Safe Sign Thomson Reuters deal shows pre-revenue AI value, founder says

The Safe Sign Thomson Reuters deal put a pre-revenue AI research startup inside one of the world’s major information companies, according to a Fortune commentary by Safe Sign Technologies’ founder. The founder said the acquisition mattered because the company had not sold subscriptions, yet Thomson Reuters valued its technical work enough to buy it.

Safe Sign Technologies was founded 20 months before the deal, the founder wrote. He described it as Thomson Reuters’ first pre-revenue acquisition in its 174-year history and called it one of the more significant European transactions of 2024, without giving the purchase price.

The founder said the company’s path to the deal was shaped by a decision to stop pursuing near-term sales and put its resources into building a proprietary AI model. That meant telling backers that meaningful revenue was not expected for a long time, a message that caused many potential investors to lose interest, he wrote.

What is Safe Sign Technologies?

Safe Sign Technologies is an AI research startup that worked on model safety, robustness and reliability, according to its founder’s account in Fortune. He said the company built a model with differentiated performance while operating on a limited budget.

The founder said the team included people from Cambridge, MIT and Harvard. He also wrote that when Safe Sign shared strong internal results, Thomson Reuters’ venture arm responded within minutes.

How did Safe Sign get funded before the acquisition?

The founder said he built Safe Sign while training as a solicitor at Allen & Overy, starting work on the company before dawn, spending the day in legal training and returning to the startup afterward. He said the arrangement was tolerated by the firm but could not continue indefinitely.

Safe Sign’s first plan was a consumer legal product, but the founder said investors would not back it. He wrote that local investors repeatedly declined, leading him to travel to New York with £200.

Most of the company’s funding ultimately came from North America rather than the United Kingdom, according to the founder. He contrasted US investors, whom he described as more willing to help, with British investors, whom he portrayed as focused on failure risks.

Why does the founder say the deal matters?

The founder argued that the acquisition challenged a common startup assumption that visible revenue and fundraising are the clearest signs of value. In his view, institutions reward underlying technical strength when they examine a company closely.

He tied that lesson to an earlier experience defending Rollins House, an Art Deco former factory in southeast London where he said he was raised. As a teenager, he wrote, he taught himself planning law and helped oppose a redevelopment plan because his family could not afford lawyers; the council panel sided with them.

The founder now invests and said he focuses on separating presentation from defensibility. He argued that investors can miss companies with strong technical positions when they judge too heavily by pitch decks or early commercial traction.

What AI opportunities does he point to next?

In the Fortune commentary, the founder said investors may be overlooking the technical layer beneath the best-known AI labs and large models. He pointed to testing infrastructure, memory systems and specialized tools for AI-for-science labs as areas that may be undervalued.

He described AI evaluation as a hard problem because model claims depend on tests that must resist being gamed. He also said continual learning remains an unresolved machine-learning challenge, while early AI-for-science labs need specialized support tools they may not be able to build themselves.

Fortune noted that the commentary reflected the author’s views and not necessarily those of the publication. The original commentary appeared on Fortune.com.

This story draws on original reporting from Fortune.