Canva, the design platform valued in the tens of billions, is running into two problems at once: the price of running AI features and a growing overlap with ChatGPT. That’s according to The Information, which reports the company has hit a speedbump as it pours resources into generative tools while OpenAI’s chatbot creeps onto its turf. Both pressures land on the same spot: the margins that made Canva such an attractive business in the first place.
This is the story playing out across the entire software industry right now, and Canva is a clean example of it.
What’s Actually Happening
Canva built its reputation on simple, cheap, drag-and-drop design. Over the past two years it went all-in on AI, rolling out image generation, text tools, and background editing, and buying its way deeper into the space with acquisitions like Leonardo.ai and Affinity.
Here’s the catch. Every AI image a user generates costs real money in compute. When your product is a low-cost subscription used by hundreds of millions of people, giving away expensive AI features to everyone can quietly eat your profits. The Information’s reporting points to exactly this squeeze.
At the same time, ChatGPT now generates images, edits visuals, and drafts marketing content on its own. For a casual user who just wants a quick graphic, the line between “open Canva” and “ask ChatGPT” is getting blurry.
Why It Matters Now
What stands out here is that Canva isn’t a struggling company. It’s profitable, huge, and well-run. If it’s feeling the AI cost pinch, plenty of smaller software firms are feeling it worse.
The dynamic breaks down into three forces:
- Compute costs scale with usage. More AI adoption means higher bills, not lower ones. That inverts the usual software model where each new user is nearly free to serve.
- Foundation model owners are moving up the stack. OpenAI, Google, and Anthropic aren’t just selling infrastructure. They’re shipping consumer features that compete with their own customers.
- Users are getting comfortable jumping tools. Loyalty to any single app weakens when a general-purpose assistant does 80 percent of the job.
This is significant because it challenges the assumption that adding AI automatically adds value. Sometimes it adds cost faster than it adds revenue.
The Two-Sided Threat
Canva’s situation captures a bind facing most application-layer companies. You can’t skip AI, because customers now expect it. But you also can’t fully control the economics, because you’re renting intelligence from a handful of labs that may compete with you directly.
Say no to AI features and you look dated. Say yes and you import someone else’s cost structure into your business. Neither path is comfortable.
Practical Takeaways
For anyone building or buying AI-powered software, Canva’s speedbump offers a few lessons:
- Watch the unit economics, not just the demo. Before shipping an AI feature, know what each use costs and who pays for it. Free-for-all generative tools can turn a healthy product unprofitable.
- Gate expensive features. Tiered access, usage caps, and premium pricing on the heaviest AI tools protect margins. Canva already limits some AI credits, and expect more of that across the industry.
- Build what the general models won’t. Defensibility comes from workflow, brand assets, team collaboration, and data that a chatbot can’t replicate. Raw image generation alone is not a moat.
- Assume your model provider is also a competitor. If your product is a thin wrapper on someone else’s API, plan for the day they ship your feature themselves.
Canva has the scale and cash to absorb this and adjust. The real question is what happens to the thousands of smaller AI-first startups whose entire pitch is a feature that ChatGPT might fold in for free. Canva’s speedbump is their cliff edge.
More details are available in the original reporting from The Information.