Why Google Antigravity 2.0 Split Its Most Popular AI Tools
TL;DR
Google Antigravity 2.0 splits its agent tooling into four pieces: desktop app, IDE, CLI and SDK. The desktop app becomes the hub for scheduling, orchestration and parallel sub-agents. The IDE still exists for coding and AI collaboration, but it is no longer the default download. Auto-updates made some users think their projects or settings had disappeared.
Nauti's Take
The strategy makes sense, but the rollout sounds clumsy. Agent orchestration, terminal control, SDK integration and coding in an IDE are different jobs; forcing them into one surface gets heavy fast.
The issue is not modularity itself, it is the upgrade experience: if users think an auto-update removed their IDE, that is a product problem, not a power-user problem. For teams, Antigravity 2.0 becomes useful only when the new structure is introduced deliberately and documented well.
Briefingshow
The split shows where Google wants agent work to go: away from one all-purpose editor and toward a toolchain for different roles. Developers may get more control, but also more setup, more handoffs and more upgrade confusion. The key test is whether Google explains the separation clearly and avoids making migrations feel like data loss.
Video
Sources
