In two separate robot-arm-controlled laptop comparisons published last week, the YouTube channel PhoneBuff found that a 500-page Microsoft Word document opens in 7 seconds on Apple laptops and 12 seconds on their Windows rivals — a result that held across two completely different chip architectures and appeared to indict the Windows software stack rather than any specific processor. The finding matters not because Apple Silicon is fast — that is well established — but because Word is Microsoft's own flagship productivity application, and the machine that opened it faster was $300 cheaper. This was reported by Windows Latest.

Word, Windows, and a Robot Arm

PhoneBuff's testing uses robot-controlled laptop testing methodology, the same approach the channel has applied to smartphones for years and has now extended to laptops. A robotic arm performs the same physical interactions on both machines simultaneously: the same keystrokes, the same file, the same scripted sequence. Human variability is eliminated.

In the first comparison, both laptops opened an identical 500-page Word document from a cold start. The MacBook Air M5 — running Apple's latest Apple Silicon chip — finished in 7 seconds. The Surface Laptop — running Qualcomm's Snapdragon X2 Plus, the latest ARM-based Windows chip, with 16GB of RAM — needed 12 seconds to open a Microsoft application on a Microsoft operating system. Nearly twice as long.

Then PhoneBuff ran a parallel comparison between Apple's MacBook Neo and Dell's XPS 13. The MacBook Neo does not run conventional Apple Silicon — it uses the A18 Pro, a chip repurposed from the iPhone 16 Pro lineup, in a laptop priced at $699. The Dell XPS 13, in the entry-level configuration Dell ships to compete on price, runs Intel's Core 5 320 — part of the new Wildcat Lake family — paired with 8GB of RAM. The result from the Neo vs XPS 13 test: the MacBook Neo opened Word in 7 seconds; the XPS 13 needed 12 seconds. The same gap. Different chips, different manufacturers, different price tiers, different operating system configuration — the same 5-second penalty for being Windows.

File extraction was even more pronounced: the Surface Laptop required over 40 seconds for extraction on the same archive task that the MacBook Air completed in under 10 seconds. Web browsing produced roughly equal results on both platforms, which itself tells the story — the gap is specific to workloads that interact with the Windows application launch stack, not to raw compute.

Why Windows, Not the Chip, Is the Problem

The instinctive explanation for any performance gap between Apple Silicon and Windows hardware is that Apple's chips are simply faster. This test complicates that explanation considerably.

The MacBook Neo's A18 Pro is a mobile phone chip. It was not designed to be a laptop processor. It has a 6-core CPU configuration and runs on a 3nm process, but it is architecturally and thermally constrained in ways that dedicated laptop silicon is not. The Dell XPS 13's Wildcat Lake Core 5 320 is a purpose-built laptop chip. The Surface Laptop's Snapdragon X2 Plus, with its Oryon Gen 2 CPU cores running at up to 4.0GHz, has been independently benchmarked as much faster than the prior Surface Laptop's Snapdragon X Elite. None of that processing capability translated into opening Word faster.

The technical explanation has been confirmed, somewhat awkwardly for Microsoft, by Microsoft's own executive vice president for Windows and Devices. On July 31, 2026, Pavan Davuluri announced that replacing WebView2 with native Windows code is now the company's top engineering priority.

Since 2021, Microsoft built large portions of Windows 11's own shell — Widgets, Windows Search, Settings pages — on top of WebView2, a component that wraps the Chromium rendering engine, the same engine that powers Chrome and Edge. As explained in detail in a prior TechTimes report on WebView2, every one of those OS components runs the way a browser tab runs: it spawns a renderer process, a GPU process, and a network process in the background, consuming memory and CPU cycles before the user opens a single application. On 8GB Windows 11 machines, independent testing has documented up to 70% idle RAM usage before any app is launched. That is the environment into which Word attempts to open.

Apple's macOS uses its own native rendering frameworks throughout its shell. There is no embedded Chromium instance in Finder, System Preferences, or Spotlight. The comparison is not between two chips; it is between an operating system with a browser engine inside it and one without.

Davuluri's July 31 announcement named three specific technical levers: a system-wide memory allocator overhaul, a migration of Windows shell components from WebView2 to the native WinUI 3 framework, and efficiency improvements to the Chromium components that cannot yet be replaced. The migration is not complete. Windows Search, one of the most frequently active OS components, still runs on WebView2 despite a visual redesign completed this past summer. General availability of the improvements is expected to begin this fall, continuing through the end of 2026.

Why Does Windows Win on Battery?

The battery results reversed the expected outcome so sharply that Windows Latest's own author noted surprise in his coverage.

In PhoneBuff's two-hour productivity battery test — a scripted cycle alternating between Word and Excel, with continuous typing, copy-and-paste work, and Spotify playback in the background — the Surface Laptop drained just 16 percent of its battery against 24% for the MacBook Air M5. The XPS 13 drained only 18 percent of its charge versus 26% for the MacBook Neo.

Both results run counter to Apple Silicon's dominant reputation for energy efficiency, and both find directional support elsewhere. Tom's Guide's separate continuous web-browsing battery test found the 2026 Dell XPS 13 outlasting MacBook Neo on battery — 14 hours and 29 minutes against 13 hours and 26 minutes.

Two technical factors explain the reversal. First, Intel's Wildcat Lake architecture — the Panther Lake-derived chip family from which the Core 5 320 is drawn — has made genuine gains in light-load efficiency that bring it far closer to ARM platforms in typical office workloads than any prior Intel generation. Second, the productivity test cycles specifically through the kinds of light, intermittent workloads — short typing bursts, copy-paste operations, background audio — where ARM chips are most battery-efficient, and where the MacBook Neo's fanless thermal design delivers modest performance but tight power envelopes. The Adobe Premiere Pro export segment, which represents sustained heavy compute, narrowed both gaps in the battery charts, suggesting Windows' battery advantage is specific to light productivity use rather than all workloads.

This is a meaningful finding for buyers who keep their laptops off a charger during workdays. In the specific type of use the battery test models — the ordinary rhythm of document work, spreadsheets, and background music — both Windows laptops showed measurably better endurance than their Apple counterparts.

Which Windows Laptop Is Actually Worth Buying?

The two Windows laptops in these tests are not in the same position.

The Dell XPS 13 is a competitive product at its current $699 price point. Its 8GB RAM configuration is a real limitation for power users, but the machine adds a touchscreen and 120Hz display the MacBook Neo does not have, expandable storage, and USB4 connectivity. The 5-second Word gap is a genuine weakness, but it is one weakness against an otherwise well-rounded machine that now demonstrably outlasts its Apple competitor on the battery life metric most buyers will actually experience day-to-day. At equal price and with comparable build quality, the XPS 13 has a reasonable case.

The Surface Laptop is a harder sell. It starts at $1,599 for the 13.8-inch model — approximately $300 more than MacBook Air at $1,299. The Surface Laptop gets a 120Hz display and touch support that the MacBook Air lacks, and its 3:2 screen ratio shows more vertical content per inch for productivity work. But it loses to the MacBook Air on app launch speed, on file operations, and on price. Microsoft's competing claim — that its Surface Laptop camera now holds the top spot in DXOMARK's laptop camera ranking, with a score of 146 against the MacBook Pro M5's 135 in fourth place — is legitimate and independently verified. Camera quality is a real differentiator. Whether it justifies a $300 premium over a machine that opens your documents faster is a decision each buyer must make.

European markets face a similar pricing dynamic: the gap between the Surface Laptop and a comparable MacBook Air M5 runs approximately €300 (about $345 at August 3, 2026 exchange rates) in most markets.

Microsoft has publicly acknowledged the software overhead problem and committed to resolving it. The WebView2-to-WinUI 3 migration and memory allocator overhaul named in Davuluri's July 31 statement are the right technical interventions — they target the exact source of the overhead that the PhoneBuff results reveal. The question is timing. General availability of the first improvements is not expected before fall 2026, meaning buyers purchasing a Surface Laptop today are paying $300 more than a MacBook Air M5, losing on app speed and file operations, winning on battery endurance and display refresh rate, and waiting for a software fix that does not yet have a published delivery date or a confirmed improvement target.

What Comes Next for Windows on This Problem?

Microsoft is not standing still. The WinUI 3 migration for Windows shell components is actively in progress: as confirmed in recent Windows shell redesign reporting, the File Explorer Properties dialog has already been rebuilt in WinUI 3, and the Widgets panel and a new Run experience are committed for migration. WinUI 3 renders natively through the Windows GPU compositor using XAML APIs, without spawning any browser subprocesses, so each migrated component directly removes a source of the overhead the PhoneBuff results measured.

NVIDIA's RTX Spark-powered Windows laptops are also expected in the second half of 2026, introducing an entirely new GPU-accelerated ARM architecture — one with dedicated CUDA cores and unified memory — into the premium Windows laptop segment. Details on the RTX Spark platform's pre-production performance appear in a prior TechTimes Surface Laptop Ultra report. The Lenovo Yoga 9n 2-in-1 is already confirmed as the first RTX Spark convertible. Whether RTX Spark devices will close the app-launch gap, or whether that gap will persist until the WinUI 3 migration is complete, depends on the timing of both the hardware launches and the software fix.

For buyers making a decision this week: the benchmark data says the MacBook Air M5 opens your documents faster and costs less. The Surface Laptop lasts longer off a charger in everyday productivity use and has a better camera. Both things are true, and the right choice depends on which of those outcomes you experience more often.

Exchange rate as of August 3, 2026; conversions are approximate.

---

Frequently Asked Questions

Why does Microsoft Word open faster on a MacBook than on a Windows laptop?

The cause is not processor speed — it is operating system architecture. Windows 11 builds portions of its own shell, including Widgets, Windows Search, and certain Settings panels, on WebView2: a component that embeds Chromium in the OS. Each active WebView2 component spawns background processes identical to open browser tabs, consuming memory and CPU resources before any user application launches. Apple's macOS does not use an embedded browser engine in its shell — its OS components render through native Apple frameworks, leaving more headroom available when an application like Word needs to open. Microsoft's EVP for Windows confirmed this architectural problem on July 31, 2026 and named replacing WebView2 with native WinUI 3 code as the company's top engineering priority, with improvements expected beginning in fall 2026, as detailed in Davuluri's July 31 blog post.

If Windows laptops lose on app speed, why do they win on battery life in this test?

PhoneBuff's productivity battery test cycles through light, intermittent workloads: typing in Word, switching to Excel, copy-pasting, with Spotify playing in the background. This is the specific category of workload where ARM chips — including Qualcomm's Snapdragon X2 Plus and Intel's new Wildcat Lake architecture — compete most favorably on efficiency, as confirmed in PhoneBuff's productivity battery results. Both platforms converge when workloads go heavy: during the Adobe Premiere Pro export segments in PhoneBuff's test, the battery gap narrowed. Apple Silicon's efficiency advantage reasserts itself under sustained compute load. For buyers who spend most of their day in documents, spreadsheets, and browser tabs rather than video export, Windows' endurance lead in this test is a real and meaningful finding.

Is the Surface Laptop worth $300 more than the MacBook Air M5?

Based on available benchmark data: for pure app-launch performance and price, the MacBook Air M5 has a clear advantage. For buyers who specifically value a 120Hz touchscreen display, the 3:2 screen ratio, the top-ranked laptop camera (per DXOMARK's laptop camera ranking), or the battery endurance edge in light productivity use, the Surface Laptop offers distinguishing features the MacBook Air does not. The $300 premium does not currently purchase faster software performance — it purchases display features and endurance, with the understanding that Microsoft has publicly committed to improving the software speed gap by the end of 2026.

How reliable is PhoneBuff's robot-arm testing methodology?

PhoneBuff developed its robot-arm testing methodology for smartphone drop and battery testing, where mechanical precision eliminates the human variability that makes most tech comparisons unreproducible. The same automated actuation applies here: the robot performs identical physical actions on both laptops simultaneously. One limitation, noted by Windows Latest's own author, is that he opened the same 500-page document on his personal PC in under 3 seconds — faster than either machine in the test — suggesting that file content, storage configuration, or device setup may affect individual results beyond the OS-level overhead. The pattern across two separate hardware pairs producing the same gap is more significant than any single result: when ARM Windows (Snapdragon X2 Plus) and x86 Windows (Intel Core 5 320) both trail their Apple rivals by the same margin, a hardware-specific explanation becomes difficult to sustain.