Around a decade ago, I’d have laughed at the idea of mobile processors being capable of running AAA games like Cyberpunk 2077, but that’s the reality. So much that was once unthinkable is now the norm, including ARM-based computers being a popular choice.
ARM-based computers aren’t exactly new, but it wasn’t until Apple’s push with the M-series chips that ARM became a serious competitor in mainstream PCs. Things have only gone uphill since then, to the point that an average PC user can consider buying a new ARM-based computer without worrying too much.
It’s no surprise: ARM is powerful, efficient, and cheaper to manufacture, making it a viable option for conventional PC use when hardware prices are through the roof. Now, ARM’s rise has sparked a heated debate in the PC community: many believe it will take over x86, but despite its growing popularity and efficiency, x86’s dominance won’t fade, even in the distant future.
Where ARM’s argument holds
Efficiency is where ARM wins
ARM’s instruction set architecture (ISA) uses simple, fixed-length instructions that a chip can decode more cheaply than x86’s. However, that isn’t where most of ARM’s efficiency comes from.
The biggest factor is the SoC design, which bundles several components and reduces motherboard complexity, latency, and space. The CPU and GPU generate less heat because of a lower TDP, so they need less cooling, making it easy to squeeze out up to 10 or 20 hours of battery life on a system without even a substantial battery.
So, ARM dominates x86 in one area: efficiency. This gives it a major advantage over x86 for simpler tasks, delivering higher performance per watt. That is why you’ll find several ARM-based options in the budget laptop range that deliver the best performance per watt.
ARM chips took a long time to gain traction in the computer market despite being around for decades because of performance concerns.
Only in recent years have they gained more power through industry-wide developments, particularly Apple’s work with the M-series chips, which has significantly improved ARM chips’ performance to the point that they’re no longer confined to phones and tablets and can handle desktop-level workloads as well.
The desktop ceiling still belongs to x86
Discrete GPUs and swappable parts sit outside ARM’s approach
Sure, Apple has done the impossible with the M-series chips in terms of power, as shown by the latest M5 Pro 18-core, which is around 26% faster in single-thread testing than the Intel Core Ultra 9 275HX (CPU Benchmarks), but that’s for the laptop market.
Even though ARM chips are able to scale, they still can’t compete directly with x86 in desktop environments because no ARM chips can match GPU performance, given the integration design and desktop-level GPU customization, since the SoC design eliminated modularity.
Consumer ARM chips generally target low-power devices.
Even if Apple’s M-series and Nvidia’s upcoming RTX Spark chips can handle heavy workloads and demanding tasks, they’re super expensive, and you can get a desktop PC with similar performance for less, with the freedom to choose different parts and upgrade easily.
The same SoC design and efficiency also mean they can’t sustain heavy workloads for CPU-intensive tasks because of lower power limits and lower clock speeds than x86.
For laptops, that is fine, but desktop PC users are generally reluctant to invest heavily in an SoC for a desktop PC that offers upgradability, since they’d have to replace the entire PC.
Decades of software aren’t getting rewritten
Gaming on ARM also isn’t just there yet
x86 has existed for decades, and we don’t see more CPUs built on the same architecture because AMD and Intel hold a duopoly for the x86 instruction set. ARM chips have only recently entered PCs, and most software over the past few decades has been written explicitly for x86.
Currently, the ARM market isn’t big enough for all developers to rewrite software for a different instruction set. Although Microsoft’s “Prism” emulation layer translates native x86 .exe files for ARM, the emulation overhead becomes noticeable in higher-end software, where you might see occasional hitches and performance drops on lighter machines.
Gaming also hits the same emulation wall, since most titles aren’t ARM-native and emulating AAA games comes with considerable overhead. Also, not all integrated GPUs in these SoCs are designed with gaming performance in mind (aside from a few instances), so they may lack key optimizations.
Also, ARM chips don’t have the higher clock speeds and extra cores needed for gaming and other heavy-load tasks, so they aren’t even feasible for gaming or similar heavy software, at least on mid-range chips. Some low-level features and anti-cheat software are also a mixed bag on ARM.
In the near future, anti-cheat support is expected to be broader for ARM PCs, but that’s not guaranteed across all games. But for the most part, compatibility is still an issue, and given how niche the ARM market is right now, few developers will prioritize ARM-native builds while the install base stays small.
Although I do see a future for ARM chips in the handheld market, which is currently booming, the tighter SoC design will make for great gaming handhelds, offering a great leap in battery life and performance like the Nintendo Switch 2, but that’s a future for PC handhelds.
ARM is a new category, not a takeover
Macs and other ARM-based computers were positioned as an alternative to x86. Apple didn’t drop Intel because ARM was better, but so it could move to a vertical design across the entire architecture, since iPhones already use ARM.
Don’t get me wrong: ARM works well for daily, average tasks and suits office and learning environments, but that doesn’t mean x86 is going away. If anything, ARM still struggles to compete at the high-end, where x86’s dominance will continue. I’d call ARM PCs a new subcategory of computers viable only for mobile laptops.
Much like Linux never displaced Windows despite predictions years ago, x86 will remain the leading PC architecture for years to come.