AI Design Evaluator DesignArena Raises $7.9M Seed Round With $60M ARR

Intelligence, the company behind the DesignArena platform, has raised $7.9 million in seed funding to scale human evaluation for AI-generated designs. The round was led by Index Ventures, with additional participation from Valkyrie, A*, and Conviction, specifically Sarah Guo and Mike Vernal.

DesignArena utilizes a crowdsourced model where 5.3 million users worldwide rank AI outputs through "A vs. B" head-to-head matchups. This preference data allows AI labs to move beyond basic functional correctness toward a more nuanced, human-centric aesthetic. Despite its early seed stage, Intelligence reports a significant commercial footprint with $60 million in Annual Recurring Revenue (ARR).

The Bottleneck of AI Taste

The platform operates as a visual benchmark similar to how LM Arena functions for text-based large language models. For frontier AI labs, the primary challenge has been the lack of scalable, honest human feedback to improve the aesthetic and functional quality of visual outputs. Automated benchmarks often fail to capture the subjective nature of design, leaving a gap in how models are refined for real-world appeal.

Grace Li, co-founder of Intelligence, noted that this human-led evaluation "was the missing bottleneck for a lot of these models to make improvements in the design space." By capturing these preferences, the platform provides a safeguard against the limitations and potential manipulation of purely automated scoring systems.

Industry Implications

The rapid growth of DesignArena underscores a massive, underserved market for high-quality preference data. The disparity between the company's seed funding and its $60 million ARR suggests that AI labs are willing to pay a premium for data that can objectively quantify "taste."

Furthermore, the platform is uncovering critical regional variations in design. Intelligence has noted that design preferences differ by geography, observing that Asian web dashboards often favor a more maximalist style compared to other regions. This suggests that the next generation of AI design tools will likely need to be culturally adaptive rather than relying on a single global aesthetic standard.

What's Next

Founded in 2025 by a group of college friends who originally sought to make an AI game engine "fun," Intelligence is now pivoting toward the broader infrastructure of AI evaluation. As frontier labs race to release more sophisticated image and UI generators, the demand for human-in-the-loop verification is expected to grow. Observers will be watching to see if DesignArena can maintain its growth trajectory and whether its regional preference data becomes a standard for localized AI design deployment.