Safe Sign: The Bold Sale of a Revenue-Free AI Startup

A Unique Acquisition for Thomson Reuters

At just 21 years old, while still training to become a lawyer in the UK, the founder of Safe Sign Technologies pulled off a bold gamble. He established an artificial intelligence research company with experts from prestigious institutions such as Cambridge, DeepMind, Harvard, and MIT. These researchers believed in his vision enough to join him on this entrepreneurial journey.

In just 20 months, Safe Sign Technologies was acquired by Thomson Reuters. This purchase marked a first in the 170-year history of the company, as it was the first acquisition of a company before it generated any revenue. Thomson Reuters was drawn to the scientific promise of the startup.

A Rocky Path

The road to this deal was not without obstacles. Safe Sign Technologies had managed to publish work placing its model among the best in legal reasoning. Furthermore, it had succeeded in training this model on a budget far lower than that of major laboratories. This approach is reminiscent of DeepSeek, a company known for its high-performing models and innovative algorithms.

Despite these successes, convincing investors proved to be challenging. Discussions often focused on the product and commercial traction, areas where the startup had yet to prove itself. Faced with skepticism from British investors, the founder had to turn to the United States to raise the majority of the necessary funds.

The Importance of Science in Innovation

Today, the founder of Safe Sign Technologies supports other entrepreneurs and remains convinced that pure science is crucial, even if it is not immediately integrated into a product. This perspective, often overlooked by investors, is, in his view, essential for the future.

The Rise of Fundamental AI Startups

In 2026, fundamental AI startups raised approximately $178 billion, highlighting the growing importance of this sector. However, a large portion of this funding, around 97%, was captured by already established players like OpenAI, Anthropic, and xAI. This leaves little room for new companies to carve out a space.

For deep-tech founders just starting out, this concentration of capital can make it seem like the race is already lost. Yet, some see this as an opportunity to stand out by building on solid foundations rather than relying on existing giants.

Building on Solid Foundations

Companies that focus on the application layer, without owning their own model, are often at the mercy of pricing and access decisions made by large firms. In contrast, those that concentrate on the foundational layer, where scientific challenges remain numerous, have a much higher potential for sustainability.

Advancements in training efficiency, model architecture, and inference cost are areas where progress will have a lasting impact. These innovations will be crucial long after current solutions have been surpassed.

Key Questions for Founders

When evaluating startups, the founder of Safe Sign Technologies asks essential questions:

• Is the technical team on par with those at DeepMind?
• Is the problem being addressed fundamental or merely an improvement on existing work?
• Will the product become indispensable in the coming years?

Companies that succeed in answering these questions are often those that, like DeepMind and OpenAI, started without an obvious product but ended up playing a major role in the industry.

The Future of Deep-Tech

Founders must focus on creating fundamental solutions, even if they do not seem immediately fundable. The deep-tech capital market is slow to react, but those who persist in this direction will be the ones the market seeks tomorrow.

The future of AI lies in deep-tech, not in superficial solutions. Founders who tackle fundamental problems today will be the leaders of tomorrow.