Google's Answer to Canva Removes the Canvas

Google shipped a design tool on September 1 that does not, strictly speaking, let you design. Google Pics reached general availability that morning, roughly three months after its debut at I/O 2026, and the first thing a new user notices is what is missing: no blank artboard, no template gallery, no layer stack. There is a prompt box.

That absence is the strategy. Google is not trying to build a better Canva. It is betting that the canvas itself — the drag-and-drop surface that made Canva a $42 billion company and forced Adobe to keep Express alive as a defensive product — is the part of design software that generative models render optional.

What Google Actually Shipped

Pics is built on Nano Banana, Google's image generation and editing model, and it lands as both a standalone app at pics.new and an overlay inside Workspace. Click an image in Docs or Slides and the Pics editor opens on top of it; Drive support follows in the coming weeks.

The feature set is more considered than a naive text-to-image toy. Object segmentation isolates one element in a generated image and changes it without disturbing the rest of the frame. In-image text editing rewrites or translates words baked into a graphic without altering the font or layout. You can upscale to 2K or 4K, stack multiple edits in one pass, and revert individual changes. Teammates can co-edit the same image.

Google's own framing is unusually blunt about the target. The company says Pics “eliminates the frustration of wrangling pixels and feeling like you need to be a graphic designer to deliver a great image.” Meg Whittenberger, the product manager who announced the launch, described a tool for generating images and making precise edits “rather than starting from scratch” — a phrase that quietly concedes the premise. As TechCrunch's Sarah Perez put it, Pics is a tool where you prompt to create, not where you design something new.

There is no free tier. Access is bundled into plans people already buy: Google AI Pro at $19.99 a month, AI Ultra at $249.99, AI Pro for Education, and Workspace Business Standard ($14 per user per month), Business Plus, Enterprise Standard and Enterprise Plus. That distribution math is the real weapon: Google does not need Pics to win on features, only to be the thing already open when a marketing coordinator needs a flyer by 4 p.m.

The Company It Is Aiming At

Canva is not a soft target. The Australian company ended 2025 with roughly 265 million monthly active users, up from about 180 million a year earlier, more than 31 million of them paying. Annualized revenue sits near $4 billion, and the last known valuation, set by an August 2025 employee share sale, was $42 billion. COO Cliff Obrecht says the company is working toward a 2027 IPO.

On September 3, Canva released its first hard numbers on the productivity push it began with Visual Suite in 2022: 4.5 billion presentations and 1.4 billion documents created, at nearly 3,000 presentations and 1,000 documents a minute. Canva is marching onto Google's lawn while Google marches onto Canva's.

Cameron Adams, Canva's cofounder and chief product officer, argues the incumbent tools are “a bit old-school” and that “productivity and creativity have been separated for a long-time for no really good reason.” But his view of AI's role is narrower than Google's. “AI is great at getting you to 80%,” he has said — and users cannot simply hit send on that 80%. They need somewhere to dive in, move things around, change the words. The goal, he says, is to bring the “human element into it” and move “beyond AI slop.”

Canva's AI transition has not been painless. Demand for its AI features ran hotter than executives expected, driving up compute costs and forcing the company to cut its revenue growth forecast to 20%. CEO Melanie Perkins conceded last month that Canva needed to reduce the cost of completing an AI task before rolling features out broadly.

Why This Matters

When the interface becomes a prompt box, the moat stops being the software and starts being the model plus the distribution. That is bad news for design SaaS in a specific way. Canva's defensibility was never the vector editor; it was a quarter-billion people's muscle memory, a template library, and a marketplace where illustrators and photographers publish work and earn royalties. Google Pics has no marketplace. It generates from a model trained on the kind of work that marketplace pays for. The economics of that substitution get less theoretical every month.

What prompting loses is harder to price but easy to feel. Direct manipulation gives you a tight feedback loop: you nudge a thing, you see the result, you nudge again. Prompting replaces that with a negotiation — you describe, the model interprets, you re-describe. The failure mode is not a wrong pixel you can drag but a wrong understanding you must argue with. Google's object segmentation and in-image text editing are an admission of exactly this: the company reintroduced direct manipulation as a feature, because pure prompting could not close the last 20%.

Adams is right that AI gets you to 80%. The open question is whether most business users ever cared about the other 20%, or were only ever paying for the fastest path to something acceptable. If it is the latter, the canvas was never the product. It was the toll booth.

What To Watch

Three things. First, whether Google extends Pics beyond images into layout and multi-page design, where Canva actually monetizes. Second, whether Canva's AI cost problem forces it to gate features Google gives away inside a $14 seat. Third, the artist question: Canva pays contributors, Google trains on them, and no regulator has drawn a line. If Canva's 2027 IPO stays on schedule, that line gets drawn in a prospectus long before it gets drawn in court.