Paris has quickly established itself as Europe's leading hub for AI startups, and a new deal now in the works could further cement that position.
Poolside.ai, a Paris-based generative AI company that's making tools to speed up software development, is raising at least $400 million at a post-funding valuation of $2 billion, sources told TechCrunch.
Bain Capital Ventures and DST are currently in talks to co-lead the round, according to sources. BCV is a previous investor in the company, while DST is a new investor. Bain's participation was previously reported, with PitchBook noting that this is an estimated $450 million round.
Poolside made headlines last August when it raised a big seed round as an AI startup in London. It raised $126 million from investors including BCV and early-stage specialists such as London's Air Street, Abstraction, Scribble Ventures, France's New Wave and Frst. Bpifrance, Felicis, Point Nine and Redpoint also participated in the round. None of the investors we contacted would comment on the matter. Poolside's CEO did not respond to a request for comment.
Foundational model companies Mistral and H are among those to have raised nine-figure seed rounds from New York City ($113 million and $220 million, respectively). At this rate, the City of Light may need to be renamed the City of AI.
Looking at the bigger picture, it can sometimes feel like the market for AI startups is heating up too quickly — with companies like Anthropic and OpenAI raising billions of dollars collectively — and you might be wondering, do we really need another foundational AI company?
There are several reasons why Poolside secured this level of funding to take on the generative AI opportunity on its own big spurt.
– The strength of the founding team and how that ties into the company premise. Both are deeply immersed in the world of developer tools and DevOps. One of the founders, CEO Jason Warner, was CTO at GitHub and led engineering at Heroku and Canonical. The other CTO, Eiso Kant, previously founded Athenian, a set of tools to help developers optimize how they build and work (the company was acquired by the Linux Foundation for an undisclosed amount). Warner was also a former VC at Redpoint, so he knows the value and language of VC-founder interaction.
– The problem it's trying to solve. Unlike companies building foundational LLM models that are more generalized in their approach, Poolside (at least for now) is addressing one use case in particular: making developers faster. This will resonate with investors, and is reminiscent of a memorable essay written by Paul Graham many years ago about the startup idea: building the tools you know you need (that is, the tools that founders and other technologists need).
Companies like Mistral are also focused on developers and developer tools, not to mention Microsoft's adaptation of OpenAI on GitHub (indeed, OpenAI is centered around APIs that developers also use). But there are some notable examples where code is (at least for now) one of the more problematic blind spots for LLM more generally. There is an opportunity now to build on this, and build something better than the more common approaches.
That's not to say Poolside isn't thinking big, either: in a three-phase plan outlined on the company's website, they hope to eventually work with developers, then anyone who wants to write code or software, and then “generalize these capabilities to all areas outside of software.” No big deal!
– an early sign of what they’re building. We haven’t confirmed if the company has released a public-facing product yet, but there is some evidence that they’re active and growing (and possibly running in private beta). Their computing supplier, IREN, reported in April that Poolside had strengthened its cloud services agreement with the company.
– And last but not least, there’s the issue of monetization. A source close to the company told me that while there are plenty of examples of AI working on various segments of the market, not many offer clear indications of, much less evidence of, monetization potential. (Indeed, OpenAI’s big deal with PwC could open the door to more enterprise business, but that could be a multi-year effort, and it remains to be seen what enterprise adoption will look like in the meantime.)
But in Poolside's case, given the opportunity and scale of building Copilot tools for developers, if done right, it could be one of the best and easiest places to apply AI. It's a clear and big need, and because all computer programming has syntax, it's not as limiting as many other areas where AI is being applied. Plus, it has constraints, performance metrics, and benchmarks, making it an attractive combination for investors and startups looking to build a business.
We will update this post as more information becomes available.