Replicate, a startup that runs machine learning models within the cloud, today launched out of stealth with $17.8 million in enterprise capital backing; $12.5 million of the overall got here from a Series A led by Andreessen Horowitz with participation from Y Combinator, Sequoia and angel investors including Figma CEO Dylan Field and Vercel’s Guillermo Rauch, while the remainder was from a previously undisclosed seed round.

The corporate was co-founded by Ben Firshman, who led open source product efforts at Docker, and Andreas Jansson, previously a machine learning engineer at Spotify. The way in which Firshman tells it, he and Jansson got here to the mutual realization that AI was accelerating at an “absurd” pace, but that technical barriers were standing in the best way of mass adoption.

Enter Replicate, which offers a library of open source models that software developers can run with a number of lines of code. The platform can robotically generate an API server for custom machine learning models, deployed on a big cluster of GPUs.

“When you get a ton of traffic, we scale as much as handle the demand. When you don’t get any traffic, we scale all the way down to zero and don’t charge a thing,” Firshman explained. “We only bill you for a way long your code is running. The choice is generally deploying models yourself on Amazon Web Services. Typically, you’d must battle with servers, Kubernetes, GPUs, API servers, auto-scaling and more.”

Core to Replicate is Cog, an open source tool that lets developers package machine learning models in a normal, production-ready container format. Firshman and Jansson developed Cog, which runs on any newer macOS, Linux or Windows 11 machine.

“AI is currently too hard to make use of for software engineers and you could have to be a machine learning engineer to make use of it,” Firshman said. “Corporations and the industry as a complete is being held back by the dearth of machine learning experts. We’re making it possible for software engineers to make use of machine learning with zero experience, with just a number of lines of code, so that they can construct products with AI and apply it to business problems.”

Replicate isn’t the just one doing this. The startup competes with vendors including Hugging Face and OctoML (and to an extent Runway ML), which collectively have raised a whole lot of tens of millions in enterprise capital. Google, Amazon and Microsoft may very well be considered rivals, as well — offering their very own solutions for developing, launching and maintaining machine leaning models within the cloud. (See SageMaker, AutoML and Azure’s no-code ML tools).

So what sets Replicate apart? Firshman claims the developer experience is “a lot better,” which after all stays to be seen — in spite of everything, Replicate is brand-spanking latest. One clear point of differentiation, though, is the expansiveness of Replicate’s AI library. The platform offers diffusion models including Stable Diffusion, models for creating and editing videos, upscaling models for images and various image-to-text and text-to-image models.

Swift, painless deployment is the main focus. Replicate’s website guarantees: “With Replicate and tools like Next.js and Vercel, you possibly can get up with an idea and watch it hit the front page of Hacker News by the point you go to bed.”

The marketing appears to be resonating with the developer community, which has enthusiastically embraced Replicate over the past few months — at the very least in keeping with Firshman. He says that the platform’s seen 149% month-over-month growth in energetic users and 125% growth in API calls for the reason that middle of last yr. Enterprise customers include Character.ai, Labelbox and Unsplash.

“We’ve effectively been indexing the expansion in generative AI,” Firshman said. “Founders are constructing tons of latest products, investors are investing in it and users are clamoring for all these latest things.”

Leaning into generative AI is actually a sensible decision on Replicate’s part. The segment — under which technologies like ChatGPT and Stable Diffusion fall — has seen a large uptick in investment over the past several years. PitchBook (via Bezinga) reports that VCs funneled 425% more dollars into generative AI in 2022 in comparison with 2020, with the space reaching $2.1 billion total capital pledged in 2022.

Firshman sees the expansion continuing — and Replicate benefitting.

“It hasn’t yet entered the enterprise’s consciousness how much generative AI goes to upend so many parts of their business: customer support, marketing, sales, content creation, and doubtless other things we haven’t anticipated yet,” he said. “Very soon, customer support will probably be mostly automated and intensely good — not the terrible chatbots of the past. Creating assets for marketing will probably be mostly automated. Many of the ads you see will probably be robotically generated and personalized. Creating assets for video games will probably be mostly automated. And that is with the technology we’ve today.”