Business thinker

Jensen Huang on first principles, new markets and showing your reasoning

Updated 4 November 2026

Not affiliated with or endorsed by Jensen Huang or their companies.

Jensen Huang co-founded NVIDIA in 1993 and has led it since. In a 2024 conversation at Stanford Graduate School of Business, published with a full transcript, he described three habits that apply well beyond chips. Learn how something is done, then redo it from first principles for today's conditions. Create the market as well as the technology when the market doesn't exist yet. And show people how you reason, rather than only handing down answers.

Redo it from first principles

The conversation opens with Huang describing how he approaches a problem: first learn how it is done, then go back to first principles and ask how he would redo it, given the conditions today, his motivation, the tools available and how things have changed. The point is not to ignore what exists, but to stop assuming it must stay that way.

Take a fictional chain of language schools that still places new students with a 45-minute paper test marked by hand, a process designed when there was no other way to score it. Asked from first principles, the grammar part could be scored online at booking, leaving teachers a short speaking check, and students could be placed within a day instead of a week.

Create the technology and the market

Huang describes NVIDIA's founding as an intersection: a technology that was hard to make cheap, and a market that didn't exist yet, the videogame market of the time, which he calls "zero billion dollars". He says the company creates technology and creates markets, building the conditions for people to buy, and that this is why it was early in fields such as autonomous driving and deep learning.

For a smaller company, the lesson is a question to ask before entering a new space: if the market doesn't exist yet, what else must exist for anyone to buy, and who will build it?

Show how you reason, and keep information flowing

The interviewer describes Huang's organization as flat, with about 50 direct reports and an invitation to people across the company to send him the top five things on their mind. Huang answers that no task is beneath him, recalling his work as a dishwasher, and that when someone asks for his input, he shows them how he reasoned through it. Seeing how someone reasons through something ambiguous or scary, he says, empowers people.

How to apply it to your own work

1. Write how a process works today and why each step exists.
2. Mark which reasons still hold given today's tools, costs and customers.
3. Redesign it from the goal, with today's tools, and mark each step kept, changed or new.
4. If the result is something new, list what must exist for people to buy it, and who builds each piece.
5. Write the reasoning, including the uncertain parts, and share it with the people affected.
6. Decide how people closest to the work will send back what they see.

What the sources say, and what they don't

The ideas above come from one public conversation, the Stanford GSB transcript of April 2024; the description of the flat organization is the interviewer's, which Huang answered. We don't describe NVIDIA's strategy or results beyond what he said there. The language-school example is ours.