Sequoia backs Pydantic to expand beyond its open source data-validation framework

A U.K.-based, open-source startup is launching its first commercial product with the backing of one of Silicon Valley’s most renowned venture capital firms.

Pydantic on Monday launched an observability platform called Logfire, five months after trialing it in open beta, and announced $12.5 million in Series A funding led by Sequoia.

However, the company is better known for its eponymous Python library and open source data-validation framework, started by U.K developer Samuel Colvin back in 2017. The project has gone from strength to strength, and is now used by developers at some of the world’s biggest companies including Meta, Nvidia, Netflix, Google and OpenAI.

Companies deploy Pydantic within applications that need to verify the type of data a user has entered — if a form requires an email address and the user instead inputs a phone number or leaves it blank, Pydantic checks this and delivers a user-friendly error message. It basically validates data structures to ensure integrity and has myriad use-cases.

For example, ChatGPT maker OpenAI introduced structured outputs for its API in August, and this feature uses Pydantic under the hood. So if a company wants to develop a chatbot that collects user details and returns them in a structured manner so the data can be easily processed by the system, it would use Pydantic.

“Where Pydantic is exciting is that it is the default way of validating the response from an LLM,” Colvin told TechCrunch in an interview last week. “So if you want to do structured output, that’s how you do it.”

Colvin launched Pydantic as a commercial entity in 2022, emerging from stealth 18 months ago with $4.7 million in seed funding from Sequoia. And it seems it’s now time to start making money — indeed, Colvin said the company is, effectively, looking to “cash-in on our credibility and our brand name,” using Pydantic as the carrot-on-a-stick for other products, rather than building on Pydantic itself.

Tried and tested

The typical trajectory for a startup building an open-source business looks something like this: Create an open source product that solves a real problem; that product gains traction with developers, becoming an indispensable tool in their stack; the startup creates commercial services and features on top of the core open-source project to make it even more useful.

It’s a tried and tested model, but the problem is that businesses are increasingly retreating from open source in one form or another, whether that’s transitioning to a less-permissive license as Grafana did, or abandoning it altogether like HashiCorp did. The reasons are generally the same — it’s all about protecting the company’s bottom line, ensuring that larger companies don’t take advantage of a product’s open source credentials.

There’s even a whole new licensing paradigm emerging to tackle the “use and abuse” problem in open source. Billion-dollar developer tooling company Sentry is pushing the concept of “fair source,” as it seeks to align itself with “open” software without actually going open source. “Open source isn’t a business model — open source is a distribution model, it’s a software development model, primarily,” Sentry’s head of open source, Chad Whitacre, told TechCrunch in an interview last month.

While using open source to ingratiate a company to the developer community is far from a novel concept, Pydantic is slightly unusual in that it’s using its open-source project entirely as a marketing tool. So rather than trying to transform Pydantic itself into a commercially viable product, it’s leaning on the project’s gravitas to sell other, not-directly-related products instead — such as Logfire.

“Instead of building the hosted version of Pydantic, the library, we’ve built Logfire, the observability platform,” Colvin said. “The trust that we have as a company from the Python community is in a different league to many other companies. We went to PyCon US this year just after we announced Logire in beta, and our booth had a cluster of people around it all week because everyone knew the library and they knew us. Whereas, if we had turned up as a brand new observability company, people would have ignored us. Pydantic is a better-known brand than almost any other in the Python world, other than the big guys like AWS and Google.”

Logfire is basically a Datadog competitor, designed to give developers insights into how their software is performing. But Pydantic wants to make the whole observability process simpler to configure. It wants to be “to Datadog what Vercel is to AWS,” as Colvin put it.

“AWS has an enormous amount of functionality and it’s incredibly complex to use,” he said. “Datadog is also an enormously complex piece of kit, so we’re trying to build a simpler experience for developers. Longer term, we want it to be so that you could go and use this [Logfire] in place of Datadog. But in the medium term, we want to be that simpler solution for smaller teams.”

It’s certainly an interesting approach to building a business — the startup is essentially using Logfire to solve a different problem for the same people that use Pydantic.

“They are different things, but where they overlap is that all the people who need Pydantic, the validation library, also need observability,” Colvin said. “So we’re targeting a solution for the same people.”

Show me the money

Back in the earlier days of Pydantic, Colvin managed to secure some decent sponsorships from some of the framework’s biggest corporate users, including Salesforce, which donated $10,000 in 2022; AWS and GitHub sponsored $5,000 and $750, respectively.

But as the business has grown and VCs have entered the picture, corporate donations have grown less frequent.

“We’ve had reasonably generous sponsorships, but more so when I was working on my own,” Colvin said. “But now that we’re backed by Sequoia, people are less handy with their wallet!”

With Logfire now in general availability, Pydantic hopes to build on the 2,000-plus developers and 150 companies it attracted during the beta phase. It now has a heavy focus on AI companies.

Aside from lead investor Sequoia, Pydantic’s Series A round saw participation from Partech and Irregular Expression, alongside angels such as Logan Kilpatrick and Jason Liu. Colvin said the fresh cash will be used primarily for salaries, and to bolster its existing headcount of 13, which are spread around the U.S. and Europe.

“We’ll use the funds for hiring, mostly developers,” Colvin said. “We’ll probably hire for sales at some point, but for now, it’s just engineering.”