Jensen Huang built NVIDIA into an AI powerhouse. But does the company's future depend on the founder or the business he created?
KEY TAKEAWAYS
NVIDIA faces significant leadership uncertainty. Reports indicate that its board is considering removing co-founder and CEO Jensen Huang, although no final decision had been announced when the note was written.
NVIDIA has substantial key-man exposure. Huang has led the company since its founding in 1993, oversees approximately 60 direct reports, and has no publicly identified successor.
NVIDIA's business has transformed dramatically. Revenue has grown from $26.9 billion for fiscal 2022 to $96.2 billion in its latest reported quarter, with data center sales accounting for approximately 92% of quarterly revenue.
NVIDIA's competitive position extends beyond its founder. Its CUDA software ecosystem, established customer relationships, and substantial demand for AI infrastructure create business advantages that would survive an immediate leadership change.
Investors face two distinct categories of risk. Existing customer commitments and installed technology provide operational continuity, while future product development, supplier relationships, and strategic direction remain exposed to a potential leadership transition.
MY HOT TAKES
The market may be confusing the value of Jensen Huang with the value of NVIDIA. Huang built the company, but billions in revenue, decades of software development, and deeply embedded customer relationships don't simply disappear when a CEO leaves.
Jensen Huang was arguably more indispensable five years ago than he is today. Back then, investors were buying his vision; today, they are buying an established business with substantial demand and a formidable competitive moat.
NVIDIA's biggest leadership vulnerability isn't losing its CEO; it's failing to prepare for the possibility. A company worth trillions of dollars should never have its organizational structure depend so heavily on one individual, regardless of how extraordinary that individual might be.
The next CEO's job isn't necessarily to preserve NVIDIA's current business; it's to build the next one. The existing order book and CUDA ecosystem provide substantial advantages, but neither guarantees the company will dominate the next generation of computing.
Index investors have more exposure to this leadership drama than many realize. When a single company represents more than 6% of the S&P 500, its succession risk becomes a portfolio risk for millions of investors who never deliberately bought the stock.
My hot takes: "Somewhere in the last five years, NVIDIA’s value quietly migrated–from vision to purchase orders, from charisma to code."
Speed dial. Early in my career–back when a Rolodex was an actual object that sat on an actual desk–the biggest rainmaker at my firm resigned on a Friday afternoon. By Monday morning, the place felt like a wake. Managers huddled behind closed doors, assistants whispered at the coffee machine, and the rest of us did the only math that mattered: how many of his connections would follow him out the door? The man WAS the business–or so we all believed. He had the relationships, the coveted golf handicap, and the corner office to prove it.
Here is what actually happened. Some clients left–a few of the big ones, in fact–and it stung. But most stayed. They stayed because their accounts were there, their paperwork was there, and the people who picked up the phone at 7:00 AM when something went wrong were there. It turned out that a good many of those clients had signed with the FIRM, not the man. That episode taught me to ask Wall Street’s oldest question whenever a star heads for the exit: does the business belong to the banker–or to the bank? I have been asking it ever since, and this morning I am asking it about the most important company in the stock market.
By now you have seen the headlines. The Wall Street Journal reported last night WHILE YOU SLEPT that NVIDIA’s board is preparing to remove Jensen Huang, the only CEO the company has had since he co-founded it in 1993. CNBC confirmed that the directors met on Wednesday to decide his future. As I write this, no decision has been announced and the company has not said a word, so let’s treat this as an IF, not a WHEN. I will leave the personal side of the story to other publications (there is no shortage of them 😉). What matters to us is the stock, which was down more than 7% in premarket trading after closing Wednesday at $173.73. And what matters even more is the vacuum. NVIDIA has no chief operating officer, no board chair, and no public succession plan. Some 60 executives report directly to Huang, who has said–out loud–that he doesn’t believe in succession planning. That–unquestionably–is a LOT of Rolodex.
Let me be clear about something before I go any further: management matters. It matters more than most investors care to admit, largely because you can’t find it on a balance sheet. I spent a good part of my career in tech M&A, and I can tell you that in many of the deals that crossed my desk, the people WERE the asset. We wrote retention packages into the paperwork precisely because the engineers and the founders could walk out the front door–and a good portion of what we had just paid for would walk out with them. In a young company, the founder IS the strategy. He picks the markets, recruits the talent, charms the customers, and convinces the capital markets to fund the whole adventure. Take him away, and you are left with a business plan and some overpriced, bougie office furniture. NVIDIA itself agrees with me–its annual report warns that the company is “highly dependent” on its long-standing executive team. And the Street is already doing the arithmetic. My friend, Dan Ives estimates that NVIDIA would be worth 25% to 30% less without Huang, and Bernstein just cut its price target in half–from $400 to $200. So no, I am not here to tell you that the man in the leather jacket doesn’t matter. He does.
But here is the question I want you to chew on: WHEN did he matter most? Turn the clock back five years. In fiscal 2022 (which, for NVIDIA, covered most of calendar 2021), the company took in $26.9 billion for the ENTIRE year, and its largest business was gaming–graphics cards for video games–at $12.5 billion. The data center business was $10.6 billion, and the notion that it would one day eat the world was, at that point, a story. A very good story, told by a very gifted storyteller–but a story nonetheless. If you owned NVIDIA in 2021, you owned a bet on Jensen Huang’s vision. You were paying for the Rolodex.
Now look at what you own today. In its most recent quarter (the one that ended in July), NVIDIA took in $96.2 billion–in ONE quarter–more than three and a half times what it collected in that entire year. Data center accounted for $89 billion of it, or better than 92 cents of every dollar. Gross margin was 75%. Management guided the current quarter to $108 billion and has told investors it sees at least $1 TRILLION in revenue from its Blackwell and Rubin chip platforms from 2025 through calendar 2027. I should be careful here–that trillion is management’s number, built on purchase orders and what the company calls “high-confidence” demand. Not every dollar is signed, sealed, and delivered. But these are orders from the largest and best-capitalized companies on earth, placed years in advance for chips that are in short supply. And then there is the moat. NVIDIA’s real weapon is not a chip–it’s CUDA (the software platform developers use to write programs for NVIDIA’s chips), which the world’s AI engineers have been building on for the better part of two decades. Every line of code written on it is one more reason not to switch. Somewhere in the last five years, NVIDIA’s value quietly migrated–from vision to purchase orders, from charisma to code. The clients, in other words, signed with the bank.
So, what can a new CEO actually break? Not the order book–customers who have waited in line for these chips are not going to step out of it because a nameplate on the corner office changed. Not the installed base, and not two decades of developer habit–at least not in a quarter or two. What a new leader CAN break is everything that hasn’t been built yet. Huang has been the one deciding what comes after Rubin. He is the one working the suppliers for manufacturing capacity that competitors can’t get. He is the hub of a wheel with 60 spokes, and wheels like that tend to wobble when you pull the hub. And he is a big reason the market has been willing to pay a premium for every dollar of NVIDIA’s earnings. The names in circulation tell you the board understands both halves of the problem. Reuters reports that analysts and investors see CFO Colette Kress, who has run the company’s finances since 2013, and Ian Buck, the engineer who joined in 2004 and helped create CUDA, as the likeliest candidates–at least for the short term. One is a steward. The other is a builder. Stewards protect what is already on the books. Builders decide what gets booked NEXT. Which one the board reaches for (or whether it looks outside altogether) will tell you a great deal about which risk it fears most.
So here are the questions you should be asking yourself this morning. First–how much of what I own is NVIDIA the company, and how much is NVIDIA the man? Remember that if you own an S&P 500 index fund, somewhere north of 6% of it is this single stock, whether you picked it or not. Second–NVIDIA’s market value has fallen from roughly $5.8 trillion at its October 6 record to about $4.1 trillion at Wednesday’s close. How much of that drop is the market pricing a person, and how much is it pricing a product cycle and a Fed that is widely expected to hike next month? Those are very different risks, and they deserve to be separated. Third–the company reports earnings on Tuesday, November 17. I will be listening less for the quarter and more for the order book. Is the trillion still a trillion?
As for that rainmaker of mine–he did just fine at his new shop, and so did the firm he left behind. The clients who followed him were the ones who had hired HIM. The ones who stayed had hired US! Nobody could tell the two groups apart until the day he cleaned out his desk. NVIDIA’s board may be about to run that same experiment with one of the most valuable Rolodexes in the history of capitalism–and everyone with an index fund has a seat at the wake. My hunch is that the phones in Santa Clara will keep ringing at 7:00 AM. They usually do. It’s the calls that haven’t been made yet that I’d lose sleep over. And a word of advice for whoever ends up in that corner office: on your first morning, check the desk. If the Rolodex is still sitting there–you’re going to be just fine. 🤣