Google Pics launches at I/O 2026 with Nano Banana 2 underneath

Google announced Google Pics at I/O 2026 on May 19–20, alongside an updated version of its image generation model Nano Banana 2. Pics is the consumer-facing product layer: an AI image creation and editing tool that integrates with Google Workspace, starting with Slides and Drive. Nano Banana 2 is the underlying model, which Google describes as faster than its predecessor with improved text rendering and instruction following.

What Google Pics actually does. Per Google's announcement, the editing feature set is built around three primary capabilities. Object segmentation lets users select an element in an existing image and move, resize, or transform it without touching the surrounding image — comparable in spirit to selective-edit tools in dedicated image editors. In-image text editing lets users modify text inside a photo (signs, labels, captions) and translate it into different languages — useful for localised marketing or social assets. Uploads mean users can bring their own images into Pics and apply the same removal, resize, and transformation tools to them, rather than generating from scratch every time.

SynthID watermarking on every image. Every image generated through Google Pics is fingerprinted with Google's SynthID watermark. For creators publishing AI imagery — particularly anyone producing client work, stock contributions, or social media assets that may face provenance questions later — this is a notable shift: the watermark is invisible to viewers but detectable. Whether platforms (Instagram, Etsy, stock libraries) start surfacing or filtering on SynthID is a separate question, but the infrastructure is now in place for them to do so.

Availability is narrower than the keynote framing suggests. Google Pics launches today to a limited group of trusted testers only, not to general users. Broader rollout is scheduled for this summer, scoped to Google AI Pro and AI Ultra subscribers. The keynote did not commit to a free-tier release, nor did it specify exactly which Workspace plans (Business, Enterprise, Education) get access at launch. If you don't have AI Pro or AI Ultra today, you cannot use Pics yet, and Google has not published a price for standalone access.

Why this matters for AI tool users. For creators, marketers, and small-business operators already paying for Google Workspace or AI Pro/Ultra, Google Pics becomes a creator tool you'll have access to without adopting a new vendor. The Slides + Drive integration matters specifically — most quick visual asset work (deck imagery, social cards, internal docs) already flows through those tools, so the value is in skipping the export/import round-trip with Midjourney, Canva, or Adobe. For independent stock contributors, the SynthID watermark is the more strategic story: provenance signals are now baked in at generation time, which over the next 12–18 months will likely affect what platforms accept. See our Midjourney review for the current independent leader on raw image quality, and the Adobe Stock Contributor review for the platform side of stock-asset distribution.

What to do this week. If you don't already pay for Google AI Pro or AI Ultra, there is no action today — Pics is not on a free tier and trusted-tester access is invitation-based. If you are an AI Pro or Ultra subscriber, the practical move is to evaluate Google Pics specifically against your current workflow this summer when general access lands: does the Slides/Drive integration save enough round-trips to justify shifting visual work in-house? For stock contributors and creators publishing AI imagery commercially, treat SynthID as a planning input now — assume provenance fingerprinting becomes standard across major AI image tools within 12 months. See the AI stock photos playbook for the current contributor workflow, and the AI Tool Pricing Database for where Google AI Pro and Ultra sit against competing creator subscriptions.