Last of the Unicorns
Is Canva, the last hero of the VC boom, in real trouble?
Until yesterday, Canva was the last hero standing from Australia’s VC boom of the 2010s.
Culture Amp, funded by Blackbird and guided by Atlassian founders Mike Cannon-Brooks and Scott Farquar, has removed its popular CEO and co-founder and still hasn’t worked out a way to make any money. Safety Culture somehow got a US$2.5b valuation despite losing yet another $49m last year for its checklist app, while its c-suite resembles a swingers party.
Atlassian (which was bootstrapped in 2003 and pre-dated the boom) and which arguably funded everything else has seen its share price drop 80% and inspired John Stensholt and Perry Williams to write a soon-to-be-released blockbuster tell-all book.
Canva stood aside as a beacon of VC returns, guided by one of the best product builders in Australia (Mel Perkins) and a super sharp operator (Cliff Obrecht). Obrecht even claimed the business had been profitable for nine years (albeit with a questionable definition of profitability).
As the SaaSpocalypse raged and enterprise rival Figma lost almost 80% of its market value, Canva maintained its US$42b valuation (making it easily Australia’s most valuable unicorn). Canva is notionally worth 50% more than the far larger (and faster growing) Atlassian.
Alas, it appears not only have the wheels come off the Canva bandwagon, but the wagon has veered off the road, fallen off a cliff and burst into flames.
This week the AFR published a series of articles, initially reporting that Canva’s growth had dropped to 30%. A devastating follow-up brought even worse news. Street Talk dropped a bombshell, declaring Canva’s growth in FY26 was not 30%, but only 25%. Even worse, the business is now forecasting a miserly 20% growth for the coming financial year.
The freight train has become a rusty kids’ tricycle.
Canva blamed AI for the horrendous result, with Perkins noting that “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.”
Obrecht claimed “we anticipated we’d be able to bring those costs down quicker than we were able to, and so ultimately we were a bit too ambitious in our assumptions, which meant we couldn’t roll out a competitive free AI product as quickly as we’d hoped.”
These high inference costs would have significantly impacted Canva’s profitability, (the business reported net profit of $26m for the 2025 financial year) but arguably, should have had a less material impact short term on revenue growth (reducing the ‘top of funnel’ free subs).
The business reported that active users had dropped from 265 million to only 208 million which backs up the importance of Canva’s ‘free’ top of the funnel plans.
And while Canva investors and management might claim this is a temporary AI related hiccup as the business reduces model costs, the longer-term revenue trend suggests otherwise.
In 2023, Canva grew at 49%, in 2024 it grew at 45% (these numbers are based on Canva’s recent ASIC filings). Growth was understood to have been 38% last year and is now forecast to fall to only 20% in FY2027.
This is not a one-off AI-induced token-maxxing blip, but rather five years of increasingly weaker growth from the one-time VC darling.
To be sure, a good deal of Canva’s predicament is bad luck. The business until recently had almost no real competitors in the consumer graphic design space (the only semblance of a foe was Adobe, which was focused on enterprise). Canva is now facing competition in its consumer market from not only AI products like Claude Design but also Microsoft’s imitator Designer software, which it can bundle for free into its Office Suite.
Canva maintains its US$42b valuation despite the terrible growth numbers; this is based on a small recent secondary round of share sales (with Robinhood Ventures acquiring US$25m of Canva shares).
But this figure appears significantly overstated given current market valuations of software businesses (let alone software businesses with rapidly diminishing growth).
While Canva hasn’t released its full FY26 (or even FY25) financials (they should be lodged shortly if the company wants to avoid another ASIC fine), based on US$922m revenue for the June quarter, we can estimate Canva’s FY26 revenue was around US$3.7b.
Figma has seen its share price slump by 78% despite reporting very strong growth numbers. Revenue grew by 46% year-over-year (up from 40% the prior quarter) while net revenue retention hit a record 139%. (Figma isn’t profitable on a GAAP basis, but it sounds like neither is Canva currently).
Figma’s market cap is US$14b on around $US1.4b annual revenue, so a 10x sales multiple. But Figma is growing far quicker than Canva (46% vs 20-25%) and should command a far higher multiple.
At US$42b, Canva is currently trading on a historical sales multiple of 11x.
If Canva’s revenue multiple was 5x (similar to Meta, which is growing at 28%, so faster than Canva) it would have a valuation of around US$19b, a massive 55% drop on its most recent valuation.
But even that is likely overvaluing the business given the rapidly slowing growth rate and significant business model risks with the rise of AI driven products like Claude Design.
Airtable is arguably a better comp - it was also growing at 20% and faced significant AI headwinds. This week it announced it was being sold to Bending Spoons for a 2.7x EV/ARR multiple (which would equate to a US$10b valuation for Canva).
Canva’s diminishing free user funnel (it has more than 220 million active users monthly but only 31 million actually paying users) casts real doubt over the company’s ability to grow significant profitability. Those 190 million free loaders had a negligible cost in the pre-AI environment. In an AI world, that top-of-funnel has a significantly higher cost. Even with Canva reducing token costs by 90% using its own models, that’s still a lot more than free users used to cost (which was almost nothing).
That massive funnel of potential paying users just got a lot more expensive. And reducing the benefits they get just means they are more likely to turn to free design products from OpenAI, Anthropic or Microsoft. In this sense, Canva feels far more vulnerable than the enterprise level Figma, which is far more differentiated and not wallowing in free consumer plans.
Adir, my pod co-host, says that owners need to ensure they never miss their golden window – where the market values your business irrationally. Miss that window and the market is likely to overcorrect on the downside.
Canva had the most golden of golden windows in 2021 and 2025, where they could have potentially sold one of the best ever Australian businesses for more than US$30b. Based on the company’s most recent data, that window appears to have slammed shut.
None of this is to say Canva itself is a bad business – it could conceivably pivot to profitably, but like Atlassian has discovered, that comes with a steep valuation discount. If Canva is growing at 10% annually and generating net profit of $200 million, that’s probably a US$6 billion business – a far cry from US$42 billion.