google i/o starts tuesday (may 19). the headline most coverage will lead with is gemini intelligence, the agentic-ai layer for android 17. the more interesting story for indie devs is the one running in parallel: android xr glasses get their first public hands-on demos at i/o, and developer preview 3 of the android xr sdk is already shipping.
what's actually new#
google has confirmed two glasses tracks. ai glasses with no display — speakers, mics, cameras, gemini as the interface. and display ai glasses that add a small in-lens screen for "private information." partner list so far: samsung, xreal, warby parker, gentle monster. first consumer hardware lands in 2027, but the developer story is being seeded now.
the sdk side is more concrete than the keynote slides:
- jetpack projected — apis for the sensors, speakers, and display on glasses hardware
- jetpack compose glimmer — a compose ui library specifically tuned for small in-lens displays
- developer preview 3 — production-ready apis, available now via the android xr sdk
google isn't shy about what it wants: app developers building xr-aware versions of existing apps before the glasses ship, so launch-day stores aren't empty.
why this matters for indie devs#
new form factors are a trap. ask anyone who shipped on watchos, tvos, or vision pro — the api is half-baked, the audience is tiny, the design problems are unsolved, and the dollar-per-hour of engineering is the worst in your portfolio. xr glasses will be the same, at least until late 2027.
but there is one specific thing worth paying attention to. jetpack compose glimmer is closer to a port than a rewrite if you already ship a compose app. that makes it the cheapest possible path to "we support xr glasses on day one" — and being on a launch-day store is the only good reason to ship early on a new apple/google platform. that's where the temporary discovery boost lives, and it's the only thing that justifies eating the cost of being a launch app.
everything else about the xr story can wait. don't redesign your app around glasses. don't burn a sprint on jetpack projected experiments. don't commit to an xr-first roadmap because a vp at google said the word "ambient" five times in a keynote.
the stora angle#
when you do ship to a new form factor, the work nobody talks about is the store presence. apple watch and wear os builds need their own screenshots, their own listing copy, their own compliance approvals. android xr will too. play console already supports per-device-type listings; "xr glasses" will be the next row in that table.
stora's screenshot engine and store listing generator already handle multiple device families per build — when google flips the switch on a glasses device type, that's one more device frame on the list, not a new pipeline to build. the part you don't want to be doing manually the week launch apps are due.
the call#
don't write a glasses-first app this year. but if you have a compose codebase, spend an afternoon next week — after the keynote dust settles — looking at glimmer and projected. file the experience. set a calendar reminder for late q4 to check what shipped and what got cut. when google ships hardware in early 2027 (which is what partner timelines are signalling), the apps on day-one stores will be the apps that did the quiet experiment in mid-2026.
watch the demos on tuesday. don't pivot.