business 8 min read

Nvidia's New Era Begins After Huang's Exit

Jensen Huang has left Nvidia, but the company's earnings suggest the AI chip machine keeps running. Colette Kress takes the helm—and the real question is whether the market can accept a CEO who never designed a GPU.

  • Corporate Governance
  • Semiconductors
  • NVIDIA
  • AI Chips

The Numbers Don’t Lie — But They Also Don’t Tell the Whole Story

Nvidia posted $115.8 billion in quarterly revenue Tuesday, crushing Wall Street’s estimate of $109.1 billion. Diluted earnings per share came in at $2.68 against expectations of $2.48. Net income reached $65.3 billion, more than doubling year over year. The cloud industry backlog grew to $2.5 trillion from $2 trillion just three months prior.

On paper, this looks like business as usual at the world’s most valuable chipmaker. In practice, it looks like something far more consequential: a company proving it can function without the man who built it.

Jensen Huang resigned last week following revelations of extramarital affairs. On Tuesday, he released a statement saying “I own my mistakes completely” and suggested he would be open to returning if the board thought it served shareholders’ interests. Harvey Jones, Nvidia’s board chairman, ended that speculation immediately — unanimously accepted the resignation, no plans to reinstate him.

Colette Kress, Nvidia’s CFO since 2013, stepped into the CEO role and delivered the earnings call herself. Her message was calibrated to project stability: “Our momentum is real. Our demand is accelerating. Our supply is growing. We’ve got this.”

That phrase — “we’ve got this” — was doing heavy lifting. It needed to reassure institutional investors who had bet billions on Huang’s vision. It needed to calm customers who rely on H-series GPUs for their AI infrastructure. And it needed to signal to employees that the culture Huang cultivated would survive his departure intact.

What followed was a meticulously choreographed display of competence. Kress answered every analyst question with precision, deflected personal inquiries about Huang’s departure with practiced neutrality, and repeatedly returned to the same triad of metrics: backlog growth, gross margin expansion, and data center revenue velocity. The tape of the call will read as proof of continuity. Those who were there noted something else — the absence of a particular energy that used to make these calls feel like events rather than reporting obligations. Huang had a habit of turning quarterly updates into cultural moments. That gravity is now missing, and the silence around it was louder than any carefully worded press release.

The One Person Who Can’t Be Replenished

Here is what the earnings call didn’t address directly, and why it matters: no one on that stage has ever owned a data center full of Nvidia’s latest GPUs the way Huang did. He wasn’t just CEO. He was the chief evangelist, the face of the product, the person every major cloud provider knew they could call when a roadmap slipped or a timeline tightened.

Kress is a formidable operator. She has been at the center of Nvidia’s extraordinary growth for thirteen years. Jones praised her understanding of the business, the customers, the operations. But there is a difference between managing financial strategy and commanding the imagination of an entire industry.

The supply chain tells a different story than the revenue numbers. Nvidia’s backlog grew by half a trillion dollars in a single quarter. That represents committed demand stretching years into the future. The question is whether the company can convert that commitment into delivered product once the leadership uncertainty settles — or whether customers will begin looking at alternatives.

AMD exists. Intel’s foundry push is accelerating. Huawei’s AI chip development continues under export restrictions. None of them are near Nvidia’s current position today. But leadership transitions create openings. And the AI infrastructure buildout isn’t going to wait for a permanent CEO appointment.

Consider the second-order dynamics. When a founder-ceo departs amid scandal, even a financially pristine quarter, the narrative around the company shifts before the fundamentals have a chance to prove themselves. Competitor executives who previously deferred to Nvidia’s market primacy may start recalibrating their own positioning. Sales teams at cloud providers who relied on Huang’s direct relationships for escalation paths now need new channels. TSMC, which manufactures Nvidia’s chips, will face its own questions about account prioritization when the person who traditionally held preferential access is gone.

The Internal Fractures That Won’t Appear in Earnings Reports

Perhaps the most significant consequence of Huang’s departure won’t show up in any quarterly filing. For nearly thirty years, Nvidia’s engineering culture has revolved around a single point of strategic conviction. Product roadmaps, R&D allocation, talent acquisition — all filtered through Huang’s preferences. He had famously resisted diversifying away from graphics into general-purpose computing until the market forced his hand, then moved with such speed that the rest of the industry spent a decade catching up.

That kind of centralized decision-making produces extraordinary results in stable times. It also creates fragility. Key engineers and architects who joined Nvidia specifically because of Huang’s vision may begin reassessing their positions. Not necessarily leaving — yet — but quietly updating their resumes, attending conferences, reestablishing connections outside the Nvidia ecosystem. The brain drain from leadership transitions is rarely immediate. It is gradual, quiet, and devastating when it finally accumulates.

Additionally, there is the matter of internal morale that no earnings call can capture. Nvidia employees have publicly praised Kress’s operational rigor. Privately, sources familiar with the culture describe a palpable sense of disorientation among mid-level managers who built their careers around Huang’s management style — direct, intense, often confrontational in ways that paradoxically created loyalty precisely because it felt genuine. Kress operates differently. She is measured, diplomatic, consensus-oriented. Neither approach is wrong. They simply produce different organizational temperatures.

Who Wins. Who Loses. What Happens Next.

The board says it won’t rush the process. “We’re not in a hurry to find exactly the right person,” Jones implied, which translates to: the search could take months.

During that window, several outcomes become likely:

First, existing customers with multiyear supply agreements extending through the end of the decade will continue operating under contract terms that were negotiated assuming Huang would still be steering the ship. Those contracts provide a floor. But they don’t prevent renegotiation pressure or scope creep in directions that benefit the buyer. Several hyperscalers are already understood to be exploring clause adjustments, particularly around delivery timelines and pricing escalators tied to performance milestones.

Second, potential new entrants into the high-performance GPU market may sense vulnerability. Nvidia’s moat has always been software as much as silicon — CUDA is the platform lock-in that makes switching costs prohibitive. But if leadership drifts even slightly, competitors gain breathing room to close the gap where they can. AMD’s MI300 series already runs CUDA-competitive workloads through forward-compatible tooling. Google’s custom TPUs serve a growing share of internal training workloads. Amazon’s Trainium chips are inching toward viability. None represent existential threats today. All represent opportunities that compound over time.

Third, the market itself will test whether the narrative holds. Nvidia shares dropped roughly 2% in Wednesday trading after the earnings release — a modest decline that suggests institutional investors are holding steady, at least for now. But stock price movements during leadership transitions are often delayed reactions. The real pressure may come later, when the permanent CEO search drags on without a clear frontrunner.

There is also a fourth dynamic worth tracking: the geopolitical dimension. China restrictions on advanced semiconductor exports have complicated Nvidia’s business in one of its historically important markets. Huang navigated those constraints with remarkable skill, finding workarounds that kept revenue flowing while staying within regulatory boundaries. A replacement CEO will inherit that same tension without Huang’s institutional knowledge of how each workaround was constructed.

The Structural Question No One Is Asking Out Loud

Huang’s resignation raises a question about Nvidia’s architecture that has nothing to do with personal conduct. For nearly three decades, the company’s strategy, product roadmap, and corporate identity have been inseparable from one person. That is true of very few public companies at this scale.

Kress’s interim tenure will reveal whether Nvidia’s systems are robust enough to absorb leadership change without strategic disorientation. The financial results demonstrate operational discipline. The backlog demonstrates market demand. But discipline and demand are not the same thing as direction.

There is a subtle distinction that matters here. Nvidia does not lack people who understand the business. It lacks the person who made the hard calls about where the business should go next — calls that frequently went against conventional wisdom, against near-term financial logic, against boardroom consensus. Huang bet everything on general-purpose GPU computing when every analyst told him to stay focused on gaming. He pivoted to AI training before anyone else saw the shape of the wave. Both moves looked reckless at the time. Both proved prescient.

A successor — interim or permanent — inherits a company whose growth trajectory is unprecedented but whose strategic independence is unprecedented as well. The challenge ahead is not maintaining the status quo. It is figuring out what comes next when the next comes from someone other than the person who imagined it first.

Nvidia is well positioned, Kress said. The board agrees. The numbers back both claims, for now.

The next six months will determine whether this positioning is durable or merely temporary.