Figma bets on human judgment at Config 2026 while the AI powering its canvas belongs to someone else

2026-06-24 · Source: The Decoder · Field: Technology & Digital — Software Development & Engineering, Artificial Intelligence & Machine Learning, Project & Product Management · Depth: Intermediate, medium

Summary

At Config 2026, Figma unveiled significant updates to its design canvas, expanding its capabilities to integrate code, animations, 3D effects, and AI agents. These enhancements, including Figma Make, Code Layers, Motion, Depth Layer, and WebGPU-powered Shaders, aim to keep users within the platform as AI competitors like Anthropic and OpenAI increasingly offer direct interface generation tools. The company also introduced collaboration features for sharing AI prompts, workflows, and custom plugins. Despite revenue growing 46% and crossing \$1 billion in fiscal year 2025, Figma faces margin pressure from high inference costs, with gross margin falling from 92% to 86% in 2025 due to reliance on external AI models from providers like OpenAI and Anthropic. Figma is addressing this by focusing on efficient token usage, such as CodeConnect, which cuts inference costs by approximately 30%. CEO Dylan Field emphasized human judgment and avoiding lock-in to single AI providers, rather than developing proprietary foundation models.

Key takeaway

For product designers and AI product managers evaluating design tool ecosystems, Figma's Config 2026 updates signal a strategic shift towards integrating AI as a collaborative, tweakable tool within the canvas, rather than a black-box generator. You should prioritize platforms that enable human judgment and efficient AI interaction to control costs and maintain design quality. Consider how your team can utilize shared prompts and custom AI skills to streamline workflows and reduce reliance on external, costly AI inference.

Key insights

Figma integrates AI into its design canvas to retain users and manage costs, emphasizing human judgment over proprietary foundation models.

Principles

• AI should be a tweakable tool, not a final output.
• Human judgment remains critical in AI-driven design.
• Efficient token usage can mitigate high inference costs.

Method

Figma integrates AI generation directly onto the canvas via Code Layers, Motion, Shaders, and Weave, allowing designers to tweak AI outputs and share prompts/skills collaboratively.

In practice

• Use Code Layers for side-by-side design and code.
• Share successful AI prompts and custom plugins.
• Implement CodeConnect to reduce inference costs.

Topics

• Figma Config 2026
• AI-powered Design
• Design-to-Code Workflows
• AI Inference Costs
• Collaborative AI Tools
• WebGPU Shaders

Best for: Investor, CTO, VP of Engineering/Data, AI Product Manager, Product Designer, Director of AI/ML

Related on AIssential

• Figma adds code layers, support for animations, more AI features in new update — Figma's update deepens design-code integration and expands AI for asset creation, workflow automation, and rapid prototyping.
• Exclusive: a conversation with Tibo from Codex on what your company has to become when the model can actually do the work — Advanced AI models shift the core challenge from capability to strategically embedding human judgment in software development.
• AI Didn’t Kill Design—It Exposed It — AI reveals how product design often prioritizes corporate incentives over genuine user needs.
• Figma Is Doomed — AI-powered design tools are rapidly disrupting established platforms like Figma, impacting market valuation and competitive landscapes.
• 🔵 Claude Code to Figma and the new era of Product Taste — Product taste and context curation are emerging as critical differentiators in an AI-driven product development landscape.
• Turning Figma designs into code collapses the handoff gap — Automating Figma-to-code conversion drastically reduces design handoff friction and development costs.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by The Decoder.