Canva

IWhat it does

Canva is a private company. It has not listed its shares on any exchange, and I found no public registration statement for it at the US Securities and Exchange Commission (SEC) as of 3 November 2026²⁴. Details are in the IPO section below. Nothing on this page is a price you can trade on a stock exchange.

Canva is an online design tool. A person with no training can make a presentation, a social media post, a poster, a short video, a document or a simple website by dragging items around a page, starting from a library of templates. It was founded in 2013 by Melanie Perkins, Cliff Obrecht and Cameron Adams¹¹, and is based in Sydney⁸.

The reach is large. By the end of 2025 Canva reported more than 265 million monthly active users and more than 31 million paying users¹²⁰. Those are Canva’s own figures as relayed by TechCrunch and by the research firm Sacra, which I could not check against a primary Canva page because canva.com blocked my requests. About 12 percent of monthly users pay (31 divided by 265, my arithmetic). The rest use the free tier.

Over time Canva has bought its way into adjacent products. It bought Pixabay and Pexels (free stock photos) in 2019²⁶, Serif, maker of the Affinity design suite, in 2024 for about US$380 million²⁷, and Leonardo AI (image generation) in 2024²⁸. In 2026 it added Cavalry and MangoAI in February²⁹, and Simtheory and Ortto in April³⁰. In October 2025 Canva said Affinity would be free³¹.

The newest push is artificial intelligence. Canva AI 2.0 launched in April 2026, described by the company as its biggest transformation¹¹¹⁴. The idea is that a user types what they want and the product produces an editable design. That is also where the recent trouble started, covered under the risks section below.

IIHow it makes money

Canva sells subscriptions. A free tier brings people in. Paid plans for individuals and teams unlock premium templates, brand controls, larger storage and more tools. A separate enterprise line sells to companies with more than 25 seats. Canva said that business-customer line doubled in 2025 to about US$500 million of annual recurring revenue (ARR, the yearly value of subscriptions in force at a point in time)¹.

The model is now changing. Canva is moving from feature-based tiers to a model built around AI credits, which Cliff Obrecht called the most significant change in its history¹¹. Reporting based on a briefing to investors says an AI add-on called AI Pass was priced at US$100 per user per month on top of Pro or Business plans, and that Canva has not disclosed how many people buy it or what margin it earns¹⁵. I treat that price as press-reported, not confirmed.

Cost of serving customers matters more than it used to. Classic software costs almost nothing to deliver to one more user. AI features cost real money each time someone presses a button. PitchBook analyst Derek Hernandez told Fortune that AI means software no longer has near-zero marginal cost, and named Canva and Figma as the biggest signals of this shift¹⁴.

Canva has been profitable for nine years, and ended the June 2026 quarter with US$1.47 billion of cash, according to press accounts of the company’s shareholder update¹⁷¹⁸. Margins are not disclosed. I cannot tell how much profit remains after the AI spending.

One number needs care. Canva said ARR was US$4 billion at the end of 2025¹. Revenue for the June 2026 quarter was US$921.9 million¹³. Multiplied by four, that is about US$3.7 billion (my arithmetic). Revenue and ARR are different measures, and the gap may reflect timing, free-tier conversion, currency or the definition used. But it means the US$4 billion should not be read as current revenue.

Canva headline numbers

IIIThe market and the competition

Canva competes for the person who needs a good-looking visual and is not a designer. Its main rivals come from three directions.

First, the design incumbents. Adobe is the large one. Adobe filed fiscal 2025 revenue of US$23.8 billion, up about 10.5 percent from US$21.5 billion (my arithmetic from its filed numbers)⁵. It sells Express as a direct rival to Canva, and sells Photoshop, Illustrator and Premiere to professionals. Figma is the other name. It sells collaborative design software mainly to product and design teams, and it listed in 2025²¹. Its filed fiscal 2025 revenue was US$1.06 billion, and its June 2026 quarter revenue was US$370 million, about 48 percent above US$250 million a year earlier (my arithmetic from its filings)⁶.

Second, the office suites. Microsoft (PowerPoint, Designer) and Google (Slides) give away or bundle presentation and document design. For plain slides, the bundled option is often good enough.

Third, and more important for the next few years, the AI labs. Press accounts name ChatGPT as an overlapping design tool¹⁹, and one investor quoted by Startup Daily said AI has routed us around Canva¹⁷. That is an opinion, not a measurement. Canva’s own counter is that it is a workspace where the output is editable, branded and shareable, not a single generated image. It has also placed itself inside ChatGPT: Canva reported more than 26 million conversations through its ChatGPT app by October 2025¹⁵.

Canva has a different shape from its listed peers. It is much larger than Figma by revenue (about US$3.7 billion annualised against Figma’s US$1.28 billion trailing twelve months²³) and growing more slowly. It is much smaller than Adobe, and reportedly growing about twice as fast.

IVThe moat and how it could be disrupted

I rate the moat contestable. Canva has real advantages, but I do not think they are durable if the way people make visuals changes.

What protects it. Scale: 265 million monthly users and 31 million payers is a lot of distribution¹²⁰. Habit and switching costs: teams store brand kits, templates and files inside Canva, and moving them is a chore. A very large template and asset library: more than 165 million templates and assets, per a trade report¹³. A low price for what it does, and a free tier that feeds the funnel. A brand people trust for non-designers.

What is weaker than it looks. A template library is a moat only while templates are how people start. If a user can describe a design in plain words to a general-purpose AI and get an editable result, the library matters less. Switching costs are modest for individuals, who make up most of the 265 million. Network effects are limited: a Canva design shared with a colleague does not make Canva better for the next user the way a social network does. Canva’s brand is strong, but its AI features lean on models built by others. Perkins said the company had been relying too heavily on frontier models¹³, which means a supplier is also a rival.

The AI turn tests the economics. Canva says it cut the cost of serving an AI task by almost 90 percent since April, and that its in-house models are 17 to 30 times cheaper than comparable frontier models for style transfer, image and video generation¹³. Those are company claims. If true, they help. But they also show how fast cost and capability are moving, and that a rival with a bigger model budget can follow.

How it could break: AI assistants absorb the starting point of design; Microsoft, Google or Adobe bundle comparable generation inside tools already paid for; enterprise buyers decide that a general AI plus an office suite is enough; AI serving costs stay high so margins thin; or growth slows far enough that the premium multiple is not defended. The 20 percent growth guide is a first sign of the last one¹⁴.

Why not thin. Canva has nine years of profit, US$1.47 billion of cash, and real share of a very large user group¹⁷. A competitor has to win those users, not just copy features. That is why I stop at contestable and not at thin.

VThe numbers we can see

Canva files no public accounts that I could find. Everything below comes from company statements relayed by the press, secondary research, or third-party marks, and I label each as such.

Valuation history. In 2021 the company was reported at US$40 billion³³. In 2022 venture investors marked it down by about 36 percent¹⁷. In October 2024 a secondary sale (existing shares changing hands, with no new money for the company) valued it at about US$32 billion, up roughly 23 percent¹⁰. In August 2025 an employee share sale at US$1,646.14 per share valued it at US$42 billion, led by Fidelity with J.P. Morgan Asset Management among new buyers, and described as heavily oversubscribed⁸⁹. Reported, but note the size of the sale was not disclosed.

Since then the marks have fallen. In August 2026 an independent valuation used to set the price of employee shares came in at about US$31 billion, about 20 percent below US$38.9 billion a year earlier¹⁶¹⁸. Early investors Blackbird and Airtree marked their holdings at US$34.9 billion, 17 percent below their prior US$42 billion¹⁶¹⁷. The broker Hiive had an offer implying about US$30 billion, a 29 percent discount, which one report notes is not evidence of a company-wide transaction¹⁶¹⁸. StockAnalysis shows a secondary-market price of US$1,300 per share at its 52-week low on 2 November 2026²⁵. Against the US$1,646.14 sale price, that is about 21 percent lower, which would imply roughly US$33 billion if share count were unchanged (my rough arithmetic, not a market figure).

The sources disagree on some points. The 2025 peak is US$42 billion in US dollars, but some outlets quote the same event as A$65 billion, and Sacra writes US$65 billion¹⁶²⁰. I treat US$65 billion as a currency mix-up, not a valuation. The 2021 figure is US$40 billion per TechCrunch³³, while Startup Daily cites a 54.5 peak that appears to be in Australian dollars¹⁰. StockAnalysis shows an implied US$37 billion in the same page where news items quote lower marks²⁵.

Revenue and growth. ARR was US$4 billion at the end of 2025, said to be up about 43 percent²⁰. Quarterly revenue in the June 2026 quarter was US$921.9 million, up 25.2 percent, and one report says it came in below Canva’s internal guidance¹³¹⁵. Growth is slowing quickly: 43 percent on ARR in 2025, 25 percent on June-quarter revenue, and a 20 percent guide for 2026¹⁴.

Public comparables, from filings and market data as of 2 November 2026. Adobe: market value about US$98 billion, enterprise value to sales 3.8 times, revenue growth about 10 percent⁵²². Atlassian: market value about US$47 billion, enterprise value to sales 7.2 times, on fiscal 2026 revenue of US$6.57 billion⁷²⁴. Figma: market value about US$11.8 billion, enterprise value to sales 8.0 times, on trailing revenue of US$1.28 billion²³. Press says Figma is more than 70 percent below its post-IPO high¹⁶.

Listed comparables (data as of 2 Nov 2026)

VIWho runs it and who owns it

Melanie Perkins is chief executive and Cliff Obrecht is chief operating officer. They co-founded the company with Cameron Adams in 2013¹¹. The ownership split is not public, and I did not find a source for it, so I do not give percentages.

Kelly Steckelberg joined as chief financial officer in late November 2024. She spent seven years at Zoom, including its 2019 IPO¹². That hire is one of the signs that an IPO is being prepared.

Early backers named in sources include Sequoia, Blackbird and Felicis³⁴, with Blackbird and Airtree still holding and publishing marks¹⁶. Fidelity and J.P. Morgan Asset Management took part in the 2025 employee share sale⁸. Press says more than 5,000 employees hold shares under the annual valuation¹⁷.

Canva set up a US parent in Delaware. One report says in early 2025 and another places the redomicile in June 2026; the Tech Times article itself says early 2025 and is dated June 2026¹². I treat early 2025 as the better reading but cannot confirm it. The company said the change helps prepare for a future IPO¹². A September 2025 Australian Financial Review headline says Canva ruled out an Australian listing³².

On governance, the founders appear to control the direction and pace. Perkins described the August 2026 growth slowdown as a deliberate decision to get the economics right¹⁷. That is clear leadership. It also means shareholders without a vote depend on that judgement.

VIIValuation: is the last price supported?

Equity value in US$ billions. Not a share price: a private company has none. Latest mark: independent valuation used to set the annual employee share price (not a trade). Last priced transaction: US$42B, employee share sale, Aug 2025, with marks of US$34.9B (Blackbird, Airtree) and a Hiive offer near US$30B in Aug 2026.

What the last valuation assumes

At US$31 billion and about US$3.7 billion of annualised revenue, the price is about 8 times sales before cash, the same as Figma and above Adobe at 3.8 times. At the older US$42 billion it is about 11 times. To justify US$31 billion on my model, you need growth well above Adobe’s for five years, plus a buyer in the public market willing to pay a software-growth multiple at the end. That means growth holding near 20 percent and AI costs staying low.

Bear, base and bull

• Bear. Start from June-quarter 2026 revenue annualised at US$3.69B (press-reported base, my arithmetic). Growth 15%, 10%, 8%, 7%, 6% over five years to about US$5.7B. Exit at 3.5x enterprise value to sales, just below Adobe’s 3.8x. Add US$1.47B cash, discount at 11% a year, 2% a year dilution. Result about US$11.5B, shown as 12.
• Base. Same start. Growth 20%, 17%, 15%, 13%, 12%, to about US$7.5B. Exit at 5.5x, between Adobe (3.8x) and Atlassian (7.2x) and below Figma (8.0x). Same cash, 11% discount rate (a typical equity return for a risky private company, my choice), 2% a year dilution. Result about US$23B.
• Bull. Same start. Growth 25%, 24%, 21%, 18%, 15%, to about US$9.4B. Exit at 8.0x, equal to Figma and just above Atlassian, justified only if AI costs fall and Canva keeps growth above 15%. Same cash, discount rate and dilution. Result about US$41B.

Comparables used

Where I come down

My base case is about US$23 billion, with a range of US$12 billion to US$41 billion. The latest US$31 billion mark sits above my base and inside my range, closer to the bull side. The US$42 billion trade of August 2025 is far above my base. The market could be right if growth stabilises around 20 percent, if the AI cost cuts hold and margins stay healthy, and if Canva’s user base turns into AI-credit revenue at a higher price. A listed Figma at 8 times sales and Atlassian at 7.2 times show that the public market still pays premium multiples for some software growth. A stronger exit could push value up. My margin of safety figure (base less 25 percent, about US$17 billion) is a stress marker, not a target. This is research, not investment advice.

What I cannot judge

I cannot see audited accounts, margins, the AI gross margin, the cap table, the share count or the preference terms. I could not open Canva’s own newsroom, which blocked automated access, so the ARR and user figures come through TechCrunch, Sacra and press. My discount rate, exit multiples and growth paths are judgement. The spread between my bear and bull is wide because public data is thin, and one forecast should not be leaned on.

VIIIThe IPO

Reported to be preparing an IPO. Not listed, and I found no public S-1, F-1 or confidential-draft notice for Canva on EDGAR as of 3 Nov 2026. EDGAR name search finds only Form D filings for Canva, Inc. (Surry Hills, NSW; filed 2012 to 2016) and unrelated shells or investor vehicles²³⁴. A confidential draft filing would not appear publicly, so this does not rule it out. Cliff Obrecht told Capital Brief in April 2026 that Canva is working toward an IPO in 2027 and is IPO ready, but wants the AI business model change bedded in first¹¹. Canva has a Delaware parent, hired a former Zoom CFO, and a September 2025 AFR headline says it ruled out an Australian listing³²¹². No exchange, ticker, banks or size reported. After the August 2026 valuation cuts, Startup Daily and Fortune say a 2027 listing may slip or wait for better prices¹⁴¹⁷.

What to watch

• A public S-1 or a bank mandate report (Reuters, Bloomberg, The Information)
• The next quarterly shareholder update: growth versus the 20 percent guide, and gross margin on AI
• Next employee share valuation and any tender, and whether prices move back toward US$42 billion
• Figma, Atlassian and Adobe multiples, which set the public market’s willingness to pay
• Disclosure of AI credit and AI Pass uptake
• Any change to the 2027 timeline from Obrecht or Perkins

IXWhat could go wrong

The biggest risk is that AI changes how design starts. I cannot judge how quickly. If the main interface becomes a general assistant, Canva becomes one of several tools that assistant calls, and that weakens pricing power.

The second is AI cost. In August 2026 Canva cut its 2026 revenue growth guide from 30 percent to 20 percent, after finding that the average cost of serving an AI task was too high¹³¹⁴. It slowed the rollout, and one report says some AI features are up to six months behind¹⁵. Canva says the cost per task is down almost 90 percent since April, but did not give the starting cost, the volume or the margin¹³¹⁶. Rapid cost cuts help, but the first quarter after a cost cut is not proof that margins stay healthy as usage grows.

The third is growth that keeps slowing. From 43 percent to a 20 percent guide in a year is a steep step down¹⁴²⁰. A business growing 12 percent is valued like Adobe, at three to four times revenue, not like a growth stock. One investor quoted put Canva near US$12 billion on that logic, a deliberately hypothetical figure, not a price anyone paid¹⁶¹⁸.

The fourth is the exit window. A listing depends on the public market receiving a design-software story kindly. Figma’s drop of more than 70 percent from its post-IPO high is a warning for any comparable¹⁶.

Things that have to go right together, my estimate. For the base case to hold, I need: growth to settle in the mid-teens (probability I put at about 60 percent), AI serving costs to stay under control while usage grows (about 70 percent), and the public market to pay a mid single-digit revenue multiple for such a business at exit (about 60 percent). Multiplying gives roughly 25 percent for all three together. These are my judgement, not data, and they are not independent. The point is that a high valuation needs several things to line up.

What an owner can take out. Canva is private and does not pay dividends that I could find. A holder gets money only through a sale: an employee tender, a secondary sale at a discount, or an IPO. A discount of 29 percent on the Hiive offer shows what a forced seller can face¹⁶. The next best alternative for the same dollar is a listed company with audited accounts and daily prices, such as Adobe or Atlassian, which carry lower multiples and visible filings²²²⁴.

What would change my view: an audited filing or public S-1 showing margins and AI costs; a priced primary round or tender at a higher level; two quarters of growth holding at 20 percent or better; or evidence that the AI add-on sells in volume. A drop in paying users or a further forecast cut would push the other way.