Independent Analysis: AI, SaaS, and Revenue Growth Strategy

* Vendor-supplied figures, unaudited.

My first encounter with Canva was a message from Guy Kawasaki in 2013. He had just joined as chief evangelist and was spreading the word with characteristic enthusiasm. The product was simple, browser-based, built on the premise that good design should not require years of training. That premise built a platform now used by a quarter billion people every month.

What I saw at The Shay Hotel in Inglewood on April 15, the day before Canva Create 2026, is a different kind of company. Co-founders Cliff Obrecht, chief operating officer, and Cameron Adams, chief product officer, opened the morning session alongside Stef Corazza, Head of AI Research, working through questions that went well beyond the keynote script. The afternoon session, led by Duncan Clark, Head of Pro-Creative, and Andrew Green, Head of Design, covered something not in the keynote at all: what Canva is doing for the professional creative who was never the target user.

I have covered Canva's trajectory across four posts this year, from the five-acquisition sprint to the Ortto and Simtheory deals to the argument embedded in Magic Layers. Today is where those threads pay out. Canva AI 2.0, unveiled at Canva Create 2026, is the company's clearest statement yet that it is no longer a design tool that added artificial intelligence. It is an artificial intelligence platform that ships design and productivity tools. That inversion has structural consequences for every enterprise that currently runs Canva alongside existing creative suites, Microsoft 365, or a dedicated marketing automation stack.

But the more aggressive move may be the one that did not get a keynote slot.

Two Years of Rebuilding, One Release

The Canva leadership team was specific this morning about what the past two years actually involved. They did not layer AI features onto an existing product. They rebuilt t
he platform from the inside out, converting every tool and every workflow into an application programming interface call. The metaphor one engineering lead coined internally: an orchestra, where the power comes not from individual instruments but from how they play together. That is the honest description of what Canva AI 2.0 is attempting.

The four pillars of the new architecture layer each carry weight on their own terms.

Conversational Design replaces the blank page and template picker with a persistent prompt interface. Describe a goal, and Canva AI generates a fully structured, fully editable design with brand elements applied from the start. The system maintains context as work evolves, supporting iteration at every step. That continuity is what separates it from a text-to-image tool that stops after the first output.

Agentic Orchestration is the coordination layer. A single instruction to create a multi-channel campaign produces outputs across every required format simultaneously. The system understands intent, selects tools, and coordinates them without additional prompting.

Object-Based Intelligence addresses the problem I wrote about in March when covering Magic Layers. Artificial intelligence generation created a bottleneck: fast to produce, impossible to edit. Canva's proprietary Design Model makes everything the system generates remain layered, movable, and editable as if a designer had built it by hand. The generative output and the editable file are the same object.

Living Memory is the least dramatic name for the most structurally significant capability. Persistent memory means the system learns how a team works, keeps projects on brand automatically, and improves with every use. Every design a team has ever made becomes signal for the next one. That is a switching cost that compounds over time.

The Cost Structure Is the Strategy

Canva cannot serve a quarter billion users with AI at scale if it routes every request to third-party frontier models. That constraint shaped the company's entire model investment strategy, and it came through clearly in the morning briefing.

The Leonardo AI acquisition two years ago started as a research team that has since grown to more than 100 researchers operating under the name Canva Original Research and Exploration, or CORE. Internal model development has produced a stack that is significantly cheaper and faster than third-party equivalents for the workflows Canva runs most. Edge deployment on mobile and desktop devices pushes certain inference costs toward zero. For tasks where frontier models remain superior, the platform routes intelligently rather than defaulting to the most expensive option.

That framing matters for enterprise buyers evaluating long-term pricing stability. Canva's AI cost structure is not entirely dependent on what OpenAI, Google, or Anthropic charge next quarter. The internal stack absorbs the bulk of the volume. Third-party models are a capability dial, not a core dependency.

Five Intelligent Workflows, One Competitive Argument

The intelligent workflow features in Canva AI 2.0 are better read as a competitive positioning statement than a feature list. Connectors integrate Slack, Notion, Zoom, Gmail, Google Drive, and Google Calendar at launch, pulling conversations, content, and schedules directly into the design context. A meeting transcript becomes a summary document. A Slack thread becomes a newsletter. A customer email becomes a sales pitch. None of it requires leaving the platform.

Scheduling lets tasks run without a user present. A weekly social content batch translated into ten languages, generated every Friday. Morning briefing documents built from calendar and email context, ready at login. This is Canva competing with workflow automation tools on their own terms.

Web Research brings external intelligence directly into the design. A competitive analysis, a market overview, or a business proposal can be researched, structured, and delivered into an editable Canva document without a separate tool or a separate tab.

Brand Intelligence is the feature likely to close enterprise deals. Every design starts on brand automatically. Applying a new brand across thousands of existing assets becomes a single instruction rather than a team project. For organizations managing creative output at scale across distributed teams, this is what moves Canva from a useful tool to a governance layer.

Canva Code 2.0 now imports HTML. Any AI-generated experience, from any tool, comes into Canva for visual editing, publishes to a custom domain with single sign-on protection, and feeds responses into Canva Sheets. Canva is positioning itself as the output and publishing layer for content generated anywhere in the AI ecosystem, including by competitors.

The Afternoon Session: Canva Moves into Professional Creative Tools

The Pro-Creative session, led by Duncan Clark, Head of Pro-Creative, and Andrew Green, Head of Design, was not on the public agenda. It covered a set of moves that will matter as much to enterprise procurement decisions as anything in the keynote.

Affinity, acquired by Canva in March 2024, went free in October 2025. The core downloadable application, covering photo editing, vector illustration, and page layout, now costs nothing and will stay that way permanently. The stated logic is exactly what it looks like: Canva wants professional designers inside large organizations, the people who control brand and have significant influence over tooling decisions, to stop paying for an expensive alternative subscription and start using Affinity instead.

The results after five months were described as strong. The more significant number, from an enterprise strategy perspective, is that Affinity users are coming back into Canva and upgrading to paid tiers to access AI features. The free professional tool is functioning as a reactivation channel for Canva's broader business.

The integration announced this week makes the strategy explicit. Canva's Brand System is now live inside Affinity. A professional designer working in Affinity sees the same brand colors, fonts, and assets that govern the rest of the organization's Canva output. Work created in Affinity exports as an SVG directly back into Canva's asset library, available immediately to every team member working in Canva. The round-trip is the point: precision design in Affinity, scale distribution in Canva, brand governance enforced at both ends.

Cavalry, the procedural motion design tool acquired in February 2026, went free the same day. It follows the same model: free forever for the core product, Canva AI credits required for AI features, accessible via a Canva Enterprise seat. Canva noted that all motion design for the Canva Create 2026 event itself was produced in Cavalry, and all static assets were produced in Affinity. The tools have been in active internal use long enough to carry that load.

New integrations with DaVinci Resolve and Capture One, announced this week, connect Affinity directly into professional video and photography workflows. Affinity files import into the DaVinci timeline with live update: a graphic changed in Affinity updates almost instantly in the DaVinci workspace. Capture One enables one-click export to Affinity for detailed editing, carrying layers, masks, and annotations intact.

An Model Context Protocol connector launching tomorrow connects Affinity to Claude on the desktop. The integration operates through Affinity's scripting system, which exposes nearly the full functionality of the application. The demonstrated capability included renaming complex layer structures, applying non-destructive edits across a multi-page document, and generating new interface features through scripting. That last one is worth stating clearly: through the Model Context Protocol connection, an AI model can extend Affinity's own feature set without a product update from Canva.

The Culture That Scale Did Not Erode

Canva is an Australian-founded, privately held company with no public market obligations and no quarterly earnings call to manage. That structural fact showed up repeatedly in the customer and partner sessions in ways that are easy to dismiss as soft and worth taking seriously as competitive advantage.

Meghan Gendelman, Chief Marketing Officer for B2B, hosted a panel with enterprise customers from Stripe and Virgin Voyages. Both described a shift from centralized creative production to distributed creation at scale. The Stripe team talked about moving from a centralized production model where quality control lived with one team, to a distributed model where brand-governed templates put creation in the hands of the people who need to act on it. Virgin Voyages, which launched its first ship in 2021 into one of the most difficult operating environments imaginable, said Canva was one of the tools that simply worked when the team was lean, moving fast, and had no room for a steep learning curve. Billy Bogan, vice president of marketing at Virgin Voyages, described the journey toward becoming an AI-native organization and said the platform had been part of the company since the beginning.

The partner session, hosted by Emma Robinson, Head of Growth Marketing, featured Design Pickle and Amazon Web Services. The Design Pickle representative said something that does not appear in any vendor briefing deck: Canva is the only company of this size they have worked with where you can send an email and someone responds. At a company now serving a quarter billion users monthly, that responsiveness is not accidental. It is the result of a leadership team that has not yet decided that scale requires distance.

The Design Pickle session also surfaced a practical insight about enterprise adoption. Onboarding is the variable that determines whether a deployment succeeds or stalls. When templates and brand kits are set up properly at the start, distributed teams adopt quickly. When they are not, teams revert to their old tools and the investment is wasted. Canva AI 2.0, with its Brand Intelligence layer and Living Memory, addresses this structurally rather than leaving it to implementation services.

None of this is guaranteed to survive at the next order of magnitude of scale. But the gap between Canva's customer relationships and those of the incumbents it is displacing is real, and it compounds in the same way that retention does. A platform that enterprise customers feel genuinely loyal to generates different expansion economics than one they use because switching costs are high.

The Open Question Is Integration Depth, Not Feature Count

The architecture Canva described across both sessions at The Shay Hotel is internally coherent. Design model, memory layer, connectors, agentic orchestration, scheduling engine, professional tools with shared brand governance: each component has a clear purpose and a plausible connection to the others. The question enterprise teams will actually test is not whether these features exist but whether they work across the seams.

Pulling a Salesforce field into a Canva presentation is a demonstrated capability. Pulling a Zoom transcript, distilling it into a briefing, applying brand templates, scheduling the output for Monday delivery, and having the professional design team's Affinity assets already loaded in the asset library requires every connector and pipeline to function without friction at the integration points. Canva has the components. The proof is in the end-to-end workflow.

The memory system was discussed candidly. Token management is a real constraint. Feeding every document a team has ever created into an AI context is technically possible and financially unsustainable. Canva's approach is selective memory compression: distill the most relevant context, call it based on the task, and iterate the feedback loop over time. That is the right approach. It is also the approach most likely to surface inconsistency at the edge cases that enterprise teams encounter first.

The research preview launches today to the first one million users who discover it on the Canva homepage. Access expands progressively from there. That is not a soft launch. It is the same scaling pattern Canva has always used: reward the most engaged users first, let usage patterns inform the rollout, expand when the infrastructure proves it can hold the volume.