Canva cut its expected revenue growth rate by a third to 20% after high generative AI delivery costs forced the design-software company to slow its rollout. CEO and co-founder Melanie Perkins revealed that user demand exceeded expectations, prompting a strategic architecture rebuild to reduce unit costs.
The Bottom Line
- Growth Deceleration: Canva slashed its projected revenue growth rate by a third, down to 20%, citing heavy underlying AI compute expenses.
- Margin Compression: Public-market parallel Figma reported its free-cash-flow margin dropping to 14% in Q2 from 27% in Q1 due to similar inference costs.
- IPO Timeline Shift: The economic recalibration directly impacts valuation metrics, as Canva protects profitability ahead of a potential public listing.
The Structural Fracture of SaaS Unit Economics
For over a decade, the software-as-a-service model thrived on a simple premise: zero marginal cost. Once code was written, serving an additional user cost practically nothing. Generative AI just broke that equation.
Here is the math. According to Pitchbook senior research analyst Derek Hernandez, inference expenses—the recurring cost of processing live AI requests—operate like fuel and maintenance costs for a vehicle, whereas initial software training represents the manufacturing phase. When users execute complex agentic workloads, server bills scale directly with usage.
That dynamic caught up with Canva simultaneously as it expanded past basic design into enterprise workflows with tools like Canva Code. User demand for AI features “significantly exceeded” internal forecasts, according to CEO and co-founder Melanie Perkins. But satisfying that demand threatened the bottom line.
“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,” Perkins stated over email.
Engineering Efficiency Under Pressure
To fix the margin leak, the company focused aggressively on efficiency. Perkins noted that Canva reduced its cost per task by nearly 90% following the April launch of Canva AI 2.0. Yet, efficiency gains were partially offset by behavioral changes: users began creating three times as many designs as they did on the previous iteration.
The operational headache is hardly isolated. Figma, often viewed as Canva’s closest public-market comparator, disclosed identical financial friction during the second quarter. Figma watched its free-cash-flow margin contract from 27% down to 14%, while simultaneously forecasting Q3 revenue growth at 36%, down sharply from 48% in June.
These parallel disclosures signal a broader industry reckoning. Software firms cannot afford to sit out the artificial intelligence wave, yet deploying unoptimized models directly erodes the lucrative margins investors have historically rewarded.
| Company | Metric Impact | Previous Period | Current Guidance / Result |
|---|---|---|---|
| Canva | Expected Revenue Growth | projected growth (Projected) | 20% (Adjusted) |
| Figma | Free-Cash-Flow Margin | 27% (Q1) | 14% (Q2) |
| Figma | Quarterly Revenue Growth | 48% (June) | 36% (Q3 Forecast) |
Protecting Valuations Ahead of Public Markets
The timing of the cost reset carries significant weight for Canva‘s capital structure. Last year, employee share sales valued the private design giant at $42 billion, with experts saying the company could go public in 2026, though Derek Hernandez told Fortune Canva might be targeting a time next year. By tapping the brakes on broad AI distribution, leadership is signaling discipline to prospective public equity investors.
As Hernandez observed, the decision to throttle deployment stems directly from a desire to safeguard profitability ahead of an eventual initial public offering. By prioritizing unit economics over unconstrained feature expansion, Canva is attempting to navigate software’s most expensive operational transition without sacrificing the financial health required by Wall Street.
As competition intensifies across enterprise productivity software, the winners will not simply be the companies with the most capable algorithms. They will be the ones that figure out how to run them without burning down their balance sheets.
Disclaimer: The information provided in this article is for educational and informational purposes only and does not constitute financial advice.
