No One in the C-Suite Wants to Admit They Don't Understand AI
Hugging Face's CEO runs a $4.5 billion company and keeps publicly saying 'I don't know.' Most C-suite executives can't bring themselves to. That's the gap.
In 2024, Clément Delangue, CEO of Hugging Face, sat for a long interview about the future of AI. He said this:
“No one knows a profitable, sustainable business model for AI.”
A few minutes earlier in the same conversation, asked about a controversial model that had been hosted on his platform, he said: “We are still just scratching the surface when it comes to ethics reviews.” Most of ML research, he added, is in the same place.
Delangue’s company hosts more than 3 million AI models. It became profitable in August 2024 while keeping most of its platform free. Its Series D, led in August 2023 by Salesforce, Google, Nvidia, and Amazon, valued the company at $4.5 billion. By any reasonable measure, Delangue is the operator best positioned to claim he knows what works in commercial AI.
He has spent the last two years saying out loud that he doesn’t.
Compare that to a typical enterprise C-suite conversation in the same window. The board asks what we’re doing about AI. The CEO answers with a roadmap. The CTO presents pilots. Somewhere in the middle of that conversation, certainty gets manufactured that nobody in the room actually has.
The contrast is the article.
“I don’t know” is a leadership move, not a confession
Delangue’s “I don’t know” statements are not hedges from an uncertain leader. They’re stated publicly, by the CEO of a category-defining company, while the company is shipping products, partnering with Amazon and Google, testifying before the U.S. House Science Committee, and raising $235 million.
The pattern is consistent: high-conviction action paired with explicit uncertainty about the destination. Announcing the Series D, Delangue said, “AI is the new way of building all software. Hugging Face intends to be the open platform that empowers this paradigm shift.” Asked about the business model in the same window: “No one knows what a profitable, sustainable business model for AI is. The beauty of the position we are in is that if you are the number one platform, there’s a sustainable massive business model around it.”
Read those sentences together. The strategy is held with conviction. The financial return on that strategy is held with humility. Both are true at the same time.
That posture is the move most executives don’t make. They collapse strategic conviction and financial certainty into a single posture of confidence. When the market shifts under them—as it has with AI in every quarter since late 2022—the posture becomes the liability.
Why most C-suites avoid it
A 20-year career in any senior role builds an identity around knowing more than the people in the room. AI dissolves that identity. A junior analyst with a good prompt can produce in 20 minutes what used to be the senior person’s defensible expertise.
The natural response is defensive: project mastery, downplay what you haven’t tried, hire a consultant whose deck restores the comfort of certainty. The result is decisions made on borrowed conviction—pilots that don’t ship, budgets approved for tools no one uses, and an org that learns to perform AI competence rather than build it.
Delangue’s posture removes that pressure. By stating publicly that no one knows, he does two things at once. He gives his own team permission to experiment without political cost. And he gives partners and investors a more honest picture of the risk they’re underwriting—which, paradoxically, increases their trust.
What it looks like in practice
Co-founder Thomas Wolf described the company’s approach: “People come to me ten times a day with a proposal for a new project, and I say yes maybe to maybe 1 percent of it. We also do a lot of small bets with maybe one or two people, to see if they can prove that it will work. We want people to have a plan—what happens if they succeed? What happens if they fail? What are the lessons?”
The structural question—what happens if it fails, what are the lessons—is the operational form of “I don’t know.” It assumes the failure case is real, expects it, and budgets for it. Most enterprise pilot reviews ask the opposite question: how do we make sure this works.
Wolf, on what the company hadn’t yet figured out at the time the case was written: “We still haven’t nailed the part that will be the most interesting in terms of value.” The company was already on its way to profitability. Both states coexisted.
The translation
For a C-suite operating in 2026, the actionable version of the Delangue posture has three pieces.
First, separate strategic conviction from financial certainty. You can be certain that AI will reshape your business without being certain when, how, or which products will pay for it. Saying both, in the same sentence, reads as more credible than projecting confidence on either alone.
Second, treat “I don’t know” as a structural input. Plan budgets, pilots, and head count as a portfolio of small bets, not a single funded program. Wolf’s “what happens if they fail” is the planning prompt; most organizations don’t have a routine slot for it.
Third, model it from the top. The reason most teams perform AI competence is that the senior leadership has implicitly required them to. Delangue inverts that. The team learns to experiment because the CEO says, in public, that he is also experimenting.
The most surprising thing about the most successful AI CEO of the last 24 months is how often he tells reporters and investors he isn’t sure. The least surprising thing is that this hasn’t slowed him down.