Substrate closes $8M round to create an API to accelerate AI adoption

Substrate Labs, an artificial intelligence infrastructure startup founded by a team from Stripe, Pulumi and Robust Intelligence, announced that it has raised $8 million to create elegant APIs that will enable developers to build more modular artificial intelligence systems, accelerating a new era of software development. The round was led by Lightspeed Venture Partners, with participation from South Park Commons, Craft Ventures, Guillermo Rauch (Vercel), Immad Akhund (Mercury), Will Gaybrick (Stripe) and others. Led by co-founder and CEO Rob Chung and co-founder Ben Guo, the company also announced the official launch of its API optimized for model ensembles.

Substrate raises $8M to build a new modular AI API. "The toolkit we've created will allow any company - not just tech giants - to run dozens of branching ML models within a single query," says Rob Cheung, CEO.

"The artificial intelligence market will inevitably move from training models to using inference to make the models work - a huge market that is too inefficient today," said Nnamdi Iregbulem, Partner at Lightspeed Venture Partners. "Rob and Ben are strong engineers with over a decade of experience developing APIs and scaling systems at organizations like Stripe, Substack, and Venmo. They are able to abstract away the complexities of operationalizing the right AI model for each task, allowing engineers to optimize development with sophisticated tools previously only available at large tech giants."

Developers implementing AI today have the most success when they break the problem down into tasks and implement the best model for each task. However, the vast majority don't have access to the back-end infrastructure to run dozens of AI models in modular pipelines, so they either try to use one big AI model for each problem or combine solutions from different API providers and GPU services. Both approaches lead to unreliable, slow, and expensive AI applications due to low GPU utilization.

In line with Stripe's developer-focused strategy, the Substrate team is developing a powerful AI API that leverages curated open-source AI models across multiple modalities such as text, images, audio, and semantic vectors, and optimizes these models to work at scale - with a particular focus on how quickly the output of one model can be used as input to another. Substrate's underlying infrastructure is deeply integrated with SDKs that make it easy to create "graph relationships" between the interd Substack and Maven are among the first customers.

"New AI models have tremendous potential, but the tools developers have access to today hinder the adoption of intelligence in software systems," said Rob Cheung, co-founder and CEO. "We built Substrate to help developers build AI-integrated software intuitively, just like they build any other software: by linking small semantic tasks together to automate work. The toolkit we've built will allow any company - not just tech giants - to run dozens of branching ML models within a single query."

"We chose Substrate as a partner because we knew they would allow us to quickly leverage the power of modern ML models without committing significant engineering resources," says Jairaj Sethi, CTO of Substack. "We've already integrated Substrate into our internal systems for categorizing and recommending content, as well as our tools for creators, and we can't wait to continue exploring new use cases."

"Over the past year, we've explored various tools to provide course recommendations in Maven," said Shreyans Bhansali, CTO of Maven. "Substrate has been a breath of fresh air for us because their product is so clearly focused on providing the simplest possible developer experience."