MBD131: Canva Locks Features You Already Paid For

17 August 2026 | Issue #131

In this issue:

Canva wants you to upgrade to access some tools you used to use
Claude watermarks its output
What if you have bad taste?
Design process for a book
Aeto iOS app coming soon
My favorite design quote!

Canva users are unhappy and they're not being quiet about it.

Users have been using Canva to create QR codes for years. Now, the QR code comes with a big, fat Canva icon in the middle. If you want to remove that, you have to upgrade your plan again to the Business plan ($25/mo or $250/yr).

To remove the logo, you have to upgrade from Canva Pro to Canva Business

Another user described logging back in after a few weeks away to find that designs they made under Pro, designs they already printed and used, wouldn't work properly anymore. The exact same content that was included two years ago now prompts a Business upgrade to touch.

Neither of these is a price increase in the way people usually mean it. Nobody's monthly bill went up. What changed is subtler and, if the reactions are anything to go by, angrier: the thing they were already paying for stopped being theirs to keep using.

Price is the risky lever. Usability isn't.

Canva already knows what a visible price hike costs. In 2024 it tried moving Teams customers off a flat $119.99/year plan onto per-seat pricing that could run a five-person team $500/year, a 300% jump. The backlash was immediate and loud enough that Canva grandfathered existing customers and published a public pricing promise.

A price increase is a single, visible moment. It's easy to screenshot, and it hits everyone on the plan at once. It creates a lot of negative brand noise.

Restricting what a feature actually does doesn't work that way. Nobody experiences a QR watermark or a shrinking AI allowance as a single event. They hit it the next time they happen to need the thing, spread out over months, easy to mistake for a bug before the pattern becomes obvious. That makes it a much safer way to recover cost than a headline price change, even if the toll on trust ends up the same or worse. It's basically shrinkflation.

This is the actual mechanism behind what's happening at Canva. Magic Studio runs on a credit system that burns on every generation attempt, including the failed ones, and heavy users can buy an AI Pass add-on for $100 a month once they run out. Generative AI has a real, variable cost per use, unlike most of the software Canva built its reputation on, where the cost of one more user was close to zero.

Metering that cost through credits and friction, instead of raising the subscription price to cover it, keeps the sticker price defensible while moving the real constraint out of view. One Reddit comment summed up the result about as bluntly as it gets: the new AI-first editing experience was described as trash, with a plain request for the old, simple edit buttons back.

Adobe is already running the same play

Canva isn't alone in this. Adobe's Creative Cloud subscribers pay through Firefly's generative credits, and those allowances have been getting tighter for existing users while the headline subscription price holds roughly steady. One longtime Adobe subscriber described their monthly credit allowance shrinking to what they called a "laughable" number, with new subscribers getting even less, and said buying extra credit packs became the only way to keep working the way they used to.

The complaint reads almost identically to Canva's. People aren't primarily upset that a bill went up. They're upset that a tool they already paid for, and a tool central to their work systems, stopped doing what it used to do at that price, without ever announcing the change. While we often see this in our personal lives, users get angrier when its tied to their work productivity.

The IPO angle

There's a second explanation going around: that this is about grooming the business for an IPO. I've seen Canva's valuation estimated ~$42 billion as of last August, with annual recurring revenue crossing $4 billion and IPO chatter pointing to a 2026 or 2027 listing. Per-seat billing and feature gating both push exactly the metrics a company wants to show off before going public: revenue per user, upgrade conversion, expansion revenue.

But that $42 billion number is already out of date. Blackbird and Airtree, two of Canva's longest-standing backers, marked the valuation down to $34.9 billion this month, a 17% cut. Canva's own internal valuation, the price employees can actually sell shares at, fell even further, from $38.9 billion to $31 billion over the past year.

The cause wasn't market jitters. A week earlier, Canva had cut its 2026 revenue growth forecast by a third, down to 20%, after second-quarter revenue landed at $921.9 million and missed the company's own guidance. CEO Melanie Perkins said demand for the new AI features actually exceeded what Canva could profitably serve, so the company chose to slow the rollout and rebuild its cost structure rather than keep scaling a product before the economics were ready.

(I'm still waiting for access to AI 2.0. Are you?)

Derek Hernandez, a senior SaaS and AI analyst at Pitchbook, put the underlying shift plainly: generative AI means software is no longer a "zero marginal cost" business, which he calls the industry's secret sauce up until now. You built the software once, and then sold it multiple times. Now there are more ongoing costs that didn't exist before.

Figma hit the identical wall about five days before Canva did. It reported 48% growth and raised its own outlook, and investors still knocked the stock down 15% once they saw its available cash fall from 27% to 14%.

Put together, this is the same story the rest of this piece is telling, just with a bigger price tag attached. Canva bet heavily on AI, underpriced what running it would actually cost, and is closing that gap two ways: taking features back from people who already paid for them, and telling its own investors a smaller growth number than it first promised. A price increase asks people to pay more. What's happening here asks them to trust that what they have today is still theirs tomorrow, while Canva's own numbers, and its own investors, say the company is still working out what the AI bet costs.

Where this is heading

If metering usability instead of raising prices is the actual playbook, expect it to spread rather than reverse. AI compute isn't getting cheap enough fast enough to make unlimited-use pricing sustainable at today's subscription levels, and no software company wants to be the one that raises headline prices into a market this sensitive to it after watching what happened to Canva, Adobe, and Figma. Friction doesn't generate a news cycle. A price increase does.

Not every company is choosing friction. Pega just moved the opposite direction for its enterprise AI agents, replacing per-token charges with a flat fee per completed task, a bet that predictable pricing is worth more to buyers than the margin it costs Pega to guarantee it. That's the exception that clarifies the rule: it works for Pega because enterprise buyers can evaluate a fixed cost per case. It's a harder sell for a consumer tool like Canva, where usage is diffuse and hard to price per unit.

What we'll most likely see in the future: subscription prices for AI-powered tools are going to look surprisingly stable over the next few years. What won't stay stable is how much you can actually do at each price point before you hit a watermark or a credit counter. The real price increase won't show up on your invoice. It'll show up as the moment you notice you're paying the same amount for less than you used to get.

THIS WEEK'S NEWS

Claude Watermarks

We all saw many takes on the announcement that Claude will watermark its output. The best explanation I saw was written by Chris Penn.

Do You Have Bad Taste?

One thing I'm seeing more often when it comes to AI is now that we all have access to the same tools, the only real differentiator is the human using the tools.

But what if you have bad taste? Kara Redman asks this question in her recent article.

BOOK DESIGN

If you're a part of the RISE Community, hopefully you got to participate in the August Book Club where we discussed Matt Wilkinson's book, The Buyer in the Loop. Host Valentina Escobar-Gonzalez asked me a couple of questions because I worked with Matt on the book's design.

For those who attended, if you want to hear the real quote that I butchered, it's further down in this issue.

The Buyer in the Loop book and interior spread

If you have a book in the works and you're interested in how to bring it to life, let me know how I can help.

iOS APP COMING SOON

This past week, I've been building the iOS app for Aeto. I didn't realize how many hoops you have to jump through to get an app submitted and approved. Luckily, I've set it up so I can still add enhancements without needing to push a new version of the app.

It came about because I was away from my computer and I needed to add a task. The mobile web interface didn't work smoothly, so I realized I needed an app (plus, I've never built an app, and I thought it would be a good learning experience).

As of this writing, the app hasn't been approved yet. But it might be by the time you read it. I hope to have more information next week.

So, What is Aeto?

You finish a task, move on, and forget to bill for it. Then invoicing day means digging back through your project board or emails (or Slack messages, or direct messages) trying to remember what you actually did. Aeto fixes that: project management, time tracking, and invoicing in one place, so if it's on the board, it's on the invoice.

Aeto.app. If you want to be a beta tester, let me know!

DESIGN QUOTE OF THE WEEK

“Perfection is achieved not when there is nothing more to add, but when there is nothing left to take away.” – Antoine De Saint-Exupery

My AI disclaimer: The main article was completely different at first. Then I asked Claude if anyone would care. It helped me shape my original idea (Canva moving some features from Pro to Business) into a larger industry trend and the reasoning behind it. It suggested additional examples such as Figma.

ChatGPT created the image.