San Francisco-headquartered AI-as-a-Service provider Runware raised $13 million in its latest fundraising round led by Insight Partners. To date, the company has raised $16 million in funding. Participating investors include a16z Speedrun, Begin Capital, and Zero Prime.
Runware will use the funds to expand its capabilities from image and video generation to all-media workflows, including audio, LLM, and 3D.
“As more and more models launch, devs can have tens or even hundreds of endpoints to integrate with and maintain. We see model providers now moving to our platform and offering their APIs from our inference pod, because we can deliver up to 90% lower inference cost than any cloud provider,” said Flaviu Radulescu, Founder at Runware.
From PicFinder to Runware: The origin story
Runware was founded in 2023 by Flaviu Radulescu and Ioana Hreninciuc, as an evolution from PicFinder, an image creation tool that aimed to speed up image generation to match the speed of image searches. Flaviu developed PicFinder after seeing the potential of StableDiffusion, a technology he considered potent.
“The product was picked up on social media, became viral, which showed the incredible need for a responsive AI experience. The inference engine powering it was the first iteration of Runware. The pivot to Runware happened after realising that the real value was in enabling more companies to access incredibly fast and cost-effective media generation through a platform – rather than just building a single image generator,” explains Ioana to TFN.
Currently, the company employs approximately 35 people remotely from 15 different nationalities. “Our gender split is 17%/83% F/M. In terms of racial diversity, ~30% of the team is of non-white ethnicity,” reveals Ioana to TFN on diversity statistics inside the company.
Building affordable AI infrastructure
As compute-intensive workloads, such as video generation, gain popularity and GPU costs consume budgets, consumer AI apps are increasingly seeking to reduce costs. The US company simplifies this by offering a single API that combines top media generation models, providing fast performance and costs up to 10 times lower than other options.
“Runware makes it possible for companies to make their products fully AI-enabled in hours instead of months,” says Ioana.
Runware can optimise hardware fundamentally, thanks to Flaviu Radulescu’s 20 years of experience in building bare-metal data clusters for clients such as Vodafone, Booking.com, and Transport for London.
Additionally, the company’s AI media generation API utilises the Sonic Inference Engine, which combines custom hardware and specialised software to enhance efficiency and speed. Runware’s API also simplifies the integration of new models by standardising data across different providers, which allows engineering teams to add models quickly with just a parameter change.
Additionally, Runware designs its own GPU and networking hardware within a proprietary inference pod. This setup is designed for quick deployment and utilises renewable energy.
Expanding beyond image and video
Runware is investing heavily to improve its inference engine and API for all types of AI media work. The company has integrated various image and video models from sources like Black Forest Labs, OpenAI, ByteDance, and Google Veo, and is also expanding into audio and large language models (LLMs).
With Runware’s API, users can quickly create a media generator or content creation tool in just minutes. By supporting all forms of media generation, Runware makes AI integration much easier. Most importantly, its API eliminates the need for multiple model integrations and costly infrastructure, enabling product teams to launch AI media features on the same day without complex setups.
“Runware aims to power the world’s intelligence. To have all AI workloads powered by the Sonic Inference Engine and be a one-stop shop for any company building with AI – while being 10x faster and more affordable than any alternative,” added Ioana.
Women in AI: Ioana’s perspective
Women in the tech industry, particularly in AI, face both challenges and significant opportunities. The AI sector is highly competitive, with innovations occurring much faster than in other fields, requiring founders to maintain a diverse skill set.
“The speed at which you need to operate as a founder, and the range of skills required, are incredibly high. To join the sector, you have to be willing to put everything else on hold or in maintenance mode for a while. The gains are incredible, but so are the demands,” says Ioana.
Despite these demands, the AI landscape presents an opportunity for women to excel, enabling them to build generational companies that reflect their values and vision.
“This should not deter capable women from joining the sector – because it’s also the best opportunity for building a global scale company in record time,” concludes Ioana.
