Google I/O 2026 Developer Preview: Firebase, Gemini 4 & Android 17

Google I/O 2026 opens on May 19 at Shoreline Amphitheatre in Mountain View, and based on the confirmed session schedule and official previews, it may be the most developer-dense I/O since 2023. Two threads run through every preview Google has published: the transition of its developer toolchain to agentic AI, and the convergence of Android, Gemini, and Firebase into a unified platform stack. The event runs May 19–20, with the Google Keynote at 10 a.m. PT carrying the Gemini model reveals and AI platform argument, and the Developer Keynote at 1:30 p.m. PT covering Firebase, Android Studio, Flutter, and the agentic coding story. This guide walks through what is confirmed, what is probable, and what developers should prepare before the keynote lands — so you can test new APIs the day they ship rather than the week after.

Firebase Goes Agent-Native: What Antigravity Actually Is

The most structurally significant announcement confirmed for I/O 2026 is a complete repositioning of Firebase from a mobile backend platform to what Google's session copy calls an "agent-native platform." That phrase is doing real work — this is not a marketing refresh. Firebase is gaining a native path from AI prototyping in AI Studio through to production deployment on Google Cloud, anchored by a new tool called Antigravity.

Antigravity is Google's full-stack application builder, described in session previews as deeply integrated into both Firebase and AI Studio. Based on what has been published, it handles scaffolding, routing, database bindings, and agentic logic from a single prompt-driven interface — essentially a Firebase-native equivalent of what Vercel's v0 does for Next.js, but with native Gemini function calling and a deployment path straight to Cloud Run. For developers already running Firebase projects, this appears to be the lowest-friction path to adding production-grade agentic workflows without switching cloud infrastructure.

The open questions going into the Developer Keynote are substantive. Whether Antigravity ships with MCP (Model Context Protocol) or A2A (Agent-to-Agent) support out of the box would determine whether Firebase agents can interoperate with the broader ecosystem of Claude Code, OpenAI Codex, and third-party agent tooling that has standardized on these protocols. Google co-authored the A2A specification and MCP passed 97 million installs in early 2026, so native protocol support is the obvious move — but it has not been explicitly confirmed.

For pricing, Antigravity and agent-native Firebase features are expected to follow the existing Spark-to-Blaze upgrade path. Developers on the free Spark tier should review their current quotas before building anything that routes Gemini function calls through Firebase at volume.

Gemini 4: What the Signals Say

The Google Keynote at 10 a.m. PT on May 19 is where Gemini model news lands. Google has not confirmed a Gemini 4 announcement, but the circumstantial case is strong. The company shipped Gemini 3.1 Ultra in early 2026 with a 2-million-token context window and native multimodal support. Gemini 3.1 Flash followed as the cost-optimized production variant. The 3.x generation is mature enough that a 4.x reveal at I/O — Google's highest-profile annual platform event — is the architecturally correct next step.

Based on analyst reporting and session previews, the watch list for the keynote includes:

• Gemini 4 base model — Improved reasoning, stronger function call reliability, and deeper multimodal understanding across text, image, video, and audio in a single inference call.
• Gemini 4 Flash — A lightweight 4.x variant priced for high-volume developer workloads, positioned against GPT-5.5 and Claude Sonnet 4.6 on the price-performance curve.
• Native reasoning (thinking mode) — The ability to allocate compute budgets for multi-step reasoning tasks, comparable to OpenAI o1 and Claude's extended thinking. This capability is confirmed on the I/O agenda under "agentic coding."
• Improved tool-calling latency — The existing Gemini API tool-calling implementation has measurable latency gaps on structured multi-step tasks compared to Claude and GPT-5.5. Session copy explicitly mentions improvements here.

What would make developers stop and pay attention is not benchmark slides but concrete API changes: reliable structured output, sub-200ms tool-call round trips, and a Flash-tier model that beats Gemini 3.1 Flash on cost per useful token for agentic workflows. Whether the I/O announcement delivers on those specifics — or remains at the vision-statement level — is the single most important question to answer by 2 p.m. PT on May 19.

Android 17: Breaking Changes Developers Must Handle Now

Android 17 enters its final developer preview this month, with stable release expected to ship alongside or immediately following I/O. Three breaking changes will affect the broadest surface area of existing Android apps.

Predictive Back Gesture Is Now Mandatory

Predictive back gesture, introduced as opt-in in Android 13, becomes mandatory for apps targeting Android 17. Apps that have not implemented android:enableOnBackInvokedCallback="true" in their manifest and registered a proper OnBackInvokedCallback will display an incorrect system animation for the back gesture — a visible regression for users on Android 17 devices. This affects every app using custom back navigation, bottom sheet dismiss, and side-navigation drawer patterns.

The fix is documented in the Android Predictive Back developer guide. Run ./gradlew :app:lint with predictive-back lint rules enabled and resolve any DeprecatedBackGestureNavigation warnings before the stable release date.

Display Refresh Rate API Changes

Android 17 deprecates Window.setPreferredDisplayModeId() for controlling display refresh rates. Apps using this API directly should migrate to the new SurfaceControl.Transaction.setFrameRateCategory() API before the deprecation enforcement deadline. Google has flagged this as a source of visible jank on high-refresh-rate devices (90Hz, 120Hz, 144Hz) if not handled. The change affects apps that manually manage animation frame rates — game engines, video players, and custom drawing surfaces.

Health Connect Moves to Core

Health Connect moves from an optional Play-distributed APK to a system service in Android 17. The practical effects for developers: the isAvailable() availability check is no longer required, background access to continuous health data streams (heart rate, step count, sleep) is available to qualifying applications without the app being in the foreground, and permission request flows use the updated system health permissions sheet. I/O includes a dedicated Health Connect migration lab with hands-on codelabs.

Flutter GenUI: AI-Generated Adaptive Interfaces

One of the more forward-looking items in the I/O session schedule is Flutter GenUI — described as a capability for building "adaptive, AI-generated interfaces dynamically." Google has been deliberately vague in session previews, but the most technically coherent interpretation is that Flutter's rendering engine gains support for consuming Gemini-generated component descriptions at runtime, adapting layout trees without a rebuild cycle or hot reload.

If that interpretation is correct, Flutter would become the first major cross-platform framework with first-party generative UI support — a meaningful differentiator over React Native and Compose Multiplatform, both of which require custom tooling to achieve comparable dynamic layout behavior. The practical use case is context-adaptive mobile apps: a dashboard that restructures itself based on what the user is trying to accomplish, or a form that adds and removes fields based on prior answers, without any pre-written layout variant for each case.

Flutter 3.30 stable is expected to ship concurrent with I/O. The release also includes Impeller rendering improvements targeting issues reported in Flutter 3.29 with Adreno GPU variants on certain Android 14+ devices.

Android Studio and Gemini: What the Agentic Coding Track Covers

Agentic coding — AI agents that plan, write, run, and debug code across multiple files without constant user intervention — is confirmed as a named track on the I/O agenda. Google's confirmed sessions cover three areas:

1. Gemini in Android Studio — Expanded capabilities including real-time multi-file code generation, automated test scaffolding, and migration assistance for Android 17 API changes. The goal is reducing the manual work of handling breaking changes to near zero for common migration patterns.
2. Firebase Genkit — Google's agentic workflow framework for Node.js and Go, with new TypeScript improvements aligned with the Google ADK TypeScript release from April 2026. Sessions cover production deployment patterns and observability tooling for Genkit-based agents.
3. Google ADK integrations — The Google Agent Development Kit, which shipped TypeScript support in April 2026, gets dedicated session time covering integration with Firebase, Cloud Run, and the new Antigravity builder.

The strategic question the agentic coding track has to answer is whether Google is positioning a unified standalone product to compete with Claude Code, GitHub Copilot Workspace, and OpenAI Codex — or continuing the embedded-capability approach (Gemini inside existing IDEs and tools). Anthropic's Claude Code crossed $30B ARR in early 2026 largely on enterprise developer adoption. A Google-branded standalone agentic coding surface would directly enter that market. Whether I/O 2026 contains that announcement will be clear by 3 p.m. PT on May 19.

The AI Studio to Firebase Production Path

Google has been explicit in session copy that I/O will cover the "full path from AI Studio prototype to production deployment on Google Cloud." The current gap — AI Studio is excellent for building Gemini-powered demos, limited for deploying production systems — is a persistent critique that competitors like Vercel AI SDK and Cloudflare AI have already addressed. Antigravity is positioned as the bridge: a tool that takes an AI Studio prototype and deploys it to Firebase Hosting and Cloud Run with a single action.

For developers currently running Firebase projects and evaluating whether to use Gemini or an external model provider, the I/O announcement of a working AI Studio → Firebase → Cloud Run deployment path would resolve the most common friction point: the gap between "this works in AI Studio" and "this runs in production with proper auth, rate limiting, and observability."

Developer Checklist: What to Do Before May 19

Based on what is confirmed for I/O 2026, here is a concrete checklist for the next two weeks:

1. Run predictive-back lint on every Android app you own. The mandatory deadline may not hit day-one of Android 17, but shipping a visible regression before you have time to fix it is worse than finding it early. The lint rules are available in Android Gradle Plugin 8.5+.
2. Set up a working Gemini API project on the free tier. Post-I/O is historically the fastest period for new Gemini API capabilities landing in the free tier. Having a working project before the keynote means you can run new features the same afternoon they ship.
3. Review your Firebase project's Blaze plan quotas. Antigravity and agent-native features will consume Gemini API tokens and Cloud Run invocations. Free-tier Spark projects will need to upgrade for production agent workloads.
4. Upgrade your Flutter project to 3.29 stable now. The migration to 3.30 (expected at I/O) will be smaller and less disruptive from 3.29 than from 3.27 or earlier.
5. Watch the Developer Keynote at 1:30 p.m. PT, not just the Google Keynote. The Google Keynote at 10 a.m. is the vision statement. The Developer Keynote is where API names, pricing tiers, deprecation timelines, and migration guides get published. For working engineers, the 1:30 p.m. session is the one that changes your to-do list.

What Google Has to Answer

Google I/O 2026 arrives at a moment when every major AI provider is making competing claims about developer platform supremacy. OpenAI landed on Amazon Bedrock on April 28. Microsoft's Agent 365 went GA with enterprise governance tooling. Anthropic crossed $30B ARR. In that context, Google's I/O narrative has to answer a direct question: why build on Gemini and Firebase when GPT-5.5, Claude Sonnet 4.6, and Bedrock Managed Agents all exist and are shipping production workloads today?

The honest answer Google has available is infrastructure integration depth. No other provider combines frontier AI models with Android distribution (3+ billion devices), Chrome, YouTube, Search, Google Maps, and Google Cloud billing under a single developer account. Firebase Antigravity, Flutter GenUI, and an agentic Android Studio are Google's way of making that integration advantage concrete — not an abstract enterprise procurement argument, but a real productivity difference for developers building apps that need to work across Android, web, and cloud simultaneously.

Whether the May 19 announcements close the implementation gap with competitors — or remain at the preview stage — is the question every developer watching the keynote should be asking. Mark your calendar: Google Keynote, May 19, 10 a.m. PT. Developer Keynote, 1:30 p.m. PT. All sessions stream free at io.google.