Question 1

Q: What did this one leave behind?

Accepted Answer

Founder and chief executive of Nvidia, still in the seat after more than thirty years. In 2006 he made a decision that looked like burning money: build CUDA, so a graphics chip could do general computation. There were no customers for it and no market. Seventeen years later the AI boom arrived and all of it was running on his foundations. He did not bet on an application. He bet on a capability, and then paid to keep it alive.

Question 2

Q: What actually happened?

Accepted Answer

In 1995 Nvidia's first product shipped and failed. Microsoft's DirectX made the approach it was built on obsolete almost overnight, and the company was close to finished. He stood in front of the staff who were left and said, in effect, that he needed their help to get Nvidia out of this. Not I have a plan, follow me. I need you. It is a small thing to notice about a man now running one of the most valuable companies on earth, and it is why there was anybody left to run it.

Question 3

The parts · Zero-billion-dollar markets

Accepted Answer

Not billion-dollar markets: markets whose size today is zero, where there are no competitors because there are no customers. Enter one and the toolchain, the interfaces and the developers' habits are yours to define. What you buy is not a head start but the seat where the rules get written. e.g. Early in a category the valuable move is not selling units. It is making your interface the default that later arrivals have to build on.

Question 4

The parts · Two lines of blood supply

Accepted Answer

A zero market pays nothing for a decade. Gaming cards were the mature business feeding cash while CUDA burned it. Without the first the second is suicide; without the second the first is only a good business. Ambition shows up as a line you let look terrible for ten years. e.g. Most companies cannot hold a dark line open, not for lack of money but because they cannot survive the quarterly question about when it will earn.

Question 5

The parts · Always thirty days from going out of business

Accepted Answer

He says the company is thirty days from going out of business, and it is not modesty. Nvidia has been near the end at least three times. A permanent sense of the edge means that when the real emergency lands, nobody in the room is stunned by it. e.g. The teams that handle a crisis worst are the ones that had never once rehearsed the sentence this could end.

Question 6

The parts · Retreat is hardest for the ablest

Accepted Answer

Nvidia left mobile chips while they were still growing, on the structure rather than the quarter, and moved the engineers to robotics. He told the Taipei graduates that retreat does not come easily to brilliant people, and that deciding what to give up is central to success. e.g. The failing project hardest to kill is the one given to the best team, because nobody wants a loss on that record.

Question 7

Q: How do I use it today?

Accepted Answer

In your next plan, add a second question to how big is this market today. Ask which underlying capability, if it kept improving for five years, would unlock things that cannot be built now — and what it would cost to keep that line alive until then. Betting on a capability usually beats betting on an application.

Question 8

Lines to keep

Accepted Answer

We go after zero-billion-dollar markets. Our company is thirty days from going out of business. Retreat does not come easily to brilliant, successful people. Run, don't walk.

Question 9

Q: What should I read alongside?

Accepted Answer

Peter Thiel — Thiel asks what important truth very few people agree with you on. Building CUDA in 2006, when the industry was certain a graphics chip was for drawing pictures, is a live specimen of the answer. Duan Yongping — Opposite entries into a new category. Huang goes in before it exists and pays for a decade; Duan waits until somebody else has proved it and then builds the better version. Andy Grove — Both walked out of the business that was making the money. Grove left memory under fire; Huang left mobile chips while they were still growing.

Question 10

Further

Accepted Answer

Andy Grove Peter Thiel Duan Yongping Elon Musk Kevin Kelly Ren Zhengfei Zhang Yiming

Jensen Huang

What actually happened?

Zero-billion-dollar markets

Two lines of blood supply

Always thirty days from going out of business

Retreat is hardest for the ablest

How do I use it today?

Deep read

Read alongside

Further

Lines to keep