Edge's Manifest V2 Reprieve Is Over

Microsoft announced on August 7 that Edge will start disabling Manifest V2 extensions this month, beginning in Canary, Dev, and Beta and rolling out to Stable with a goal of finishing the consumer transition by the end of 2026. Enterprise-managed devices get a stay of execution until early 2027. If that sounds familiar, it should: it's Chrome's playbook from 2024–2025, run a year later with a politer blog post.

The surprise isn't that this happened. It's that anyone treated Edge as a refuge in the first place.

The fork you were never going to get

When Chrome started flipping the MV2 kill switch — dev channels in mid-2024, Stable through 2025, with uBlock Origin pulled from the Web Store and the last flag-based workarounds stripped out by mid-2026 — a chunk of users migrated to Edge specifically because it still ran their MV2 ad blocker. That was always borrowed time. Edge is a Chromium browser, and the extension platform lives in Chromium, not in Microsoft's fork. Keeping MV2 alive indefinitely would have meant maintaining an ever-growing patch set against upstream, alone, forever — for a feature that mostly benefits users who block ads on a browser owned by a company that also sells them. Microsoft carried the divergence for about a year past Google. That's roughly what carrying a dead upstream API is worth to a downstream vendor, and it's a useful calibration for anyone currently betting on Opera or Vivaldi holding out much longer.

Microsoft's framing leans on ecosystem stats: 95% of the most popular MV2 extensions in the Edge Add-ons store already have MV3 versions, only 58 MV2 extensions with meaningful usage remain, and just three of those lack an MV3 replacement. Those numbers are probably accurate and also beside the point. The controversy was never about whether your PDF annotator would survive the migration. It's about the one category where MV3 changes what's technically possible.

"Successfully transitioned" is doing heavy lifting

Microsoft says ad blockers have "successfully transitioned to MV3." What actually happened is that ad blockers shipped constrained rewrites. MV3 replaces the blocking webRequest API — which let an extension inspect every request in flight and decide, in code, what to do with it — with declarativeNetRequest, where the extension hands the browser a rule list and the browser does the matching. That's genuinely better for security and performance: no extension process sitting in the hot path of every request, no remote code deciding what to block.

But the limits are structural. An extension gets 30,000 guaranteed static rules plus a share of a global pool (up to ~330,000 if nothing else is competing for it), 30,000 dynamic rules, and a 1,000-rule cap on regex rules. Filter lists ship inside the extension package, so updating them means shipping an extension update through the store review pipeline rather than fetching fresh lists hourly — a real problem when ad networks iterate against blockers daily. That's why full uBlock Origin has no MV3 version and never will; uBlock Origin Lite exists instead, minus dynamic filtering and the on-demand list updates that made the original effective against adversarial ad tech. Adequate for most people, and a permanent step down from what MV2 allowed. Both things are true, and vendor blog posts only ever say the first one.

What to actually do about it

If you ship an Edge extension still on MV2: the store stopped accepting new MV2 submissions back in July 2022, and as of this month Partner Center shows deprecation warnings on MV2 listings. A "TBD" date will cut off MV2 updates entirely. The migration itself is the standard Chromium checklist — background page to service worker (your extension now gets killed and restarted, so persistent in-memory state has to move to chrome.storage), blocking webRequest to declarativeNetRequest, no remotely hosted code. If you already shipped MV3 for Chrome, your Edge build is likely a Partner Center resubmission and a smoke test. If you haven't, you're now two browsers past the deadline, and Edge's schedule ("gradually turned off by default," completing this year) means your users start seeing your extension disabled in the next few months, not in some abstract future.

If you're an enterprise admin: the ExtensionManifestV2Availability policy (Edge 123+) keeps MV2 alive on managed devices, and there's an EnableForForcedExtensions mode that Microsoft documents as available "regardless of the manifest migration state" — worth noting if you force-install an internal line-of-business extension nobody has budget to rewrite. But Chrome's equivalent policy had an expiry date too, and enterprise deprecation "expected to begin in early 2027" is your actual planning horizon. Treat the policy as a migration buffer, not a strategy.

If you're a user who switched to Edge for uBlock Origin: that trade is done. Within Chromium your options are uBlock Origin Lite or a browser like Brave that does its blocking natively, below the extension layer. The only place full uBlock Origin keeps running is Firefox, where Mozilla supports MV3 alongside a maintained MV2, blocking webRequest included.

The quiet consolidation

Here's the judgment call: MV3 is a real security and performance improvement, Microsoft's rollout is about as developer-considerate as these things get — four years of warning, staged channels, an enterprise buffer — and the outcome is still a loss. Not because any single decision was wrong, but because every Chromium browser inherits Google's answer to the question "how powerful should content blocking be?", and Google is structurally the wrong company to answer it. With Edge's move, roughly nine in ten desktop users worldwide now sit on that one answer. Firefox's continued MV2 support just became less of a compatibility footnote and more of the only genuine second opinion left in the market — which is a lot of weight to put on a browser with single-digit share.

The MV2 era ends in Edge the way it ended in Chrome: not with a fight, but with a settings page notice and a link to a lighter alternative.