Canva Cuts 22% of Workforce as AI Serving Costs Balloon Past Projections

Canva is laying off 1,271 employees — 22% of its workforce — after AI feature serving costs ballooned far beyond projections, forcing a dramatic cut to its revenue growth forecast. The company that became a household name by making design accessible is now paying a steep price for deploying AI at scale without fully accounting for the infrastructure bill.

What Happened

Canva announced on November 4, 2026 that it is eliminating 1,271 positions across the company. CEO Melanie Perkins described it as "the hardest moment in our thirteen-year history." The cuts span engineering, product, and operations teams globally.

The trigger: AI feature costs. Canva has been aggressively rolling out AI-powered design tools — background removal, image generation, copy suggestions, presentation layouts — features that require significant compute every time a user clicks a button. The cost per AI task was running far higher than internal models predicted.

The Numbers

The financial picture is stark:

• 1,271 employees laid off — 22% of the total workforce
• Revenue growth forecast cut from 30% to 20%
• AI serving cost reduction of nearly 90% since April 2026 — but still not enough to offset scale
• Canva serves more than 200 million monthly active users, making even small per-task costs enormous in aggregate

Perkins acknowledged that despite cutting per-AI-task costs by roughly 90% over seven months, the aggregate cost burden at Canva's scale still required structural changes. When you're running AI inference for hundreds of millions of users, a 90% cost reduction still leaves a significant bill.

Context: The AI Cost Problem Is Wider Than Canva

Canva is not alone. The consumer AI tooling industry has been grappling with the gap between what AI features cost to build and demo versus what they cost to operate at production scale.

Companies including Figma, Adobe, and dozens of smaller SaaS (software as a service) platforms have been adding AI capabilities to stay competitive — often pricing them into existing subscription tiers or offering them free to drive adoption. The assumption was that model costs would fall fast enough to absorb the load. For many, that assumption proved wrong.

AI inference costs — what it costs to run a model every time a user makes a request — have dropped dramatically industry-wide. But Canva's situation illustrates the floor problem: even at dramatically lower per-unit costs, the volume of requests across a 200-million-user platform can generate a cost structure that doesn't fit older revenue models.

What This Means for the Industry

For SaaS companies: The Canva situation is a warning signal. If you've added AI features to a product that was priced before AI, your unit economics may be broken. Canva had the advantage of being profitable and large enough to absorb the hit and restructure. Smaller companies running AI features at a loss may not.

For workers in AI-adjacent roles: Counterintuitively, this round of layoffs was not because AI replaced workers' outputs — it was because the cost of running AI for users consumed budget that would otherwise fund headcount. AI serving costs and AI-driven layoffs are two different phenomena; both are happening simultaneously.

For investors: Revenue growth guidance cuts from 30% to 20% will raise questions across the SaaS sector about whether companies with heavy AI feature exposure have properly modeled their cost of goods sold going forward.

What to Watch

Canva has said it is rebuilding its AI infrastructure architecture — suggesting the company believes it can reduce serving costs further through engineering rather than feature reduction. Watch whether the restructured cost base allows Canva to restore its growth trajectory in 2027, or whether more cuts follow if model costs remain sticky. The broader question for the AI tools industry: how many other companies are carrying similar cost exposure without yet disclosing it?