When Qualcomm's Snapdragon X laptops arrived under the Copilot+ banner, the pitch was simple: phone-class efficiency, laptop-class performance, all-day battery. The skepticism was equally simple: Windows on Arm had failed twice before on app compatibility.
Two years on, both the promise and the skepticism were partly right — and the most important outcome is the one nobody predicted: x86 got dramatically better in response.
Why it matters
The Windows laptop is a market of hundreds of millions of units a year, and for two decades it ran on one architecture. A credible second architecture changes supplier leverage, pricing and design priorities across the whole industry — even for buyers who never purchase an Arm machine.
It also matters as a template: if Arm can take share in Windows laptops, the same playbook applies to desktops, servers and eventually the AI PC's NPU-centric future, where the instruction set matters less than the accelerator stack.
How it works
Modern Windows on Arm runs three kinds of software: native Arm64 applications, which perform well; x86-64 applications through the Prism emulation layer, which now handles most productivity software acceptably; and the remaining problem set — kernel-level drivers, anti-cheat systems, some VPN and security agents, and niche peripherals — which still requires native support and still blocks some buyers.
The hardware story is efficiency architecture: big-little core arrangements, aggressive power gating and integrated platforms where the memory is soldered close to the die. The trade is upgradeability and, at the low end, price.
Evidence
Market trackers put Windows on Arm at a meaningful minority of new Windows laptop sales, concentrated in the premium thin-and-light segment — real share, not a rounding error, but not the displacement early forecasts implied. Native Arm64 versions now exist for the major browsers, Office, the Adobe core apps and most mainstream developer tools, and Microsoft's Prism emulator benchmarks show the compatibility-tax shrinking generation over generation.
The counterfactual evidence is Intel's and AMD's latest mobile parts, which matched Arm laptops' battery life within a generation — a response that would not have existed without the competitive pressure.
The competing read
Arm advocates argue the architecture's advantage compounds: efficiency headroom grows with each node, and the NPU-centric future favors the phone-derived design philosophy. The x86 camp counters that the gap has already closed enough to neutralize the pitch, and that the remaining compatibility friction is structural — enterprise driver ecosystems move in years, not quarters. The market data so far supports a third reading: Arm wins specific segments, x86 keeps the default, and the buyer benefits either way.
What happens next
Watch Qualcomm's post-exclusivity competition — MediaTek and others entering Windows Arm silicon — and whether Microsoft commits to Arm-native versions of its remaining first-party stack. The enterprise fleet decision cycle, which runs three to five years, is where the architecture question gets answered for real.