Figma's Canvas Becomes the Operating System for Code

Executive Summary

Config 2026 is Figma's decisive move to stop being a design tool and start being the multiplayer operating surface for product development. The canvas now hosts motion, 3D, shaders, code layers, and AI agents as first-class materials, with Figma Make pushing toward real code generation and editing rather than static handoff. Q1 2026 revenue of $333.4M at a 79.4% gross margin, layered on 4 million-plus paid seats and a new AI credits monetization stream launched in March, gives Figma both the distribution and the unit economics to compound this expansion. We land constructively bullish: the collaboration layer is the moat, and competitors built around files, IDEs, or single-player AI generators are attacking the wrong surface.

The Business

Figma sells a cloud-native, multiplayer product development platform priced primarily by paid seat, with a new usage-based AI credits layer introduced in March 2026.

Subscription Seats

The core engine is per-seat subscription revenue across design, prototyping, dev mode, FigJam, and design systems, now spanning over 4 million paid seats. That scale is the distribution flywheel: every Config 2026 capability ships into an installed base that already includes designers, PMs, marketers, and a growing cohort of engineers, which is why feature launches translate directly into expansion ARR rather than greenfield sales cycles.

AI Credits

Starting March 2026, Figma began charging for AI credits, layering usage-based revenue on top of seats and explicitly tying earnings commentary to AI adoption. With 79.4% gross margins on the base business, incremental AI revenue at software economics should drop disproportionately to gross profit as adoption ramps, particularly as agents, shaders, and Figma Make all consume credits.

The Canvas as Moat

The durable advantage is not any single feature but the multiplayer canvas itself. Real-time co-editing across roles, with code, motion, and AI agents now sitting natively on that canvas, is structurally hard to replicate from a code-first IDE or a single-player generator. This is what we mean by workflow lock-in: Figma is the place where the work actually happens, not where artifacts are exchanged.

The Numbers

Q1 2026 revenue came in at $333.4M with a 79.4% gross margin, a -41.2% operating margin, and a -42.7% net margin, producing net income of -$142.4M and diluted EPS of -$0.27. The gross margin is the headline: at nearly 80%, Figma is operating at the upper end of SaaS economics, and the operating loss is an investment choice, not a structural problem.

Q3 2025 showed $274.2M in revenue, a 69.4% gross margin, and a net loss of $1.1B (diluted EPS -$2.72) with a -400.1% net margin. That figure looks catastrophic in isolation, but it reflects a one-time accounting event tied to Figma's corporate transition rather than recurring operating economics. The cleaner read is the gross margin path: from 69.4% in Q3 2025 to 79.4% in Q1 2026, a 1,000 basis point expansion in two quarters that points directly at operating leverage as AI revenue scales on top of the subscription base.

The narrative that Figma is a money-losing company should be evaluated against that gross margin and the 4 million paid seat base. The operating loss is funding the canvas expansion announced at Config; the gross margin says the unit economics underneath are pristine.

The Thesis

The bet is that Figma becomes the multiplayer front-end for product development, and that the collaboration layer, not the feature set, is the moat that compounds.

Why now: Config 2026 explicitly positions the canvas to "design in code on the canvas," with code layers, prompting-based interactions, and AI-driven detail editing reframing code as just another design material on a shared surface. Figma Make is moving from static mockups toward generating and editing real code, signaling that the canvas is becoming a code-adjacent execution environment rather than a handoff tool. At the same time, AI agents convert repeatable team workflows into native-feeling plugins via simple descriptions of inputs, rules, and outputs, which means organizational process IP gets encoded directly into Figma.

What has to be true for it to pay off:

1. AI credits adoption scales meaningfully on top of the 4 million-plus paid seat base, which the March launch and Q1 earnings commentary already suggest is underway.

2. Design-to-code expansion drives seat penetration into engineering teams inside existing customers, where Figma already has design and PM beachheads.

3. Gross margin holds at or near the 79.4% Q1 2026 level as AI workload grows, with operating margin inflecting as the investment cycle moderates.

The stock is acting like this is starting to be priced. FIG is up 19% since the early-March baseline against 7% for the S&P 500, a 12-point outperformance worth roughly $4.8B in market cap. At $38, we think the market is rewarding the AI monetization story but still underappreciating the structural moat being built at the collaboration layer.

Where We Differ

The consensus read on Figma is that it is a high-multiple design SaaS company with an AI overlay, exposed to horizontal AI tools (Cursor, v0, Lovable, and the like) that can generate UI without ever touching the canvas. Bears point to the -41.2% operating margin and ask when the spending ends.

Our divergence: the consensus is mispricing the collaboration layer. Horizontal AI generators are single-player by design, optimized for an individual prompting a model. Product development inside any organization of meaningful size is a multiplayer activity across designers, PMs, engineers, and marketers, and Figma is the only platform with native real-time collaboration across all of those roles at 4 million-plus paid seats. Code-first IDEs cannot retrofit that surface because their data model is files and branches, not a live canvas. The market is pricing a feature war; we think the war is over workflow ownership, and Config 2026 is Figma planting the flag.

The Open Questions

Does AI credit revenue scale into a material line item, or stay a power-user curiosity?

The Bull View: With 4 million-plus paid seats and credits required for agents, shaders, and Figma Make, even modest per-seat consumption compounds quickly at 79.4% gross margins. Figma already tied earnings commentary to AI adoption in March, suggesting management sees the curve bending.

The Bear View: Usage-based AI in design has historically lagged because designers ration credits and IT budgets cap them. If adoption concentrates in a small power-user cohort, the line stays small and the AI narrative deflates.

Our Read: The credits stream scales, because Figma is making code, motion, and shaders credit-consuming workflows rather than optional novelties. As the canvas becomes the build surface, credits become non-discretionary for teams using the platform.

Can code-first dev tools attack the collaboration layer?

The Bull View: GitHub, JetBrains, and cloud IDEs are built around files, repos, and branches. Their entire data model is asynchronous. Bolting on multiplayer canvas semantics is a ground-up rewrite, not a feature ship.

The Bear View: GitHub already has Copilot, Codespaces, and significant AI investment. If Microsoft decides multiplayer design is strategic, it has the distribution and the capital to force a fight.

Our Read: Even a credible Microsoft response takes years and lands in the wrong audience. Figma's seat base spans designers, PMs, and marketers, not just engineers. Code-first tools attacking from the engineer side hit a structurally smaller wedge.

Does design-to-code expansion meaningfully grow seats inside existing customers?

The Bull View: Figma already has designers and PMs at most customers. Selling Make and code layers into engineering teams in those same accounts is the most efficient seat expansion path in SaaS, with no new logo acquisition cost.

The Bear View: Engineers are loyal to their IDEs and skeptical of design-tool-origin code. Adoption could stall at a thin handoff layer rather than convert into full engineering seats.

Our Read: Figma does not need to replace the IDE. It needs to be where the spec, the prototype, and the first working code live before the IDE picks it up. That is enough to justify engineering seats, and the gross margin makes the math work.

Risks

Code-first incumbents push collaboration into the IDE

GitHub/VS Code, JetBrains, and cloud IDEs could invest aggressively in real-time multiplayer and AI in the code layer, attacking Figma's expansion from the opposite direction. Magnitude is meaningful but slow: rebuilding around a live canvas data model is a multi-year project, and the audience reached is engineers, not the cross-functional base Figma already owns.

Operating losses persist longer than investors tolerate

A -41.2% operating margin in Q1 2026 requires investor patience. If the AI investment cycle extends without a visible inflection in operating margin, multiple compression is possible even with revenue growing.

AI agents stay confined to power users

If agents and Make remain niche tools used by a small cohort rather than becoming standard operating practice for mid-to-large product teams, the process IP lock-in thesis weakens and credits revenue stays small.

Horizontal AI bypasses the canvas

Tools that generate working UI from prompts could capture early-stage and prototype work without ever touching Figma, particularly for smaller teams and startups. The risk is real at the bottom of the market but limited where multiplayer collaboration is non-negotiable.

Execution risk on UX balance

Building developer-grade code tooling without breaking the designer-friendly UX that built the franchise is a non-trivial design problem. Get it wrong and Figma alienates its core audience while failing to win engineers.

What Would Break This

• Gross margin falling below 75% in any quarter through 2026, signaling that AI workload is structurally lower-margin than the subscription base.

• Paid seat count failing to grow past 4.5 million by year-end 2026, indicating the design-to-engineering expansion is not landing.

• Operating margin failing to improve by at least 1,000 basis points off the Q1 2026 -41.2% level by Q4 2026, suggesting the investment cycle is not converting to leverage.

• A credible competitor (GitHub or otherwise) shipping multiplayer canvas-plus-code with material seat traction at Figma customers within 18 months.

• AI credits commentary disappearing from earnings calls, signaling adoption stalled and management deprioritized the usage-based motion.

The Takeaway

We are constructively bullish on FIG: Config 2026 is the clearest articulation yet of Figma's push to own the multiplayer surface where product development actually happens, and the 79.4% gross margin and 4 million-plus paid seat base give it the economics and distribution to make the bet pay off. The shares are up 19% since the March AI-credits baseline versus 7% for the S&P 500, and we think that still under-discounts the moat being built at the collaboration layer. The single most important reason is structural: horizontal AI generators and code-first IDEs are attacking the wrong surface, and rebuilding around a live canvas is not a feature ship. Watch the gross margin and AI credits commentary in the next two earnings prints. If both hold, the thesis is playing out.