Canva AI
Last reviewed
Sources
21 citations
Review status
Source-backed
Revision
v3 ยท 4,794 words
Improve this article
Add missing citations, update stale details, or suggest a clearer explanation.
Suggest edit
CtrlK
Last reviewed
Sources
21 citations
Review status
Source-backed
Revision
v3 ยท 4,794 words
Add missing citations, update stale details, or suggest a clearer explanation.
Canva AI is the suite of artificial intelligence features built into Canva, the Australian online design platform, centered on the Magic Studio toolkit launched on October 4, 2023 and reorganized in 2025-2026 into the unified "Canva AI" assistant and Visual Suite 2.0. [1][9] It spans more than 20 tools for text generation, image and video creation, photo editing, design automation, and code generation, drawing on models from OpenAI, Google, Anthropic, and Canva's own Leonardo.ai acquisition. [1][4] By the end of 2025 Canva served more than 265 million monthly active users and recorded roughly 800 million AI tool uses per month, making it one of the most widely adopted generative AI creative platforms in the world. [2][3]
Canva AI is built on a multi-model architecture rather than a single in-house model: Magic Write uses an OpenAI GPT-4 class large language model, Magic Media video uses Google Veo 3, image generation runs on Leonardo.ai's Phoenix model, and Canva Code uses Anthropic's Claude. [1][4] This page covers the platform's history, its acquisitions, the individual Magic Studio features, the 2025-2026 Canva AI and Visual Suite launches, pricing, and competition.
Canva's AI push began with Magic Write in December 2022 and crystallized with Magic Studio, announced on October 4, 2023 at the first Canva Create event, on the company's 10th anniversary. [1] Canva described Magic Studio as "the power of AI, all in one place" and shipped roughly 10 new AI products at once, alongside a $200 million Creator Compensation fund for creators whose work helped train its models and a content-safety layer called Canva Shield. [1]
The platform was reorganized in 2025 and 2026. At Canva Create 2025 (April 10, 2025) Canva launched Visual Suite 2.0 with a conversational AI assistant and Canva Code. [9] At Canva Create 2026 (April 16, 2026) Canva unveiled Canva AI 2.0, described as a re-architecting of the platform around agentic, autonomous design in which the assistant interprets a goal or brief and coordinates multiple tools to produce a complete outcome built from individual, editable objects rather than a flat image. [21]
Canva was founded on January 1, 2013, in Perth, Australia, by Melanie Perkins, Cliff Obrecht, and Cameron Adams. The idea for Canva grew out of Perkins and Obrecht's earlier venture, Fusion Books, which they started in 2007 while Perkins was a student at the University of Western Australia. At the time, Perkins was teaching design programs like Adobe Photoshop to fellow students and noticed how difficult the software was for people without formal training. Fusion Books allowed schools to create yearbooks using an online drag-and-drop editor, and it eventually expanded to New Zealand and France.
Perkins recognized that professional design software like Adobe Photoshop was too difficult for everyday users, and she envisioned a platform that would make graphic design accessible to everyone. In 2012, Perkins traveled to San Francisco to pitch venture capitalists and was rejected by more than 100 investors. She spent months living on her brother's floor while chasing meetings across the West Coast. Through Google Maps co-creator Lars Rasmussen, who had been advising the project, Perkins and Obrecht were introduced to Cameron Adams, a former Google designer who joined as the third co-founder and chief product officer.
By the end of 2013, Canva had raised $1.6 million from angel investors, including Bill Tai and Lars Rasmussen, along with venture capital firms such as Blackbird Ventures, Matrix Partners, and InterWest Partners. The platform launched publicly in 2013 and attracted more than 750,000 users in its first year.
Canva grew rapidly through the mid-2010s and 2020s, completing multiple funding rounds that brought in a total of approximately $572 million in venture capital across 19 rounds (including 3 seed rounds, 5 early-stage rounds, 7 late-stage rounds, and 4 debt rounds). Key investors include Sequoia Capital, Bessemer Venture Partners, General Catalyst, T. Rowe Price, Blackbird Ventures, Felicis Ventures, Bond, and the Ontario Teachers' Pension Plan. Notable angel investors include Bob Iger, the former CEO of The Walt Disney Company.
Notable funding milestones include a $15 million round in 2015 led by Felicis Ventures, a $15 million Series B in 2016 led by Blackbird Ventures, a $40 million Series C in January 2018 led by Sequoia Capital, a $70 million Series D in May 2019 led by Bond and General Catalyst, and a $200 million Series F round in September 2021 at a $40 billion post-money valuation led by T. Rowe Price.
The company is headquartered in Sydney, Australia, at 110 Kippax Street. It employs approximately 5,500 people across offices in Austin, San Francisco, London, Beijing, Manila, and other locations. Sydney remains the largest physical hub with around 935 employees, while over 2,100 additional employees work remotely. A new flagship Sydney campus in Surry Hills is set to open in 2027 after Canva purchased a 12,000-square-meter office building.
Canva reached $2.8 billion in annual recurring revenue (ARR) at the end of 2024 and hit $4 billion in ARR by the end of 2025, growing roughly 35% year over year. [2] Monthly active users grew from about 180 million at the end of 2024 to 265 million by the end of 2025, with over 31 million paid subscribers. [2][3] Only about 12% of monthly active users are paying customers, leaving significant room for monetization growth. AI tool usage within the platform surged to 800 million uses per month, and Canva reported over 10 billion cumulative uses of its AI tools by late 2025. [2][15]
Canva's enterprise business (companies with 25 or more seats) reached $500 million in ARR with 100% year-over-year growth, representing roughly 12.5% of total company revenue. [2] The majority of Canva's revenue still comes from users in North America. AI search engines have become a meaningful distribution channel: Canva reported that by October 2025 users had held more than 26 million conversations with the Canva app inside ChatGPT, and Canva was one of the top 10 most-referred domains from ChatGPT. [2]
Canva's valuation has fluctuated across private market transactions. In September 2021, the company was valued at $40 billion. It later saw valuations as high as $65 billion in August 2025 before a secondary share sale in late 2025 valued Canva at $42 billion. [10] Blackbird Ventures, Canva's largest institutional shareholder, told its limited partners that Canva was "ready" for an IPO in the second half of 2026. [11] Canva hired Kelly Steckelberg, the former CFO of Zoom who helped take that company public, as its chief financial officer in late 2024, a move widely interpreted as preparation for going public. The company is expected to list on the Nasdaq.
Canva has completed a series of acquisitions to strengthen its design capabilities and AI technology stack. The company had made eight known acquisitions as of 2024.
| Year | Company | Description | Reported Price |
|---|---|---|---|
| 2018 | Zeetings | Interactive presentation platform | Undisclosed |
| 2019 | Pexels | Free stock photo platform | Undisclosed |
| 2019 | Pixabay | Free stock media library | Undisclosed |
| 2021 | Kaleido.ai | Background removal AI startup (Austria) | Undisclosed |
| 2021 | Smartmockups | Product mockup generator (Czech Republic) | Undisclosed |
| 2022 | Flourish | Data visualization platform | Undisclosed |
| 2024 | Serif (Affinity) | Professional design software suite (UK) | ~$380 million |
| 2024 | Leonardo.ai | Generative AI image platform (Australia) | ~$320 million+ |
The acquisition of Serif in March 2024 brought the Affinity suite of professional design applications (Affinity Designer, Affinity Photo, and Affinity Publisher) into Canva's portfolio for roughly $380 million, along with Serif's team. [14] These applications compete directly with Adobe Illustrator, Adobe Photoshop, and Adobe InDesign. Following community concern over subscriptions, Canva and Serif pledged to keep Affinity's perpetual licenses and to offer any future subscription only alongside a perpetual option. [14] At Canva Create 2026 the company went further and announced that Affinity would be made free for all users. [21]
The acquisition of Leonardo.ai in July 2024 was a turning point for Canva's AI strategy. Leonardo.ai, also an Australian company, had built the Phoenix foundational model for image generation and had more than 19 million registered users just two years after launching, with its tools used to create over a billion images. [4] The acquisition brought all 120 of Leonardo's employees, including its executive team, into Canva and directly powered the launch of Dream Lab in October 2024. [4][6] Canva co-founder Cameron Adams said the deal was a mix of cash and stock; according to reporting from Startup Daily citing Blackbird Ventures documents, it was worth at least $320 million, though a Blackbird spokesperson stated the actual sale price was higher. [4][5] Canva said Leonardo would continue to run independently with a focus on research and development. [4]
Magic Studio is Canva's centralized hub for AI-powered design tools. It was announced on October 4, 2023, at Canva's first Canva Create event, coinciding with the company's 10th anniversary. [1] The launch introduced roughly 10 new AI products designed to work together within the Canva editor, and Canva described it as "the power of AI, all in one place." [1]
Magic Studio was named one of TIME's Best Inventions of 2024 and won Fast Company's 2024 Innovation by Design Award in the artificial intelligence category. [8][19] As of late 2025, Canva reported over 10 billion cumulative uses of its AI tools. [15]
Magic Studio draws on multiple AI providers and models rather than relying on a single vendor:
| Component | AI Provider / Model | Function |
|---|---|---|
| Magic Write | OpenAI GPT-4 | Text generation, summarization, rewriting |
| Magic Media (images) | Leonardo.ai Phoenix, others | Text-to-image generation |
| Magic Media (video) | Google Veo 3 | Text-to-video generation |
| Dream Lab | Leonardo.ai Phoenix | Advanced image generation with style control |
| Canva Code | Anthropic Claude | Interactive widget creation from natural language |
| Background Removal | Kaleido.ai technology | Automatic background removal |
This multi-model architecture allows Canva to select the best available AI for each specific task, and it provides flexibility to swap in newer or better-performing models as the landscape evolves.
Magic Write is Canva's AI text generation tool, powered by an OpenAI GPT-4 class model. Originally launched in December 2022 using GPT-3, it was one of the first mainstream design platforms to integrate large language model text generation. Magic Write can draft first versions of blog posts, social media captions, cover letters, marketing briefs, and other written content based on user prompts.
Beyond generation, Magic Write supports several editing operations: users can highlight existing text and choose to rewrite, summarize, expand, change tone, or convert between formats (for example, turning a paragraph into a bullet list). The tool also includes a Brand Voice feature that allows organizations to define their writing style, tone, and terminology so that all AI-generated text conforms to brand guidelines. Magic Write supports over 100 languages for both generation and instant translation. According to Canva, users have written more than 10 billion words using Magic Write.
Free plan users receive a limited number of Magic Write uses per month, while Canva Pro subscribers receive 500 monthly credits.
Magic Design is an AI-powered layout generator that creates complete, ready-to-edit designs from a text prompt or an uploaded image. Users describe what they want (for example, "Instagram post for a summer sale at a clothing store") and Magic Design generates multiple design options with appropriate layouts, typography, color schemes, and placeholder content.
The tool can also work from uploaded media: users can drop in a photo, and Magic Design will generate designs that incorporate that image with complementary templates and text. Magic Design draws from Canva's library of over 2 million templates and more than 140 million stock photos to create its outputs. This feature is particularly useful for users who lack design experience, as it eliminates the blank-canvas problem and provides a polished starting point that can be customized further.
Magic Media is Canva's text-to-image and text-to-video generation tool. On the image side, users type a text prompt describing what they want and select a preferred style (photorealistic, watercolor, 3D render, illustration, and others) along with a desired aspect ratio. Magic Media then generates multiple image options that can be inserted directly into a design and further edited using Canva's other tools. [17]
The video generation capability uses Google's Veo 3 model to create video clips of up to eight seconds in 16:9 format from text prompts, with synchronized audio including dialogue, sound design, and music. A separate Magic Video feature can automatically assemble multi-scene vertical videos of up to 60 seconds from user-provided clips or photos, adding templates, transitions, and background music. This is designed for quick social media video production.
Free plan users can generate up to 50 AI images per month, while paid plan subscribers receive expanded generation limits.
Dream Lab is Canva's advanced image generation tool, launched in October 2024 following the acquisition of Leonardo.ai. [6] It runs on Leonardo's Phoenix foundational model, a machine learning system trained on image data, and is accessible from the Canva homepage rather than from within the editor.
Dream Lab offers capabilities beyond standard text-to-image generation. It supports 15 distinct output styles, including photorealistic renders, 3D models, and various illustration styles. Users can upload an existing image as a style reference to influence the generated output, enabling consistent visual aesthetics across multiple generations. Dream Lab handles multi-subject compositions and produces photorealistic portraits with higher fidelity than Magic Media's standard image generation. The tool generates multiple variations of each prompt, allowing users to compare and select the best result.
Dream Lab is available to Canva Pro and Canva Teams subscribers.
Magic Edit allows users to modify specific parts of an image using text prompts. Users brush over the area they want to change and then type what they want to appear in that region. For example, a user could brush over a plain wall in a photo and type "brick wall with ivy" to have the AI replace that area with the described content. The surrounding context is preserved, making the edit appear natural. This feature functions similarly to Adobe's Generative Fill in Photoshop but is accessible through Canva's simpler browser-based interface.
This feature is available on Canva Pro and higher plans.
Magic Expand uses generative AI to extend an image beyond its original frame. This is useful for adjusting aspect ratios (for example, converting a portrait-oriented photo to landscape), fixing tight crops, or creating wider backgrounds for designs that require more space. The AI analyzes the existing image content and generates plausible continuations of the scene along the expanded edges. The feature is comparable to Adobe's Generative Expand and is commonly used when repurposing content across different social media platforms that require different image dimensions.
Magic Grab enables users to select and isolate individual objects within a photo. Once an object is "grabbed," it can be repositioned, resized, rotated, or deleted while the background is automatically filled in. This brings functionality previously limited to advanced photo editing software into Canva's browser-based editor, requiring no manual masking or layer management. The AI uses image segmentation to identify object boundaries and inpainting to fill in the background where the object was removed.
Magic Eraser removes unwanted elements from photos. Users brush over a person, object, shadow, watermark, or any other distracting element, and the AI fills in the space intelligently by matching surrounding textures, colors, and patterns. The tool uses contextual analysis of the surrounding image to generate seamless fill content. This feature is available to Pro and Teams subscribers.
Magic Switch is a format conversion tool that transforms a design from one format into another. A presentation can be converted into a document, a blog post, social media graphics for multiple platforms, or even translated into a different language. The AI handles reformatting, resizing, and restructuring the content to fit the new format while preserving the core message and visual identity.
This was described at launch as a "world-first" capability that eliminates the need to manually recreate content for each platform or channel. For international teams, the translation component is particularly useful: Magic Switch can translate both the text content and adapt the visual layout for different languages in a single operation.
Magic Animate adds motion to static designs with a single click. The AI analyzes the layout and applies appropriate animations to individual elements (text flying in, images fading, icons bouncing) to create polished motion graphics. This is designed for social media content, where animated posts tend to receive higher engagement than static images on platforms like Instagram, LinkedIn, and Pinterest.
Canva's Background Remover uses AI (built on technology from the Kaleido.ai acquisition) to instantly remove the background from any photo with one click. The isolated subject can then be placed on a new background, used in a collage, or layered into a design. This feature is available on paid plans.
At Canva Create 2025 (held on April 10, 2025), Canva unveiled Visual Suite 2.0, described as the company's biggest product launch since its founding. [9] Visual Suite 2.0 introduced a unified workspace where users can work across presentations, documents, whiteboards, spreadsheets, and other formats within a single design canvas, eliminating the need to switch between tools or juggle separate files.
Key additions included:
Visual Suite 2.0 positioned Canva as a broader productivity platform rather than simply a design tool, placing it in more direct competition with Microsoft 365 and Google Workspace for certain use cases.
Canva Code, introduced at Canva Create 2025, allows users to create interactive elements by describing them in plain English. [18] Powered by Anthropic's Claude model, Canva Code interprets natural language prompts and generates HTML, CSS, and JavaScript to produce functional widgets directly on the Canva canvas.
Examples of what Canva Code can create include interactive quizzes, calculators, countdown timers, to-do lists, mini-games, pricing tools, and data entry forms. The feature is aimed at educators building interactive learning experiences, small business owners creating pricing calculators, nutritionists designing personalized meal plans, and content creators who want interactive functionality without coding knowledge. Canva Code is not a full software development environment; it is a widget builder designed for embeddable interactive elements within Canva designs and websites.
At Canva Create 2026 (April 16, 2026), Canva announced Canva AI 2.0, which it framed as a re-architecting of the platform from a design tool with AI features into an AI-native platform with design at its core. [21] Users can describe a goal, share a sketch, or provide a brief, and the assistant interprets the intent, selects the right capabilities, and coordinates them into a complete outcome such as a launch campaign, a client presentation, or a social content series. [21]
A central change is what Canva calls layered object intelligence: everything Canva AI 2.0 generates is built from individual, editable objects rather than a flat locked image, so a user can ask for one element to change (swap an image, adjust a headline, refine a font) without affecting the rest of the design. [21] The release also introduced Ask @Canva, an on-demand design assistant that users summon by tagging @Canva in a comment, as they would a human collaborator, to get feedback and smart edits. [21] Alongside Canva AI 2.0, Canva expanded its Visual Suite with marketing tools and made the Affinity professional apps free for all users. [21]
Canva offers four main pricing tiers. AI features are distributed across all plans but with varying usage limits.
| Plan | Price | AI Features Included |
|---|---|---|
| Canva Free | $0 | Basic Magic Write (limited uses), up to 50 AI image generations/month, Magic Design, limited Magic Media access, 5 GB storage, 2M+ templates |
| Canva Pro | $15/month or $120/year (single user) | 500 Magic Write credits/month, expanded AI image generation, Magic Edit, Magic Expand, Magic Grab, Magic Eraser, Background Remover, Dream Lab access, Brand Kit (up to 1,000 kits), 140M+ premium stock assets, 1 TB storage |
| Canva Teams | $10/user/month or $100/user/year (3-user minimum) | All Pro features plus shared Brand Kits, approval workflows, real-time collaboration, team folders, comment threads, permission controls, asset libraries |
| Canva Enterprise | Custom pricing (contact sales) | All Teams features plus SSO, SCIM provisioning, advanced admin controls, multi-team management, custom support, dedicated account manager |
Canva also provides free access for verified K-12 educators and students through Canva for Education, and registered nonprofits can apply for free Canva Teams subscriptions for up to 50 users.
In September 2024, Canva increased the price of its Teams plan significantly. In the United States, some subscribers on older pricing plans saw the annual cost for a five-person team jump from $119.99 to $500 per year (with a 40% discount for the first 12 months bringing the first-year cost to $300). This represented a price increase of roughly 300%. [12][13]
Canva cited the growing suite of AI tools, including Magic Studio, as justification for the new pricing. [13] The change generated significant backlash from small businesses and freelancers, who argued the increase disproportionately affected the smaller teams that benefit most from the platform. Canva communicated the changes through customer emails rather than a public announcement, which added to the frustration. The price hike came as Canva was preparing for a potential IPO and had just completed its acquisitions of Serif and Leonardo.ai in the same year.
Canva AI operates in a competitive landscape alongside several established and emerging players in the design and creative tools market.
| Competitor | Primary Strength | AI Capabilities |
|---|---|---|
| Adobe Creative Cloud | Professional design, photo/video editing | Adobe Firefly (image generation), Generative Fill, Generative Expand, AI in Premiere Pro and After Effects |
| Adobe Express | Simplified design (competes directly with Canva) | Firefly-powered templates, text-to-image, background removal |
| Figma | UI/UX design and prototyping | AI-assisted design, Figma Sites, Dev Mode |
| Microsoft Designer | AI-powered design integrated with Microsoft 365 | DALL-E integration, AI layout suggestions |
| Midjourney | High-quality AI image generation | Text-to-image via Discord and web interface |
| Kittl | Template-based graphic design | AI-powered logo and design generation |
Canva's primary competitive advantage lies in its accessibility and breadth. While Adobe Firefly and Creative Cloud target professional designers and Figma focuses on UI/UX workflows, Canva serves the much larger market of non-designers, marketers, small business owners, educators, and social media managers. The integration of AI image generation directly into a familiar, template-based editor lowers the barrier to using generative AI for creative work.
Adobe has responded with Adobe Express and deeper Firefly integration across its Creative Cloud suite. Figma, which Adobe attempted to acquire for $20 billion in 2022 before the deal was abandoned in December 2023 due to regulatory concerns, continues to expand beyond UI/UX into broader design territory with features like Figma Sites.
In 2025, designers and teams increasingly adopted hybrid workflows, using Canva for quick content creation and social media graphics, Figma for UI/UX prototyping, and Adobe for advanced photo and video editing. Rather than choosing a single tool, many organizations use all three for different stages of their creative process.
Canva's AI infrastructure relies on a combination of third-party AI models and proprietary technology. The platform uses OpenAI's GPT-4 class models for text generation, Google's Veo 3 for video generation, Leonardo.ai's Phoenix model for image generation, and Anthropic's Claude for code generation. [1][4] The company also uses its own machine learning models for features like layout recommendation, template matching, and Brand Kit enforcement.
All AI processing happens server-side, with the Canva editor sending prompts to backend services that route requests to the appropriate AI model. Generated outputs are then delivered back to the browser-based editor for immediate use in designs. This architecture means users do not need powerful local hardware to access AI features; everything runs in the cloud and is delivered through a standard web browser.
Canva has stated that its AI models are not trained on user documents or designs. The text generation model was trained on publicly available web data. User privacy and data handling remain important considerations as AI becomes more deeply embedded in the platform.
Canva AI has received several notable recognitions:
The platform's AI tools have been used more than 10 billion times cumulatively, with roughly 800 million AI tool uses per month as of late 2025. [2][15] Canva credits its AI features as a significant driver of user growth, with the company's monthly active user base increasing about 20% year over year to 265 million by the end of 2025. [2] The company has also reported that LLM-generated referral traffic to its platform is rising, indicating that AI chatbots are increasingly directing users to Canva when asked about design tools. [2]