Glonce

Design Arena raises $7.9 million seed round for AI feedback

businessThe Glonce editorsTechCrunch🔥66Glonce rating66Impact65Novelty50Relevance90Surprise50Players50Credibility80
Design Arena raises $7.9 million seed round for AI feedback

As co-founder Grace Li tells it, her company started a few weeks before graduation in 2025, when a handful of college friends were trying to make their AI game engine work. The models could produce functional games, but none of them were fun, which raised a harder question: how do you tell if a game is fun at all? The team decided there was no substitute for human judgment and began looking for ways to gather honest human feedback at scale. That effort became Design Arena, an AI tool now used by 5.3 million people worldwide.

It turned out plenty of AI companies wanted scalable user feedback and were willing to pay for it. "It was the missing bottleneck for a lot of these models to make improvements in the design space," Li says. "About a week later, we closed our first major deal with a frontier lab, and the rest is kind of history."

On Monday, the company behind Design Arena, called Intelligence, announced a $7.9 million seed round led by Index Ventures, with participation from Conviction (Sarah Guo and Mike Vernal), A*, Valkyrie and others.

For non-enterprise users, Design Arena works like a sophisticated model router: a ChatGPT-style prompt window with separate dropdowns for websites, images and a dozen other visual formats. After entering a request, format and style, users are shown a series of "A vs. B" choices until they have ranked the outputs from best to worst. The real value, though, is on the enterprise side, where participating AI labs treat the platform as a source of endless instant feedback for their media-generating models. Users tend to be indifferent to which model produced which output, Li says, so their rankings give the labs a clean read on what people actually want.

Li says the site is currently generating $60 million in annual recurring revenue, which she frames as solidifying its position as a key source of human-led evaluation data for the AI industry. Because users have to log in to see their results, Intelligence can also track how tastes shift across continents and over time; Li notes that web dashboards designed in Asia tend to favor a more maximalist style. This human-judgment layer serves as an important complement to automated benchmarks, which can run at greater scale but are more exposed to gaming or manipulation, as a Hugging Face breach demonstrated the previous week.

Crowdsourced human feedback is not an automatic path to a sustainable business, however. Yupp shut down earlier this year, less than a year after launching, despite having raised $33 million from a16z crypto's Chris Dixon and signing up some frontier-model customers and, by its own account, over 1.3 million users. By contrast, LM Arena, which applies a similar approach to text-based responses, raised $150 million in a Series A in January, just four months after formally launching its paid product.

Key facts

  • Intelligence, the company behind Design Arena, raised a $7.9 million seed round led by Index Ventures, with Conviction (Sarah Guo and Mike Vernal), A*, and Valkyrie also participating.
  • Design Arena has 5.3 million users and, according to co-founder Grace Li, is generating $60 million in annual recurring revenue, largely from frontier AI labs paying for feedback data.
  • Non-enterprise users rank AI-generated outputs, such as websites and images, through 'A vs. B' comparisons; AI labs use the resulting rankings as feedback for their media-generating models.
  • Rival Yupp shut down earlier this year despite raising $33 million and reaching over 1.3 million users, while another competitor, LM Arena, raised $150 million in a Series A round in January.
  • Human evaluation serves as a complement to automated benchmarks, which scale further but are more vulnerable to gaming, as shown by a Hugging Face breach from the week before the announcement.

Why it matters

AI labs increasingly need scalable human judgment on qualities, like design taste, that automated benchmarks struggle to measure and that can be gamed, as the recent Hugging Face breach showed. Design Arena turned that need into revenue, growing out of a failed AI game-engine project into a business Li says generates $60 million in annual recurring revenue, mostly from frontier labs paying for feedback on their models' outputs.

Who it affects

Frontier AI labs building design and media-generating models get a paid channel for structured human feedback. Everyday users get a free-to-use tool for comparing AI-generated websites, images and other visual formats. Investors backing the round include Index Ventures as lead, plus Conviction (Sarah Guo and Mike Vernal), A* and Valkyrie. The story also touches competitors in the same space: LM Arena, which is thriving, and Yupp, which folded.

How to use it

Non-enterprise users open a ChatGPT-style prompt window, pick a format such as websites or images from a dropdown of about a dozen options, submit a request, and then rank a series of 'A vs. B' outputs from best to worst. Enterprise customers, mainly AI labs, pay to plug their models into Design Arena and receive the resulting rankings as feedback data. The source gives no pricing details for either tier.

How solid is it

The round is led by a named institutional investor, Index Ventures, alongside recognizable individual backers Sarah Guo and Mike Vernal at Conviction. The $60 million ARR figure and the 5.3 million user count both come directly from co-founder Grace Li in the article and are not independently verified. A comparable company, LM Arena, raised $150 million in a Series A four months after launching its paid product, lending some support to the category, though a direct competitor, Yupp, failed despite similar early traction.

Risks and caveats

The $60 million ARR and 5.3 million user figures are self-reported by the co-founder, not audited. The article does not name which frontier labs pay for Design Arena's data, nor does it give Intelligence's total funding to date, valuation or headcount. It also does not explain what the Hugging Face breach involved beyond calling it an example of benchmark gaming. The category itself is unproven: Yupp raised $33 million and reported over 1.3 million users before shutting down less than a year after launch, having failed to build a sustainable business.

“It was the missing bottleneck for a lot of these models to make improvements in the design space.”

— Grace Li, co-founder of Intelligence