Runware raises $50M Series A to help make image, video generation easier for developers

Flaviu Radulesc started Runware successful 2023 erstwhile helium was investigating a text-to-image institution and realized that, though genAI tech was powerful, it was dilatory successful generating images. So Radulesc teamed up with Ioana Hreninciuc and launched Runware arsenic a dev instrumentality level that specializes successful generating images, videos, and audio successful existent time.

The institution has seen overmuch maturation since it archetypal launched. It has powered much than 5 cardinal creations for much than 100,000 developers, the institution told TechCrunch.

The merchandise lets developers integrate Runware’s API into their apps and past make media assets done 1 interface, truthful they don’t person to acceptable up immoderate caller infrastructure oregon support abstracted integrations. It has a customized AI inference infrastructure for open-source models, and provides day-zero entree (meaning arsenic soon arsenic a exemplary is released, it tin tally connected Runware) and competitory pricing, Hreninciuc, who focuses connected operations and GTM, told TechCrunch.

On Thursday, the institution announced a $50 cardinal Series A successful a circular led by Dawn Capital. Dawn Capital Partner Shamillah Bankiya is joining the board. Others successful the circular see Insight Partners and a16z Speedrun. Runware has raised $66 cardinal successful backing to date.

Hreninciuc said the institution remains competitory done its pricing, which is “more-cost effective,” successful summation to having a afloat unified API. She said the institution does this with its Sonic Inference Engine, which runs connected customized AI hardware. It besides partners with third-party AI unreality providers truthful that it tin reroute workloads automatically successful lawsuit much representation is needed.

“On the bundle side, we heavy optimize exemplary loading and offloading, which lets america enactment implicit 400,000 models and marque immoderate of them disposable for inference successful existent time,” she continued.

Startups that absorption connected dev tools for representation and video person been a peculiarly blistery marketplace for VC involvement lately. Fal.ai, for instance, conscionable raised $140 cardinal astatine a $4.5 cardinal valuation, its 2nd elephantine rise successful a substance of months. Fal.ai focuses connected the breadth of exemplary offerings versus customizing for speed. So, Hreninciuc considers her competitors to beryllium wide open-source exemplary makers, specified arsenic Hugging Face; Replicate, a startup that runs open-source models successful apps with conscionable a fewer lines of code; and Together AI, a akin genAI exemplary hoster.

These companies, arsenic Radulescu antecedently told TechCrunch, merchantability based connected GPU compute time. Runware alternatively leans toward the exemplary of Stable Diffusion and Flux, offering a much cost-per-image generated truthful that radical tin wage for what they request alternatively than buying a artifact of compute time.  

Hreninciuc said the caller superior volition beryllium utilized to support expanding the company’s infrastructure and it hopes to usage its Sonic Inference Engine to powerfulness implicit 2 cardinal models. The large extremity is to beryllium the API for each AI — truthful that immoderate generative AI exemplary tin and does tally connected the platform.

“We’re besides expanding rapidly into caller modalities,” she said, adding that the institution is going to grow the existent squad of astir 25 to assistance marque that happen.

Overall, Hreninciuc hopes Runware continues to “make it imaginable for applications to standard to millions of users portion really keeping their margins,” she said, adding that it helps marque the marketplace much affordable. That “benefits everyone,” she continued. “From the app builders to the extremity users, and puts almighty AI into much people’s hands globally.”