🌏 中文版
Deal Terms
| Item | Value |
|---|---|
| Company | Melius (New York, US) |
| Round | Series A (alongside a parallel seed) |
| Amount | $25M total ($20M Series A + $5M seed) |
| Lead | CRV (Series A), General Catalyst (seed) |
| Participants | Not further disclosed |
| Valuation | Undisclosed |
| Total raised | $25M (its first disclosed external funding) |
| Founded | 2025 (relaunched as a generative ad-creative platform out of stealth in July 2026) |
| Headcount | Not precisely disclosed |
What the Company Does
Melius builds AI-generated advertising creative — letting brands and agencies type a prompt and get a finished ad campaign, image, or video, rather than just a tool for optimizing where ad spend goes.
Its three co-founders, Joowon Kim, Young Kim, and Arnav Ramu, all worked as engineers at Ramp, the corporate spend-management company. Melius's original product was a performance-marketing tool for optimizing ad spend; after more than six months of building it, the founders decided the idea "didn't have legs" and scrapped it, relaunching from stealth in July 2026 with the generative creative platform it runs today.
Two months after the relaunch, annualized revenue had already passed $1 million. The category's biggest player is Higgsfield, valued at $5.4B with more than $700M in annualized revenue as of August 2026; smaller competitors include Krea and Flora AI. Kim says Higgsfield is "growing like mad," but believes the market is large enough to support several winners.
What This Round Signals
What It Means for the Agent Ecosystem
This round shows the generative ad-creative category has moved past "is anyone doing this" into "who can execute fast enough to grab share." Scrapping a product and still finding product-market fit and fresh funding within two months reflects how much AI tooling has compressed the product-iteration cycle.
What Investors Are Betting On
CRV and General Catalyst each leading a separate tranche of the same round — Series A and seed respectively — is an unusual structure, and it suggests investors are buying the team's execution and early revenue traction rather than the original idea, which didn't work out. "The product failed, but the team is still worth backing" is a pattern that shows up more often in AI tooling than in traditional SaaS, because rebuilding fast on top of a new generation of models is now so much cheaper.
Numbers Worth Watching
- Crossing $1M ARR within two months of relaunch is a faster start than most AI tooling startups at the same stage
- Compared to Higgsfield's three years to reach $700M+ ARR and a $5.4B valuation, the category already has a clear incumbent — Melius still has to prove it can differentiate rather than simply race on speed
- Pivoting away from a six-month-old product this quickly signals how much shorter the trial-and-error cycle has gotten for AI tooling startups
Watchlist Status
Melius is not currently on the watchlist. Recommend adding it to section C6 (Multimodal AI Generation Tools) — tracking how it differentiates from Higgsfield, Krea, and Flora AI, and how customer retention compares across the generative ad-creative category.
Today's Takeaway
A team that scrapped its product after six months still managed to raise $25M and hit $1M ARR within two months of relaunching — a sign that investors are judging AI founding teams less by "did you already validate product-market fit" and more by "can you rebuild something viable fast using the latest generation of models." Execution speed itself has become an investable signal.
References
Loading...