A company nobody would call an AI lab just built its own frontier model for forty million dollars.

Thomson Reuters announced it on Monday. The model is called Thomson. Developed in house, fully owned and controlled by them.

Forty million dollars covering talent and compute. The labs it gets compared against spent billions.

They did not start from scratch. They started from an open source foundation and trained it on their own material. Westlaw. Practical Law. Checkpoint. Reuters. Decades of edited legal and tax content, shaped by hundreds of their own subject matter experts.

So far they have used less than ten percent of what they hold.

Ninety percent is still sitting there.

The expensive part of AI was never the model. Foundations are open and available to anyone. The expensive part is the decades of carefully edited proprietary material underneath it, and no funding round buys that.

Thirty years of building and Investing experience later, the question I keep asking founders is what they own that money cannot buy.

Most of them answer with their product. Products get copied by Friday.

The real answer is usually something boring that took years to pile up and nobody thought to write down.

What does your business own today that a better funded competitor could not buy?

Krishna Lakamsani | Entrepreneur · Venture Studio Founder · Investor. Writing at the intersection of AI, capital, and the future of work.#MyLifeLessons #Reflections #InspireFounders #BuildInPublic #AiImpact