Project Starline was Google's trade-show toy for years: booths where two people sitting thousands of miles apart talked to each other feeling like they were in the same room, in 3D, without headsets. The problem was always the same: it was a prototype. At I/O 2026 Starline gets renamed Google Beam, and Google's story shifts.
What Beam is now
Beam is an AI-first 3D video communication platform. Google's volumetric AI model turns standard 2D video streams into realistic 3D experiences from any angle. It's designed to power a new generation of AI-enabled devices — Google no longer talks about custom booths costing hundreds of thousands of dollars, but about a software technology that will live on different hardware.
The piece arriving now: translation in Meet
The concretely available part isn't Beam itself, but some of its capabilities Google is bringing to Google Meet. The headline: real-time, low-latency voice translation that preserves the speaker's tone and expression. It's not automatic subtitling — it's the person's voice, dubbed into another language, retaining prosody and timbre.
For anyone managing distributed teams, it's the kind of feature that actually changes how a call works. International meetings stop being a linguistic compromise: people who don't speak fluent English can speak up in their own language and be understood on the fly.
What to expect
Google didn't share a general availability date for Beam as a platform — full deployment is expected later this year, with the first hardware partners. Voice translation in Meet, on the other hand, is the first concrete thing out of the lab.