Revenue jumped 48%. AI credits monetized for a full quarter for the first time. Then the next morning, Figma's stock dropped over 16%.

3-Second Summary
Revenue +48% First full quarter of AI monetization Stock -16% Culprit: R&D +101% How to bet on AI without losing margin

Everyone Believes This

The formula for tech earnings season is simple, right? If a company shows AI features turning into revenue, the market rewards it. Figma's Q2 2026 (April–June) results looked like the textbook example.

Revenue hit $370.1 million, up 48% year-over-year, marking the third straight quarter of accelerating growth. More importantly, this was the first full quarter where AI credit monetization actually showed up in results. Customers with $10K+ ARR grew 34% to 15,964, and over 80% of them were consuming AI credits weekly. Net dollar retention — the metric that shows how much customers expand their contracts — held strong at 136%. Figma even raised full-year guidance by $40 million. By every conventional signal, this was "AI is finally paying off."

The Numbers Say Otherwise

Except the day after earnings, Figma's stock dropped over 16% in after-hours trading. Despite beating estimates.

+48%
Revenue growth (YoY)
16%→10%
Non-GAAP operating margin (QoQ)
+101%
R&D cost growth (YoY)

The culprit was cost. Non-GAAP operating margin fell from 16% the prior quarter to 10%, and on a GAAP basis Figma posted a $117.3 million operating loss. R&D spending jumped from $83.1 million to $167.3 million — up 101% year-over-year. Revenue grew 48%; costs grew more than twice as fast. Sales and marketing rose 59%, general and administrative rose 169%.

What the market expectedWhat actually happened
Revenue-to-stock linkProven AI monetization = stock up+48% growth, stock -16% anyway
Operating marginMonetization starting = margin improvesFell from 16% to 10%
Cost structureRevenue growth offsets cost growthR&D +101%, double the pace of revenue

Here's What's Really Going On

Trefis summed it up best in one headline — "Figma is paying for the products it has not started charging for".

Figma Agent, Code Layers, Motion, and generative plugins — most of the new tools unveiled at Config 2026 are still in beta or early access. They don't consume paid credits, so they don't show up as revenue. But the AI inference bills to run them are 100% on Figma's books. CFO Praveer Melwani admitted as much on the earnings call, saying the company is "absorbing inference costs for beta-stage products without passing them to customers". Revenue books today, but costs get charged for tomorrow's products — that timing mismatch was the real trap in this earnings report.

There's a second layer here too. Figma leans on outside AI models from Anthropic and OpenAI — the same companies that are becoming its competitors. Anthropic's Mike Krieger stepping down from Figma's board right after Claude Design launched is a telling sign. Figma is responding with a model-agnostic strategy (Weave) plus first-party model investment, and its CodeConnect feature has already cut inference costs by roughly 30%.

Protect Your Margin First

The lesson from Figma is clear: shipping an AI feature and managing that feature's cost inside your margin are two completely different problems. If your team is investing in AI features, here's how to protect margin along the way.

  1. Set a hard date for monetization
    Decide upfront how long a beta feature stays free. Leave it open-ended like Figma did, and costs just keep piling up.
  2. Build a cost dashboard per feature
    Track inference cost against actual revenue contribution for each feature separately. Total revenue alone won't show you which feature is bleeding money.
  3. Cut costs with model routing first
    Don't send every request to your most expensive model. Figma's CodeConnect routes requests smartly and cut inference costs by 30%.
  4. Spread your vendor risk
    Your AI supplier today can be your competitor tomorrow. Don't bet everything on one model — keep an alternative path open.
  5. Watch usage before the financials do
    Weekly active usage and credit consumption are leading indicators. Check them before margin starts slipping.

Go Deeper

Figma's Q2 2026 earnings release The primary source for every official revenue, margin, and guidance figure investor.figma.com

StockTitan — detailed cost breakdown Line-item growth rates for R&D, sales/marketing, and G&A stocktitan.net

Trefis — "paying for products not yet charged for" The sharpest analyst take on this quarter's core contradiction trefis.com

the-decoder — Config 2026 and vendor risk How Figma's own AI suppliers are turning into competitors the-decoder.com

Full earnings call transcript Read CEO Dylan Field and CFO Praveer Melwani's answers verbatim investing.com

TechCrunch — how Figma Agent launched The origin story of the feature now at the center of the cost debate techcrunch.com