Google Launches Antigravity 2.0 Coding Tool at Google I/O

Google launched Antigravity just six months ago, in November 2025, as a direct response to Cursor, a fast-growing tool that lets software teams use AI to write and manage code. Version 2.0, announced at Google I/O on May 19, 2026, goes considerably further than the original. The headline feature is the ability to run multiple AI agents working in parallel. Previously, you would start one task, wait for it to finish, then start another. Now, a developer can assign different tasks to different agents at the same time and let them run, including on a scheduled basis while nobody is watching. For any organisation that uses software developers, this matters because it compresses how long a project takes. The update also brings tighter connections between Antigravity and Google's other developer products. Projects started in Google's AI Studio app can now be exported directly into Antigravity on a desktop, carrying all the context with them. Google Cloud customers get a direct path to build and deploy those projects from inside the same tool. For teams already using Google's infrastructure, this removes a lot of switching between separate platforms. The most eye-catching piece for non-developers is what Google is doing with Search. Antigravity's capability to generate working software in real time will power a feature where ordinary users can build small interactive apps directly inside a search result. Someone researching a topic, say, comparing insurance plans or calculating shipping costs, will be able to spin up a custom tool on the spot, without writing a single line of code. This is how AI coding tools stop being a developer story and start being a business operations story. On the competitive picture: Cursor reached $2 billion in annual revenue and was used by more than half of the Fortune 500 by early 2026. Anthropic's Claude Code hit $1 billion in run-rate revenue within six months of launch. The AI coding market is now worth an estimated $12.8 billion globally. Google was late to this space. Antigravity launched with 6% developer adoption by January 2026, well behind Cursor and Claude Code at 18% each. Version 2.0 is Google's attempt to close that gap by tying the tool tightly to infrastructure that Cursor and Anthropic simply do not have: Google Cloud, Google Search, and Android. The pricing changes are also worth noting. Google added a new $100 per month tier and cut its top plan from $250 to $200. Every major AI provider now sits at the same three price points: $20, $100, and $200 per month. Google also changed how it counts usage, moving from a simple limit on the number of questions asked to a measure of how much computing work is actually done. That is more honest for coding tools specifically, because a coding task uses far more computing power than a simple chat message. When you hit the cap, Google's system automatically shifts you to a faster, lighter model rather than cutting you off. For anyone who manages software teams or commissions software work, the practical upshot is this: the tools your developers use are changing faster than most procurement cycles can track. A project that took four weeks last year may take two weeks today, not because your developers got smarter, but because the tools running in the background did. That compression affects how you scope work, how you price it if you are selling it, and how you staff for it.