The data sitting in your company right now could be a new business.

Strava sells fitness routes to governments. Mastercard sells transactions to retailers. What could you do with what you already have?

In 2006, Fei-Fei Li, the same woman who launched World Labs in 2024 to create models for the visual world and raised $1 billion on the first year, had an idea that nobody wanted to fund: ImageNet.

She wanted to build the world’s largest image dataset: 14 million photos across 21,000 categories, hand-labeled by humans. The subset used for competition: 1.2 million images. Her peers thought it was grunt work. Not science. Not worth the investment.

She built it anyway. She called it ImageNet.

In 2012, a neural network trained on ImageNet crushed every benchmark in computer vision. It was the starting gun for modern AI. Every model you use today (every product recommendation, every fraud detection system, every autonomous vehicle) traces its lineage back to that dataset.

Fei-Fei Li didn’t build a better algorithm. She saw the value of data before anyone else did.

ImageNet

You don't need to build the next ImageNet. But her instinct: that data others overlook has enormous value, is one every business leader should borrow."

Example 1: Strava

Strava built a fitness app for runners and cyclists. Millions of people tracked their routes. That GPS data was a byproduct: valuable to users, but invisible as a business asset. Until city planners came knocking. Strava turned that data into Metro, a platform now used by urban governments worldwide to design bike infrastructure. The data was already there. They just asked: who else needs this?

Strava routes map in London

Example 2: Mastercard

Mastercard processes billions of transactions every day. That’s operational infrastructure. But someone looked at that flow and saw something else: the most accurate real-time picture of consumer spending on the planet. They turned it into SpendingPulse: a market intelligence product sold to retailers, investors, and governments. Same data. Entirely new revenue stream.

In both cases, the data already existed. No new collection. No new technology. Just a different question: what is this worth to someone else?

Mastercard

The question most CEOs never ask

Most AI conversations in the boardroom start with tools. Which platform. Which vendor. Which model. That’s the wrong starting point.

The right starting point is your data. Not the data you plan to collect: the data you already have. Data that crosses your organization every day without being captured, valued, or understood.

Fei-Fei Li didn’t wait for better AI. She built the fuel first.

The question isn’t whether you have data. You do. The question is whether you have the vision to see what it’s worth.

The companies that will win the next decade aren’t the ones that buy the best AI tools. They’re the ones that understand what they already own.

Question for this week

“What data already flows through my organization, that I have never stopped to value?”