Design the Machine First
Most founders start with an idea. Jensen started with a system. He asked: what is this machine we're building? What goes in? What comes out? What conditions does it operate in?
That's first principles thinking. Don't copy another company's org chart. Don't follow industry norms. Figure out what the machine needs to do, then build backward from there.
"With respect to building a company, the first thing you have to do is start from first principles. What is this machine that we're trying to create? What is its input? What is its output?"
Only Hire People Who Want Impossible Problems
Jensen didn't want a big company. He wanted a small team of people obsessed with solving problems nobody else could touch.
NVIDIA's mission acts as a filter. If a problem can be solved by normal computers, they pass. That single rule attracts a specific type of person. The kind who wants to push boundaries, not punch a clock.
"Our company's mission is to solve computing problems that are barely possible. If a problem could be solved by normal computers, we don't do it."
Small Beats Big. Every Time.
Jensen wanted NVIDIA as small as possible. Not as large as possible. The company should be as large as necessary to do the job well. Not one person more.
Naturally, you want to empower people.
"You want a company that's as small as possible, not as large as possible. It needs to be as large as necessary to do the job well, but to be as small as possible."
Kill the Pyramid
If you want command and control, build a pyramid. Militaries have done it since the Roman Empire. But Jensen didn't want obedience. He wanted empowerment.
So he made NVIDIA flat. No business units. No divisions. Everyone works as one team. Information travels fast because there are no layers to slow it down.
"If you want to empower people, you want to make it as flat as possible so that information travels quickly. We don't have business units. We don't have divisions. Everybody works as one."
Shape the Company Around One Strength
NVIDIA is shaped to do one thing better than anyone on earth: accelerated computing. Jensen admits the company would fail at anything else. Fried chicken? No chance. Swedish meatballs? Forget it.
That's the point. The entire organization is optimized for a single purpose. Every hire, every process, every decision flows from that constraint.
"The company is shaped in a way that allows us to build accelerated computing best. If you asked me to go do fried chicken, we'd have a hard time. Swedish meatballs? No chance. But accelerated computing? Very well."