Figma’s AI Credits Are Working, but the Bigger Story Is a Widening Product Stack and a C-Suite Shakeup
Q2 2026 earnings call, August 5, 2026
Figma’s first full quarter of AI credit monetization delivered exactly the proof point investors were waiting for: usage-based revenue is additive to gross margin, not dilutive. Revenue reached $370 million, up 48% year-over-year, marking a third straight quarter of accelerating growth. Non-GAAP gross profit grew 40% year-over-year, nine points faster than the prior quarter, and gross margin expanded 2.5 points sequentially to 85%. CFO Praveer Melwani was explicit that this inflection was AI-driven: “The acceleration in gross profit growth and the improvement in gross margin this quarter is the result of our first full quarter of AI credit monetization.” That is the single most important data point in the quarter — it validates the credit-based pricing model Figma rolled out in March 2025 actually converts usage into profitable revenue, not just top-line growth funded by inference losses.
Guidance Deliberately Leaves Money on the Table
Despite the beat, Figma’s third-quarter guide of $373 million to $375 million (36% growth at the midpoint) struck several analysts, including Goldman Sachs’ Gabriela Borges and Citigroup’s Tyler Radke, as conservative relative to the magnitude of the Q2 beat, and the stock fell in aftermarket trading following the print. Management’s explanation is a familiar one for Figma: newer surfaces — Figma agent, Code Layers, Make on local code, Motion, and generative plugins — are all live in beta or early access but not yet drawing down paid credits. Melwani was blunt about the guidance philosophy: “Today, at this moment, we are not taking credit for the products that are in early access programs or beta in our full year revenue outlook. I think that represents upside as we transition from these periods where the credits are not drawing down paid credits.” In other words, management is intentionally under-promising on products it hasn’t proven can monetize yet, which explains both the raised full-year guide (to $1.463 billion–$1.467 billion, up $40 million) and the seemingly light sequential step-up. Investors should read the softer near-term guide as a funding decision, not a demand problem — Figma is holding full-year operating margin guidance flat at 9% even as it leans into inference spend and new product investment, effectively banking the beat into future R&D rather than near-term profit.
Code Layers and the “Code as a Commodity” Thesis
The most important strategic articulation on the call was CEO Dylan Field’s framing of where value accrues as AI commoditizes code generation. “As AI makes code easier to write, everything though is starting to look the same. These models are trained on what already exists. So what you get back is in distribution. It's the expected answer,” Field said. His argument is that Figma’s moat shifts from being a design tool to being the canvas where taste, iteration, and differentiation happen once code itself is cheap. Code Layers — announced at Config in June and rolling into early access soon — puts interactive, editable code directly on the Figma canvas, letting teams compare AI-generated variations side by side rather than converging on generic output. This is a meaningful extension of Figma Make, which already lets teams ship directly into production codebases; Field noted 1Password now runs a full AI-assisted prototyping pipeline through Make, complete with custom MCP integrations, and has opened Make access to licensed engineers rather than just designers.
MCP Adoption Is a Leading Indicator Wall Street Is Underweighting
Figma’s Model Context Protocol server — which lets external AI tools push work into and pull work out of Figma — is quietly becoming one of the more important adoption metrics on the call. Write-to-Figma MCP usage grew 75% quarter-over-quarter (Field clarified this was sequential, not annual, after an analyst mischaracterized it), and management pointed to Clay’s design team as a concrete example: designer Alex Fortney used MCP to pull legacy components out of an old codebase and into Figma for a full visual audit, calling it a tool that “has saved me countless hours of manual labor on all the design system files.” The strategic tension worth watching is what Field flagged himself — as Figma agent absorbs more of the end-to-end workflow natively, there could be a trade-off with MCP usage over time, since MCP today partly serves as a bridge for work that starts outside Figma. Management says it’s watching this dynamic but hasn’t seen cannibalization yet.
Agent Adoption Is Real, But Not Yet Monetized
The Figma agent, in open beta since June, is already used weekly by more than 50% of paid customers with over $10,000 in annual recurring revenue, and more than 20% of weekly credit-consuming users on paid plans are exclusively using AI through the agent — a sign it’s expanding Figma’s user base beyond traditional credit consumption habits rather than just cannibalizing existing usage. None of this is in paid credit consumption yet, meaning the reported 48% revenue growth understates the agent’s current engagement. Management frames the agent as a “capable design intern,” and its ability to spawn generative plugins — custom, reusable tools built on natural-language requests — saw weekly plugin creation more than double after launch as of July 31. This is arguably the most forward-looking data point on the call: a expanding surface area for consumption that hasn’t yet hit the income statement.
Unexpected Executive Turnover Adds a Layer of Uncertainty
Buried beneath the product narrative is a notable amount of leadership churn for a company just over a year into its life as a public company. CTO Kris Rasmussen is shifting to a new Chief Architect role focused on the Figma agent, with Figma opening a search for his replacement and AI/editor engineering teams reporting directly to Field in the interim. Chief Product Officer Yuhki Yamashita, a seven-year veteran, is departing entirely, with Chief Design Officer Loredana Crisan (who joined from Meta less than a year ago) absorbing the product organization. CMO Sheila Vashee is also leaving at the end of August, replaced by long-time Chief Communications Officer Nairi Hourdajian. Four senior leadership changes in a single quarter is a lot for any company to absorb, let alone one navigating a fast-moving AI product roadmap, and it’s reasonable for investors to ask whether this reflects normal post-IPO team evolution or something more disruptive to execution. Field characterized the moves in uniformly positive terms, but the concentration of departures — especially in product and marketing — is worth monitoring into the second half.
Unit Economics: The Real Test Is Still Ahead
Management’s answer on long-term gross margin cadence was candid about the mechanics: model-agnostic routing, cost optimization across providers, and increasing use of first-party models trained on Figma’s proprietary design corpus are the three levers being pulled to control inference costs as usage scales. Field noted the company is also still leaning on frontier lab models for certain use cases, particularly full design-to-code implementation, and expects that partnership to continue for now. Melwani was direct that margin will stay lumpy: “Over the long term, we expect additional usage to drive revenue and gross profit dollar growth,” but near-term margin will keep fluctuating as new beta products convert to paid tiers at different times. This is a business still calibrating its own pricing model in real time — this week alone, Figma began rolling out user-level credit limits to give enterprise admins more granular control, a direct response to customer feedback that they want predictability alongside consumption flexibility.