AI startup DesignArena has raised $7.9 million in seed funding to expand a platform built around a simple idea: models can generate output, but people still define what feels genuinely good.
The company began in 2025, when co-founder Grace Li and a group of college friends were building an AI game engine. The system could produce playable games, yet the results lacked the spark that makes a game enjoyable. That gap pushed the team toward a new approach centered on human feedback at scale.
That experiment evolved into DesignArena, now used by 5.3 million people worldwide. Users compare outputs in an "A vs. B" format across websites, images, and other visual styles, helping the platform collect preference data that AI companies can use to refine their models.
The company's parent, Intelligence, said the new round was led by Index Ventures with participation from Conviction, A*, Valkyrie, and other investors. Li says the platform has become a valuable source of evaluation data for frontier labs, especially in design and media generation.
Beyond consumer use, the enterprise side appears to be the core business driver. Intelligence says DesignArena is generating $60 million in annual recurring revenue, while also tracking how design preferences vary across regions and over time.
As AI systems continue to improve, platforms that translate human taste into structured feedback may become a key layer in model development. The next phase of AI could be shaped not only by scale and speed, but by how well systems learn what people truly prefer.
