Canva has long distinguished itself among high-growth startups by demonstrating that rapid expansion and profitability can go hand in hand. Generative artificial intelligence, however, is presenting a new and costly challenge.

The design-software company reduced its expected revenue growth rate by a third to 20% after the unexpectedly high cost of delivering AI features compelled it to slow its rollout. Canva CEO and co-founder Melanie Perkins told Fortune that user demand for new AI capabilities "significantly exceeded" the company's projections.

"This validated the demand, but also showed us we needed to reduce the cost of completing an AI task to support a broad rollout," Perkins said over email. "Rather than broadly rolling out a product before the underlying economics were ready, we decided to slow the rollout while we rebuilt the architecture, reduced unit costs and strengthened the business model."

The cost challenge arrives at a critical juncture for Canva, as AI is central to its strategy of evolving into a broader workplace-software platform. Perkins had previously told Fortune that the AI market was too fragmented, and Canva has since introduced tools such as Canva Code as it seeks to expand beyond design into enterprise workflows. The company competes not only with Figma but also with Adobe, whose own generative AI suite, Firefly, is embedded across Creative Cloud and has been cited by Adobe as a factor in its pricing strategy.

The situation highlights a wider dilemma spreading across the software industry: companies cannot afford to sit out the AI boom, yet adopting it can erode the lucrative economics of the very businesses they are trying to protect. The issue extends beyond design tools — Microsoft has reportedly subsidized the cost of Copilot for certain enterprise customers, and model providers such as OpenAI and Anthropic have steadily lowered API prices as competition and efficiency gains progress, though inference at scale remains expensive for application-layer companies consuming those models.

"AI is making SaaS no longer a zero marginal cost solution, which has really been what I would call a lot of software's secret sauce up until now," Derek Hernandez, PitchBook's senior research analyst covering the intersection of SaaS and AI, told Fortune. "People want a much more capable product and solution, which through today's technology means cost of usage is becoming a really global challenge for all of these companies."

Perkins stated in her email that Canva has reduced the cost per task by nearly 90% since launching Canva AI 2.0 in April, described as an agentic upgrade to the Canva platform. However, with Canva AI users now creating three times as many designs as in the previous version, the company remains focused on improving its unit economics.

Figma, Canva's public-market counterpart, has disclosed its own version of AI-related trade-offs. According to Figma's second-quarter 2026 financial results, its free-cash-flow margin fell to 14% in the second quarter from 27% in the first, while third-quarter revenue growth was forecast at 36% — a deceleration from its June quarter figure of 48%.

AI costs compress margins for SaaS

Hernandez told Fortune that Canva and Figma represent the "biggest signals" that AI is disrupting SaaS's traditional model, as rising inference expenses — the recurring cost of processing AI requests — manifest as slower growth for Canva and margin compression for Figma. Traditional SaaS companies have historically enjoyed gross margins above 80% because serving an additional user carried negligible infrastructure cost; generative AI breaks that assumption by requiring compute-intensive model inference for each interaction.

"If you have a basic analogy of a car, everything it takes to build a Ford F150 would be training, and then gas, mechanic costs, and anything else would be inference, because that's the point of using the product," Hernandez explained. "Canva and Figma both hit the same wall about five days apart, but they cited it in different places."

The AI cost reset carries particular significance as Canva evaluates a potential IPO. Fortune reported last year that an employee share sale valued Canva at $42 billion, with experts suggesting the company could go public in 2026. Hernandez now tells Fortune that Canva might be targeting a listing sometime next year instead. Public-market investors typically scrutinize both top-line growth and the path to durable profitability, and Canva's decision to slow its AI rollout to protect margins may shape how analysts model its long-term unit economics.

By "making the decision to basically tap the brakes" on the AI rollout, Canva appears to be prioritizing its appeal to future investors.

"I'm sure they're trying to protect their profitability, especially if they want to go to public investors," Hernandez said.

This story was originally featured on Fortune.com.