On 12 November, Jensen Huang stepped down as CEO of Nvidia.
He had run the company for 33 years. He co-founded it over coffee at a Denny's in San Jose. He took it from graphics chips to the center of the AI boom, and from about $400 billion in value to more than $4 trillion. Then, in roughly two weeks, a scandal about his personal life went from a newspaper story to an internal investigation to a resignation.
I'm not going to pick over the scandal. Other people will do plenty of it. I care about one line buried in NPR's report on his exit:
"There's no obvious replacement."
Stephen Witt said it. He wrote a book about Nvidia. Analysts quoted in the same piece said Huang had a blind spot for succession and had never set out a clear plan in public. The board named the CFO, Colette Kress, as interim CEO and announced a search.
One of the most valuable companies in the world had no named successor. I'd bet your team doesn't have one either.

The founder trap isn't only for founders
When one person is the face, the brain and the final word, the company and the person blur into one thing. People stop asking "what does the company think?" and start asking "what does Jensen think?"
It feels efficient. One person decides. Everyone moves. No committees.
Then the person leaves. Sometimes the cause is a scandal. More often it's illness, burnout, a better offer, a family crisis, or retirement. The company finds out it never stored its judgment anywhere except in one head.
You don't need to run a $4 trillion company to fall into this. The same pattern shows up in a ten-person startup or a 200-person engineering org:
- One person approves every architecture decision.
- One person holds the customer relationships.
- One person knows why the billing system works the way it does.
- Everyone else waits for them to get back from vacation.
You've built a company with a single point of failure, and the failure point has a pulse.
Engineers have a name for this
Software people call it the "bus factor" or the "truck factor." It's the number of people who would need to get hit by a bus (or quit) before a project stalls.
In 2016, a team of researchers measured it. They built a method to estimate the truck factor and ran it against 133 popular projects on GitHub. Their finding: 65% of those projects had a truck factor of two or less.
Popular, widely used software. Two people or fewer holding the knowledge needed to keep it alive.
If open source projects with thousands of users look like this, what do you think your internal tools look like? Or your sales pipeline? Or your leadership team?

Boards talk about this for about two hours a year
Here's the part I find hard to believe. Succession is the single biggest risk most companies face, and most boards barely spend time on it.
Stanford's 2010 survey on CEO succession planning asked CEOs and directors at large and mid-cap public companies in the US and Canada about their plans. Here's what they found:
- 69% said a successor needs to be "ready now." Only 54% were grooming anyone for the job.
- More than half had no name for a permanent replacement if the current CEO became incapacitated tomorrow.
- 39% had "zero" viable internal candidates.
- Boards spent an average of two hours a year on CEO succession.
- 65% had never asked their internal candidates whether they even wanted the job.
Two hours a year. On the question of who runs the company if the boss walks out.
I'm not sure about this part: the Stanford survey is from 2010. I found no free, recent survey with comparable numbers. I'd be surprised if boards have changed much, but I haven't seen the data to prove it either way.
Why smart leaders avoid this
I don't think most leaders skip succession out of ego. Some do. Most avoid it for quieter reasons.
It feels like planning your own funeral. Nobody enjoys sitting down to write the document titled "What happens when I'm gone."
Being needed feels good. When the team brings you every hard call, you feel important. Handing those calls away feels like handing away your worth.
Training someone is slower than doing it yourself. This is true on any given Tuesday. It's false over any given year.
Naming a successor creates politics. Pick one person and three others feel passed over. So you pick no one and tell yourself you're keeping the peace.
In an October interview with Bloomberg, quoted by NPR, Huang said he hoped the next CEO was already inside the company... but 20 years from now, if he stayed healthy. Both hopes are reasonable. Neither is a plan.
What to do on Monday
You don't need a board committee for this. You need a few honest conversations and some writing.
1. Run the bus test on yourself
Pick a date two weeks from now. Pretend you're unreachable from then for a month. No phone, no Slack, no "quick question." Write down every decision, relationship and system which would stall.
The list is your succession plan's to-do list.
2. Name a deputy for each item
Not one heir for everything. One person per area. The architecture calls go to your strongest senior engineer. The key customer goes to the account lead. The budget goes to whoever already builds the spreadsheet.
Tell them. Out loud. The Stanford data says most companies never ask their candidates if they want the job. Ask.
3. Let them decide while you're still here
A deputy who only steps in during an emergency isn't a deputy. They're a spare tire nobody has checked for air.
Give them real decisions now. Let them make calls you would make differently. Back them in public when they do. This is the expensive part, and it's the only part which works.
I wrote earlier about how innovation dies when only one brain decides. Succession is the same problem with a longer fuse.
4. Write down the "why"
The hardest thing to replace isn't what a leader decides. It's the reasoning behind it. Why did we pick this vendor? Why do we never discount for this customer? Why does the deploy run at 4pm?
Start a decision log. A sentence or two per call. Future you will thank you, and so will whoever replaces you.
5. Put it on the calendar
If the board of a public company manages two hours a year, you beat them with one hour a quarter. Review the list. Check the deputies. Move one more thing off your plate.

The real test of a leader
Here's the uncomfortable truth. If your team falls apart the week you leave, you haven't been leading them. You've been carrying them.
Carrying people feels like leadership. It even looks like leadership from the outside. You're busy, you're essential, everyone comes to you. But it leaves the team weaker every year you do it.
Good leaders build people who don't need them. It's the job. Every decision you keep for yourself is a decision someone else never learns to make.
Our research on bad bosses at Step It Up HR found 99.5% of people have had at least one bad boss. Plenty of those bosses were the "indispensable" kind. They hoarded the calls, the context and the credit. Their teams paid the price when they left, and often while they stayed.
Nvidia will most likely survive. It dominates the market for the chips behind AI. Your team has no cushion like it.
So here's my question for you: if you vanished for a month starting tomorrow, who would make your next ten decisions? If no names come to mind, you know what to work on this week.