Poindexter Labs, an AI data company focused on improving the quality of training data used by frontier artificial intelligence models, announced it has raised £2 million in an oversubscribed seed round. The funding was led by Episode 1 and First Cheque, with participation from partners at Venture Friends and other angel investors.
Notably, the round also included investments from a group of Poindexter’s own contributors, including mathematicians and scientists who perform expert data annotation work on the platform and invested their personal savings into the company.
The funding comes as Poindexter signs its first direct contract with a major frontier AI lab, marking its transition from intermediary partner to named vendor. The company said it bootstrapped to $1.6 million in revenue during its first six months of operation before raising outside capital.
Founded by Jocelyn D’Arcy, a former mathematics teacher and former chief of staff at the National Mathematics and Science College, Poindexter Labs was created to address what the company describes as a structural problem in AI training data production. According to the company, frontier AI models require expert-generated reasoning data that traditional data-labeling platforms were not designed to produce.
Poindexter argues that many existing platforms were built for relatively simple tasks such as image labeling and text categorization, while modern AI systems increasingly depend on sophisticated, step-by-step reasoning across fields including mathematics, science, law, medicine, finance, and engineering.
The company has built a network of Olympiad medalists, PhDs, and professors from institutions including Oxbridge, Ivy League universities, and MIT. Contributors are admitted through a live subject-matter interview process and collaborate through a peer-review workflow designed to improve data quality and consistency.
According to Poindexter, its system delivers more than 95% of assigned data and achieves a 99.5% acceptance rate on completed work, compared with industry averages ranging from 5% to 60%.
The company’s proprietary platform, currently in beta, supports both Poindexter’s internal data production services and a software licensing model for enterprises and government organizations developing their own AI systems. The platform enables organizations to use collaborative workflows and peer-review processes similar to those employed by Poindexter’s contributor network.
Poindexter said the new funding will support team expansion, deeper relationships with leading AI laboratories, and the commercialization of its platform for enterprise and government customers. The company is scheduled to officially launch at Maths World in London on June 3, 2026.
KEY QUOTES:
“The workflow that has defined the industry since its inception was designed for a factory, not for knowledge creation. As a result, a huge chunk of training data is discarded not because it is wrong, but because adversarial review processes actively incentivise discarding tasks rather than improving them. This is a workflow problem, and we built Poindexter the way academics build knowledge: collaboratively, transparently, with peer review at every step. We have an accepted ACL paper, a growing research pipeline, and a commitment to building the kind of high-quality sovereign datasets the UK’s AI future depends on. Our proof point is our revenue. If we weren’t 5x more efficient than the platforms we replaced, we simply wouldn’t exist.”
Jocelyn D’Arcy, Founder and CEO, Poindexter Labs
“Poindexter has done something rare, built a business that is both technically differentiated and commercially validated from day one. The traction speaks for itself. We backed them because the problem is real, the solution is working, and Jocelyn is unlike any founder we have ever met.”
Adam Shuaib, General Partner, Episode 1
