Apple - Apple Announces Mac Transition to Apple Silicon. Apple's June 2020 announcement of its transition away from Intel and toward a common architecture across Apple products.
https://www.apple.com/newsroom/2020/06/apple-announces-mac-transition-to-apple-silicon/
[13] Apple - Apple Unleashes M1. Apple's November 2020 announcement of its first M1 processor and first generation of Apple Silicon Macs.
https://www.apple.com/newsroom/2020/11/apple-unleashes-m1/
[14] Amazon Web Services - AWS Graviton Processors. ARM-based processors designed by AWS for cloud workloads, illustrating ARM's expansion into datacenter computing.
https://aws.amazon.com/ec2/graviton/
[15] NVIDIA - NVIDIA and Microsoft Open New Frontier for Windows PCs and Agents With RTX Spark. Official 2026 announcement of NVIDIA's new Windows-on-ARM platform developed with MediaTek.
https://nvidianews.nvidia.com/news/nvidia-microsoft-windows-pcs-agents-rtx-spark