# Example 3 — Is Microsoft's "Return to Native Apps" for Real? Verifying Narratives with AI A walkthrough that applies the practices of Chapter 11, "Verifying Narratives with AI," to a case that touches both technical and investment judgment: **the gap between a large enterprise's strategic slogan and the actual scope of its implementation.** ## What this page demonstrates On April 29, 2026, at Microsoft's Q3 earnings call, CEO Satya Nadella delivered a striking message: **"win back fans."** It was widely read as a strategic declaration that Microsoft would acknowledge user dissatisfaction with Windows performance and return to native apps. Right after, Rudy Huyn, who leads Microsoft Store development, announced the formation of a new team to deliver a **"100% native"** experience, and David Fowler, a key engineer on .NET, declared on social media: **"Native apps are BACK!"** The developer community lit up. Against this narrative, we set up the hypothesis that **"it is hard for Microsoft to return to native apps quickly,"** and we verify it with AI in five steps. After WordPress (over-concentration on one person), Node.js (no one is responsible in a distributed setup), and Linux distros (corporate stewards rewriting their promises), a **fourth governance-failure pattern — "the structural gap between strategic slogan and implementation scope" —** comes into view. > Each looks like a separate phenomenon, but the common thread is > that **the surface narrative covers up the actual situation.** Riding > the surface narrative leads to large mistakes in technology > selection, investment decisions, and vendor evaluation. Verify with > AI, and the cover comes off. --- ## The surface narrative (the April 2026 announcements and how they were received) Let us organize the narrative being put out by Microsoft itself and related media and developers. 1. **"Microsoft is returning to native apps"** — Satya Nadella's declarations of "win back fans" and "foundational work" 2. **"Deliver a 100% native experience"** — Rudy Huyn (Microsoft Store / File Explorer development lead) announcing the new team 3. **"Native apps are BACK!"** — David Fowler (.NET / ASP.NET Core designer) declaring on social media 4. **"Windows 11 had lost its way, but is course-correcting"** — the tone of various tech outlets (TechSpot, OC3D, Windows Latest, etc.) 5. **".NET 10 Native AOT and WinUI 3 will bring a complete native renaissance like the Win32 era of the 1990s"** — expectations from developer blogs Each one sounds powerful. The narrative is "Microsoft has finally turned back to face the user side." Let us verify it with AI. --- ## Step 1: Extract and classify the claims > Please classify the claims in the Microsoft "return to native" > narrative above into "objective fact," "evaluation," and > "metaphor / rhetoric." Where the scope of a claim is not made > explicit, please point that out. Key results of the classification: | Claim | Classification | Verifiability | |---|---|---| | "Microsoft is returning to native" | Evaluation + factual claim, **scope unclear** | Verifiable — scope must be identified | | "100% native" | Factual claim, **target unclear** | Verifiable — target must be identified | | "Native apps are BACK!" | Metaphor / rhetoric | At the level of evaluation | | "Windows 11 is course-correcting" | Evaluation | The substance of "course-correcting" is unclear | | ".NET 10 / WinUI 3 will bring a renaissance" | Factual claim + evaluation | Partially verifiable | Even at this step, an important finding emerges. > **When they say "100% native," what is 100%?** > The whole OS? The whole application? Some system component? > Microsoft itself does not state the scope. Identifying **the scope of "return to native"** is itself the entry point to verification. --- ## Step 2: Cross-check factual claims against primary sources ### The actual scope Rudy Huyn's team is rewriting > Please put together Rudy Huyn's role inside Microsoft, the scope of > the new team, and his past statements in chronological order. > Identify which Windows components are actually being natively > rewritten. What Claude pulls together (key points): - **Rudy Huyn** — a developer who has long worked on Microsoft Store and File Explorer development - The new team's scope is **"the OS shell"** — Start menu, taskbar, File Explorer, the Settings panel, etc. - The May 2026 Windows 11 update (KB5083631) **replaced a 31-year-old legacy property dialog with a WinUI 3 based one** and fixed memory leaks in File Explorer - These are real progress, but **the effort is limited to the shell layer, not the whole Windows OS** In other words, the **target of "100% native" is the OS shell layer**, not the entire application surface running on Windows. ### The path the Microsoft 365 organization chose > Please put together the strategic direction of the Microsoft 365 > division (Outlook, Teams) chronologically from 2024 to today. Show > whether there are any plans to migrate to WinUI 3 / Native AOT, or > whether they are choosing web technologies instead. What Claude pulls together (key points): - **The new Outlook (Project Monarch)** — a plan to phase out the old Win32-based classic Outlook and migrate fully to the new Outlook - However, the new Outlook is **not a native app**. In reality it is based on Outlook Web App (OWA), and on the desktop **renders as a wrapper app via WebView2** - **Microsoft Teams 2.0 / 3.0** — moved away from the heavy Electron framework that drew so much criticism, but **the destination is not WinUI 3 / C++; it is WebView2.** They simply switched the underpinning to Microsoft Edge WebView2 - Teams' new features being rolled out through 2026 (multi-tenant / multi-account MTMA, optimizations for VDI environments such as Omnissa (formerly VMware), etc.) **are all built on the assumption of a WebView2-based, cloud-connected architecture** > **The OS division goes "100% native"; the 365 division goes "100% > web technology" — within the same Microsoft, the strategic > directions are exactly opposite.** This is a fact that collides head-on with the single narrative of "Microsoft is returning to native." ### Microsoft's stance toward third parties > Please summarize the resources Microsoft has invested in the > "React Native for Windows" project from 2024 to 2026, along with > the latest releases. Key points: - Microsoft itself continues to push hard on the open-source project **"React Native for Windows"** - In 2026, versions 0.81 and 0.82 were released - They are **strengthening support for Windows desktop app development with web technologies** — migration to Meta's new Fabric architecture, sample apps for AI image classifiers on ARM64 devices, and so on > Microsoft, toward third parties, **continues to offer "web-based, > cross-platform development" as a realistic answer.** They are not > consistently demanding "100% native." --- ## Step 3: Check time-series consistency ### Consistency of the Outlook migration schedule > Please summarize Microsoft's public announcements about the new > Outlook (Project Monarch) migration schedule chronologically, from > 2023 to today. Key points: - **Original plan**: begin the "opt-out (default-on, manually reversible)" phase of the new Outlook in enterprise environments in **April 2026** - **Late February 2026**: via the admin center (MC949965), Microsoft announced the start of the opt-out phase would be **postponed by about a year, to March 2027** - **Support for classic Outlook itself**: secured to continue **through at least 2029** for both subscription and perpetual-license editions > Even within Microsoft's flagship application, **the migration from > native to web technology is extremely difficult.** Full migration > takes years; a "reverse migration from web to native" happening > quickly is even less likely. ### A "history of wandering" through Windows GUI frameworks > Please summarize the evolution of Windows GUI frameworks > chronologically from the 1990s to today. For each framework, show > how Microsoft itself recommended it and how it was treated > afterward. The progression Claude shows: | Framework | Introduced | Recommended period | How it ended | |---|---|---|---| | **Win32 / MFC** | 1990s | Long | Kept around as "legacy" | | **WPF** | 2006 | About 10 years | "Effectively maintenance only" | | **Silverlight** | 2007 | About 5 years | **Effectively abandoned in 2012** | | **WinRT / Metro / Windows Store apps** | 2012 | About 3 years | Folded into UWP | | **UWP (Universal Windows Platform)** | 2015 | About 6 years | **Collapsed with the end of Windows Phone. Microsoft's own flagship apps did not adopt it** | | **WinUI 3 (Windows App SDK)** | 2021 | In progress | Undetermined | > Former Microsoft engineers (Jeffrey Snover and others) describe > this as a **"history of wandering."** With recommended frameworks > changing one after another from WPF to UWP and then to WinUI 3, and > **past investments going to zero each time,** many developers carry > the suspicion that "even if I invest in Microsoft's new native > framework today, in a few years it may again be replaced by another > architecture." WPF, when it appeared, was strongly recommended as "the next generation of Windows GUI." UWP was loudly announced as "the future, shared across PC, mobile, and Xbox." **The credibility of the "return to native" narrative needs to be evaluated against past declarations of the same kind.** --- ## Step 4: Match against verifiable third-party records ### What web wrapper technologies are actually consuming The biggest motivation pushing "return to native" is dissatisfaction with the resource bloat of web wrapper technologies. We confirm this against third-party measurements. > Please summarize the measured resource consumption of major web > wrapper desktop apps, from third-party benchmarks and reviews. | Application | Architecture | Measured resource use | Main user-experience friction | |---|---|---|---| | **WhatsApp (desktop)** | WebView2 (PWA) | Up to 600 MB RAM even at idle (8 GB RAM environment) | Background residency squeezes overall system memory | | **Discord** | Electron | Up to 4 GB RAM when active | Crashes from memory leaks; degraded performance when running heavy games in the foreground | | **Microsoft Copilot (Windows)** | WebView2 | 500 MB in background, up to 1 GB in use | Despite being a built-in OS feature, the load is comparable to a standalone browser | | **Microsoft Store** | Mixed (with web components) | Several seconds of delay per page transition | Noticeable load delays perceived as a regression from native UWP apps | > On office PCs and education-sector devices where 4 GB RAM is still > common, **resource bloat from the web architecture has hit a > ceiling** — borne out by Satya Nadella's specific emphasis on > "performance improvements for low-memory devices." This is a fact that can be widely confirmed. **The return to native has rational technical motivation.** ### .NET 10 Native AOT — what has been demonstrated > Please summarize the technical maturity and actual performance > improvements of .NET 10 Native AOT, drawing on developer-community > measurements and official benchmarks. What Claude pulls together (key points): - **Released in November 2025**, .NET 10 Native AOT does not use a JIT (Just-In-Time) compiler; instead it produces platform-specific native machine-code binaries directly at build time - **Notable advantages** (developer community measurements): - **Instant startup**: with no JIT compilation needed, startup is nearly instantaneous - **Major reduction in memory footprint**: for small apps the binary fits in a few MB, and runtime memory consumption drops dramatically - **Self-contained execution**: runs even in environments where the .NET runtime framework is not installed - The May 2026 Windows 11 update (KB5083631), which fixed File Explorer memory leaks and replaced the 31-year-old legacy property dialog with WinUI 3 based components, shows that **these tech stacks are not theoretical — they are starting to contribute to real OS stability improvements** > Technically, the return to native is **definitely happening at the > OS shell layer.** But that does not mean **a rollout across the > whole OS or the whole application layer.** ### The economics of third-party developers > From an engineering-economics perspective, please summarize how much > incentive independent software vendors (ISVs) and startup companies > currently have to build "100% native" Windows-only applications > from scratch. Claude's summary (key points): - Almost every globally popular desktop application — Discord, Slack, Notion, Spotify — uses web-based technology like Electron or React Native - The reason is simple: **maintaining a single web codebase lets the same development team deploy applications for Windows, macOS, Linux, and even mobile and the web browser, simultaneously** - For developers to choose WinUI 3 / Native AOT, the case has to be more than emotional ("for the Windows fans") — there needs to be **overwhelming economic rationale**: - Overwhelming productivity gains in WinUI 3's development tools (Visual Studio) - Clear algorithmic preferential treatment for Microsoft Store user acquisition - Overwhelming performance differences on low-memory PCs and ARM-based devices (Snapdragon X-equipped Copilot+ PCs and the like) that win new user segments - At present, these **"measurable advantages" are not adequately provided** > Until developer economics change, the third-party layer **will keep > using web wrapper technologies.** Microsoft's "100% native" has no > coercive power over third parties. --- ## Step 5: Organize what we've learned / what we don't yet know | Item | Conclusion | |---|---| | Going native at the OS shell layer | **Definitely in progress.** Rudy Huyn's team is producing concrete results | | Going native for Microsoft 365 apps | **Will not happen.** Both the new Outlook and Teams 2.0/3.0 are WebView2-based | | Going native at the third-party layer | **No prospect of it.** Economic rationale unchanged | | Technical maturity of .NET 10 Native AOT | **Demonstrated.** Contributing to OS stability improvements | | Meaning of "100% native" | **Limited to the OS shell layer.** Not the whole application surface | | End of classic Outlook | **Continues at least through 2029.** Even web migration is difficult | | Long-term durability of WinUI 3 | **Undetermined.** Given the WPF / UWP precedents, the developer community is cautious | | Meaning of "Native apps are BACK!" | **True in a limited sense.** Only the OS core | And **what we cannot yet verify**: - Whether the OS-architecture shift in the AI PC (Copilot+ PC) era changes third-party developer economics (e.g., competitive advantage from leveraging the NPU) - Whether agentic AI such as Microsoft Agent 365 brings about "the collapse of SaaS" — Satya Nadella himself has predicted that "the concept of conventional business applications may collapse" - In 5 to 10 years, whether WinUI 3 will be abandoned like WPF / UWP, or actually be established as a long-term investment target --- ## The "gap between strategic slogan and implementation scope" comes into view In WordPress, **an individual** carried too much; in Node.js, **no one** took responsibility; in Linux distros, **corporate stewards rewrite their promises.** With Microsoft's "return to native," we see a fourth governance-failure pattern: **the scope of the strategic slogan and the scope of the implementation diverge.** This is not a "lie." > **Microsoft is genuinely returning to native — but only within the > extremely limited scope of the OS shell layer.** The single, powerful slogan "return to native" covers, all under the same words, - The OS shell layer (100% native rewrite in progress) - The first-party enterprise / productivity application layer (locked in on web technology; a return to native does not happen) - The third-party application layer (Electron / React Native continues; status quo) — **three different layers of strategy.** The listener takes it as "the whole of Microsoft is going back to native," but reality is "only an extremely limited part of the OS." ### Microsoft's true strategic landing point The true strategy that emerges from verification is, in a sense, extremely sophisticated. > **Take the system resources (RAM and CPU/NPU) freed up by going > native at the OS shell layer, and use them as headroom to keep > these massive, web-based productivity apps running stably.** > > Then, dedicate that secured hardware headroom to the cloud-attached > giant WebView2 apps, and to the **"autonomous AI agents"** that > will, in time, reshape what business software itself even is. In other words, **the OS itself stops wasting resources and becomes a transparent backstage as infrastructure.** This is the true meaning of Satya Nadella's "foundational work." The "win back fans" message is real, but the way fans are won back is not "a nostalgic renaissance of the 1990s Win32 era." It is **the completion of an extremely sophisticated hybrid structure for the AI PC era.** ### Comparing the four governance-failure patterns Lining up the four cases so far (WordPress / Node.js / Linux distros / Microsoft return-to-native), the structural contrast becomes even clearer. | Case | Direction of failure | Typical damage | How verification exposes it | |---|---|---|---| | **WordPress / Mullenweg** | An individual carries too much (excessive concentration) | The whole organization is jerked around by one person's mood or feuds | Lay individual statements out chronologically; contradictions appear | | **Node.js / npm** | No one manages the whole (distributed absence) | Supply-chain incidents, burnout, unclear lines of responsibility | Organize the governance structure from primary sources; the responsible party is split | | **CentOS / Red Hat / Ubuntu** | Corporate stewards rewrite their promises | Multi-year long-range plans suddenly collapse; migration cost | Lay the policy changes of the past 5–10 years chronologically | | **Microsoft "return to native"** | Gap between strategic slogan and implementation scope (a partial truth presented as the whole) | Misjudging scope in tech selection causes both over- and under-investment | Identify the "target scope" of the slogan from primary sources | These look like separate phenomena, but the common thread is that **the surface narrative covers up the organization's internal contradictions or the limited scope of its implementation.** Verify with AI, and the cover comes off. --- ## Implications for adoption and investment decisions The questions that come out of verification, regarding Microsoft's "return to native" narrative: ### Choosing development frameworks - **Building Windows-only system utilities**: WinUI 3 + .NET 10 Native AOT is **the right choice.** Because Microsoft itself keeps showing implementation patterns at the OS shell layer, long-term support is likely - **Building cross-platform productivity apps**: Electron and React Native (which Microsoft itself is pushing) **remain the right answer.** The Microsoft 365 division's choice confirms it - **Do you need to migrate off Electron because of Microsoft's "return to native"?** **No.** Since Microsoft itself is choosing WebView2 in the 365 division, third parties have no economic rationale to migrate ### Implications for Microsoft stock and related investment - An investment thesis built on the simple narrative of "Microsoft has technically gone back to the right direction" is **risky** - The true strategy is **"turn the OS into infrastructure and yield headroom to the NPU and AI agents"** — this is **investing in the Agentic AI strategy** - With Satya Nadella predicting "the collapse of SaaS," investing in SaaS-class companies needs a different angle ### Enterprise IT strategy - A migration plan premised on "the end of classic Outlook" can be **eased off through at least 2029** (Microsoft itself postponed by one year) - The expectation that "the new Outlook is native, so it's fast" is wrong. It is WebView2 based, and you do not get native responsiveness - Redesigning business workflows that depend on COM Add-ins **does not yet need to be rushed** (revisit when the opt-out phase begins in March 2027 or later) ### The habit of verifying slogans - "100% X," "return to Y," "complete migration to Z" — for any large enterprise's slogan, always check the **target scope and time frame** - A slogan whose scope is not made explicit is highly likely to be **a partial truth being inflated as the whole truth** - One hour of verification with AI will reliably reveal the gap in scope --- ## The power of narrative verification > "Microsoft is returning to native," "Native apps are BACK!" — when > these narratives spread through the media, they affect technology > decisions, investment decisions, and adoption decisions all at > once. > But verify with AI, and you can see that they are **a truth limited > only to the OS shell layer**, that **the Microsoft 365 division is > moving in the opposite direction**, and that the third-party layer > is at **status quo**. The more powerful a narrative, **the more it tends to blur its target scope.** When you hear "100% native," first ask AI: "What is the 100% measuring against?" **The strength of a slogan and the range of its truth are often inversely correlated.** That is the practical value of verifying narratives with AI. **Seeing the structure before adoption** changes the quality of work after adoption. Seeing the scope before investing changes the return after investing. > Speaking the strategic slogan is what enterprises are good at. > Carving out the scope of the strategic slogan is what AI is good at. > Putting AI on the side that does the carving is **our choice.** --- ## Related - Chapter 11 main text: [Verifying Narratives with AI](/en/ai-native-ways/verify-narratives/) - Example 1: [Should we adopt Node.js for production work](/en/ai-native-ways/verify-narratives/example-1/) - Example 2: [Which Linux distribution should we choose](/en/ai-native-ways/verify-narratives/example-2/) - Structural-analysis series: [Security Design for the Mythos Era](/en/insights/security-design/)