Canva’s Layoffs Came From Its Own Bets, Not From AI

ByGeoff Dyers
Interior of a Sydney office where designers work at desks, the setting of Canva's 2026 workforce cuts

Canva told staff on 3 November that 1,271 people, 22% of the company, will lose their jobs, change roles or be asked to relocate. The headlines called it AI’s doing. PYMNTS ran “Canva Cuts Jobs as AI Challenges Design Platform’s Growth“, and the AFR’s editorial board said Canva had “joined a list of technology firms in Australia and overseas that are scaling back their workforces in response to AI-driven challenges.”

Canva’s own memo never mentions AI. A spokesperson said the changes are not “using AI to do the work of people who are leaving.” The record points somewhere less flattering: a company that bet its roadmap on AI, lost the people who built its product, misjudged what its own AI would cost, squeezed its users on price, and then met an economy shaken by war. Its rivals in the same market, facing the same AI, kept growing.

The bet: Canva rebuilt itself around AI in April

In February, Canva was still selling growth. Co-founder Cliff Obrecht reported about US$4 billion in annual recurring revenue at the end of 2025, more than 265 million monthly users and over 31 million paying ones. In the same announcement he called chatbots “top of the funnel acquisition platforms”, places Canva would find new customers.

On 16 April, at its Create event in Los Angeles, Canva launched Canva AI 2.0: agents, connectors to Slack, Gmail and Drive, and memory. Capital Brief reported the next day that Canva was moving from feature-based subscription tiers to a business “built around artificial intelligence credits“, and that its stock-market listing had slipped to 2027. Obrecht said: “This transition is nothing but net positive.”

Then the people who built the product left

Within five weeks of the launch, three of the people who built Canva’s product announced they were going:

Novaknown reported the walkout in May. A design company lost its head of design and its chief technologist in the month it bet its future on a new product.

The bill: Canva misjudged its own AI

The bet came back as a bill. In August The Information reported that “Canva underestimated usage of its new AI features“, and that the company was moving toward usage-based pricing. Days later Canva cut its 2026 growth forecast by a third, from 30% to 20%. CEO Melanie Perkins explained:

“We decided to slow the rollout while we rebuilt the architecture, reduced unit costs and strengthened the business model.”

That is a company describing its own pricing mistake. It launched AI features it could not afford to serve at the price it charged, then pulled back. Its backers marked it down. Blackbird and Airtree cut their valuation 17%, from US$42 billion to US$34.9 billion, and US funds T. Rowe Price and Franklin Templeton cut their marks too.

Then it squeezed its users

Canva had already shown how it recovers costs. In September 2024 it moved US Teams customers from US$119.99 a year to US$300 for the first year and US$500 after that; in Australia a five-person team went from A$39.99 a month to A$13.50 per person per month. Canva put it down to “the value of our expanded product experience”.

In 2026 it reached for the same lever. Pricing tracker PricingSaaS recorded Canva Pro rising from US$144 to US$180 a year in the week to 3 August, then falling back to US$144 by mid-September. On 9 September the free plan’s AI allowance fell from 200 standard uses to 20. The direction is the one Capital Brief described in April: from a flat subscription to paying for each AI credit.

And its users left

By late October, AFR Street Talk reported internal data showing Canva’s annualised revenue flat at about US$3.69 billion since January and monthly active users down 12 million year on year. Canva disputes those figures and has not published its own. In February it had reported monthly users growing 20%.

Put the year in order. Canva raised Pro in late July and cut the free plan’s AI allowance in September, and by October the internal data showed 12 million fewer monthly users than a year earlier. Canva charged its users more and gave them less, and in the same months they left.

The Information also reported that some users moved design work to ChatGPT. Every design company faces that chatbot. Only one of them stalled.

Same market, same AI: Figma and Adobe kept growing

Company and period Growth
Canva, 2026 revenue forecast Cut from 30% to 20%
Canva, January to late October (AFR, disputed) Annualised revenue flat
Figma, Q3 2026 revenue Up 43%, to US$391.4M
Adobe, Q3 FY2026 revenue Up 13%, to US$6.76B

Figma paid the same AI serving costs; they cut its free-cash-flow margin from 27% to 14%, and it kept growing anyway. Adobe’s consumer segment, which includes Express, its direct Canva competitor, grew subscription revenue 16% to US$1.91 billion. Anthropic’s Claude Design and Google’s Pics went after the same posters and slides as Canva, and both rivals still grew. If AI were eating design software, it would be eating them too. What Canva did differently was its own choices.

The economy did the rest

Canva made those choices into a worsening economy. The IMF’s April outlook was titled “Global Economy in the Shadow of War” after the US-Iran war began on 28 February; Brent crude spiked above US$126 a barrel, and by October the IMF expected global inflation of 4.7% this year. In Australia, growth was 0.4% in the June quarter as “households continued to behave cautiously”. On 3 November, the day Canva announced its cuts, the Reserve Bank raised its cash rate to 4.85%, noting a slowing economy and rising unemployment.

Australian tech had already been shedding jobs. RationalFX counted 4,450 Australian tech job cuts in the first ten weeks of 2026, more than in all of 2025.

The economy did not single Canva out. Figma and Adobe grew through the same war and the same interest rates. What a weaker economy does is make a company’s own mistakes more expensive to carry, and decide when the cuts land.

Why “AI” gets the blame

Blaming AI is the flattering version of a layoff. Oxford Economics found firms “don’t appear to be replacing workers with AI on a significant scale“: AI accounted for 4.5% of US job losses in 2025, and citing it “conveys a more positive message to investors” than “excessive hiring in the past”. Sam Altman conceded in February that “there’s some AI washing where people are blaming AI for layoffs that they would otherwise do“. MIT’s Paul Osterman put it plainly: “AI is a perfect excuse to justify big layoffs.” In September, US employers attributed more job cuts to a downturn in demand than to AI, according to Challenger, Gray & Christmas.

Canva’s memo says what went wrong in its own words. “Our growth has slowed for the first time in a decade.” Work had become “organised around team structures and narrow areas of ownership.” The fix is to “refound Canva, getting back to our roots of small, empowered teams”: product, growth and platform merge into one organisation, and most staff return to the office two to three days a week from January. That is a company describing its own bloat. The memo’s only nod to the fashionable excuse is one line: “New technology is transforming what people can create.”

Canva’s plan now is to sell its new in-house image model, Craft, inside ChatGPT and Gemini. The 1,271 people paying for this year were not replaced by a chatbot. They were caught in a strategy their leaders chose, priced wrong and are still betting on.

Further Reading