loading...
WHICH IS A SIGN THAT DEMAND EXISTS.
“There are a lot of players”
Higgsfield reached $700M ARR and a $5.4B valuation by solving one problem: making AI video generation fast enough for professionals. It built a single-model interface, got the quality right, and scaled. The market signal is clear. What it did not solve – and what is now the obvious next problem – is what happens when a creative team needs to chain six different models together to produce a single campaign asset.
Melius is betting on that gap. The company, founded by ex-Ramp engineers, builds a node-graph creative canvas that lets users connect AI models into reproducible pipelines. Generate an image with GPT Image or Flux, pass it to Seedance for motion, add a voiceover via ElevenLabs, grade the output with a color LUT node – all within a single workflow that can be saved, forked, and run again for the next campaign variant. The canvas supports GPT Image, Flux, Seedream, Seedance, Veo, Kling, and ElevenLabs at launch.
The founders scrapped their first product – a different AI tool – and rebuilt around this architecture. The pivot appears to have worked. Melius came out of stealth and crossed $1M ARR within two months. The Series A, a $20M round led by undisclosed investors, closed in October 2026, bringing total funding to $25M.
The pricing is intentionally accessible: $17 per month for the Creator plan. That entry point is a distribution bet. At $17, individual creators and freelancers can self-serve, which seeds the workflows that creative directors later want to bring into team accounts. The model is reminiscent of how Figma priced against Sketch – cheap enough to spread, structured enough to upsell.
The founder acknowledged the crowded market directly: 'There are a lot of players – which is a sign that demand exists.' The honest read on this space is that it is crowded at the model layer and underdeveloped at the workflow layer. Single-model tools are commodity. Multi-model pipeline orchestration with a usable canvas interface is not.
For builders in adjacent spaces: the node-graph approach to AI creative production is not limited to general-purpose tools. E-commerce brands producing thousands of product image variants, game studios generating NPC dialogue and environmental art, real estate platforms automating listing media – each of these is a vertical version of the same workflow problem. The infrastructure Melius is building at the general layer will not be optimized for any of them. That is the opening.