Back to articles

Chrome 164 ships March 2, 2027 — and February 8 decides what's in it

Chrome 164 release date: March 2, 2027, straight from Google's calendar. Branch point Feb 8, first beta Feb 10 — what extension developers should do now.

Maxim Kosterin
8 min read
A hairline timeline with evenly spaced ticks; partway along, a soft orange watercolor stroke branches upward from it, in front of a faint ink enso circle.
A hairline timeline with evenly spaced ticks; partway along, a soft orange watercolor stroke branches upward from it, in front of a faint ink enso circle.

Chrome 164 reaches the stable channel on March 2, 2027. That's not a rumor and not an estimate: it comes straight from Google's own machine-readable release calendar, a feed my forecast pipeline first logged for this milestone on August 15. I fetched it again on November 18. Nothing has moved.

Why write about a release three and a half months out? Because it's the one clock in this ecosystem you can set your watch by. And because the date that decides what 164 contains comes three weeks earlier: the branch point, February 8.

This train carries the whole store. Our catalog tracks 330,566 Chrome Web Store listings with a combined ~2.48 billion store-reported users. Every two weeks, all of them wake up inside a new browser build. Chrome 158 went stable on November 17; five more stable releases, 159 through 163, ship before 164 does.

The schedule, confirmed

From the Chromium Dash milestone schedule for 164. My pipeline logged these five dates on August 15; all five read the same on November 18:

  • Feature freeze: January 18, 2027. In the feed, but not the cutoff. More below.
  • Branch point: February 8. The release branch is cut from main. What hasn't landed waits for 165.
  • Earliest beta: February 10.
  • Final beta: February 17.
  • Stable: March 2.

This isn't a leak or a roadmap promise. It's Google publishing its own calendar as JSON. Chromium's release-cycle doc says branch point dates are fixed, while release dates can still move (a late-breaking regression is its example); holiday adjustments are already in the schedule.

The feed has held up so far: Chrome 155 landed on the exact day my pipeline had logged 99 days earlier (retrospective here).

Two audiences, two deadlines

If you commit to Chromium upstream — the small minority who build the APIs everyone else later types into a manifest — February 8 is your date. The release-cycle doc is blunt: features that aren't code complete by branch point "should be punted to the following milestone."

Each milestone gets two weeks on main, starting at the previous one's branch point. So a change that lands after January 25 (163's branch) and before the February 8 cut gets 164 as its first release. Land later and you ride 165, stable March 16.

This is not hypothetical. runtime.onExtensionLoaded landed on September 30, two days after Chrome 156 branched, and its first stable became 157 (full story here). Two days cost a release cycle.

What about the January 18 feature freeze? Chromium Dash lists one for every other milestone I checked (156 through 166, even numbers only), and the release-cycle doc doesn't define it. It doesn't decide what ships: a declarativeNetRequest change merged on October 21, sixteen days after 158's listed freeze, and Chromium Dash shows 158.0.8118.0 as its first build. I'd treat January 18 as the date to have your change up for review.

If you ship extensions but never touch Chromium — nearly everyone — circle February 10. The earliest beta opens just under three weeks of integration testing before stable. Whatever 164 carries, that's when you can run against it without betting production users on the answer.

And if you follow the W3C WebExtensions Community Group: two proposals I've written about still carry needs-triage labels from Chrome, Firefox and Safari as of November 18 — the permission lifecycle for AI agents (my write-up) and per-call data limits for unlimitedStorage extensions (explainer). Chrome hasn't taken a position on either, so neither is on any train yet. Nothing rides 164 until it merges.

March 2 is a start, not a finish

A stable date is when a version starts shipping, not when your users get it. The rollout is staged; the release-cycle doc's own estimate is one to two weeks to reach everyone.

StatCounter's weekly desktop numbers show it for the last three releases, as a share of desktop Chrome page views:

Version (stable date) Release week One week later Two weeks later
Chrome 155 (October 6) 6% 45% 51%
Chrome 156 (October 20) 9% 44% 50%
Chrome 157 (November 3) 7% 40% not in yet

Same curve every time. Single digits in release week, 40–45% a week later, about half the week after — exactly when the next version ships and starts replacing it. None of the three held much more than half.

And the tail is long: in the week 157 reached 40%, 156 still had 17%, and versions older than 156 added up to roughly four in ten.

Project that onto 164. Single digits in the week of March 2. About half of desktop Chrome by mid-March, just as 165 arrives on March 16. So if you build against anything new in 164, feature-detect — if (chrome.runtime.whatever) — and ship the fallback on day one.

What rides the train

Scale, with the fine print. The 330,566 figure counts Chrome Web Store listings in our catalog as of November 16, including the 45,568 (a combined ~283M store-reported users) already unavailable in the store at our last check.

The ~2.48 billion sums each listing's store-reported user count, so it is not unique people — one person with five extensions counts five times. And the count checks nothing beyond "it's a Chrome Web Store listing": no manifest version, no API usage, no code. It's the population riding this train, not a prediction of how many listings 164 will change.

That population includes Adobe Acrobat, Google Translate, Grammarly, uBlock Origin Lite, Adblock Plus, AdGuard, Chrome Remote Desktop, NordVPN, MetaMask and Claude. Most milestones change nothing for most of them, but some aren't quiet: 154 shipped an opt-in manifest key for events lost to async listener registration; 155 added runtime.onEnabled.

You can't know yet which kind 164 will be. You can know exactly when to look.

What to do before March 2

  1. Committing upstream? Work back from February 8. A change that lands after the M164 branch point ships in 165 at the earliest. Use January 18 as your marker for having it in review.
  2. Put February 10 in your calendar. The earliest 164 beta opens your integration-testing window.
  3. Feature-detect, never version-sniff. A new version is on single digits in its first week and never on everyone, so a capability check beats parsing the user-agent string.
  4. Check your code against the changes we track. npx @extenshi/cli@latest advise . checks your extension against the Chrome, Edge and Firefox changes we list and prints the file and line of each match — local, no account, no API key. For a full pre-upload security scan, @extenshi/cli runs 3 scans + 10 reads free, one-time, with a free account; prepaid credit packs cover you past that.
  5. Know where your listing stands — manifest version, permissions and store status, one search away in the catalog.

See where your extension stands in the catalog → — manifest version, permissions and store status on Chrome, Firefox and Edge. Between now and March 2, five more release trains leave the station. Know what's riding yours.

Sources


Catalog numbers come from Extenshi's scans of Chrome Web Store listings and their store-reported user counts, as of November 16, 2026; the count includes listings already unavailable in their store, and figures shift as we re-scan. Adoption percentages are my arithmetic on StatCounter Global Stats weekly desktop data, fetched November 18, 2026: each version's share divided by the total for all Chrome versions StatCounter lists. StatCounter measures page views, not users. Release dates come from Chromium Dash, fetched November 18, 2026; March 2, 2027 for Chrome 164 is a scheduled date and can still move. Extenshi does not independently verify all claims made by third parties, and references to specific extensions reflect their size in our catalog, not accusations of wrongdoing. If you believe any information is inaccurate, please contact us at [email protected].

Related Articles