Nvidia After Jensen Huang

Summary

  • Huang’s departure puts Nvidia’s leadership and export oversight in focus. Longtime CFO Colette Kress takes over as interim CEO while the board searches for a permanent successor. Questions about whether advanced chips were diverted to China remain unresolved.

  • Nvidia’s competitive advantage remains substantial. Its combination of chips, CUDA software and networking is hard to replicate. The immediate test is whether new leadership can secure supplier commitments and keep the annual product cycle on track.

  • Nvidia can stay on top in AI and still face pressure on margins and valuation. Its largest customers are developing their own chips, while smaller cloud providers rely more on outside financing. 

What happened

On November 12th, Jensen Huang resigned as Nvidia’s CEO, president and a member of its board, effective immediately. He had led Nvidia since co-founding it in 1993. The board named longtime CFO Colette Kress interim CEO while she continues as CFO, appointed Harvey Jones as board chair, and began a search for a permanent successor. The board says it expects Nvidia to keep its customer commitments and product roadmap on track. Kress has been with Nvidia since 2013 and knows the business through several very different phases of its growth.

The resignation followed reports about Huang’s personal relationships, including one involving an executive at a customer, and questions about whether advanced chips sold to a Singapore-based buyer were diverted to China. Nvidia’s independent directors retained outside counsel after reports that Huang had been targeted by a foreign intelligence effort. Nvidia’s transition announcement gave no reason for his departure. For shareholders, the open questions are what Nvidia knew about where its chips were going, how it handled any warning signs, and whether the findings could change where, how and to whom it sells.

Why it matters

Huang’s biggest contribution to Nvidia was seeing where computing was going years before the demand was obvious. Nvidia built its name selling graphics chips for gaming. With CUDA, it gave developers the tools to use those chips for a much wider range of computing tasks. It then bought Mellanox, whose networking technology helped Nvidia link thousands of chips into a single system. When ChatGPT arrived, Nvidia already had the chips, the software and the networking needed to train large AI models at scale. That combination helps explain why customers paid a premium and why Nvidia could earn gross margins above 70%.

Huang also coordinated the supply chain behind that growth. A new generation of Nvidia hardware requires chip manufacturing, advanced packaging, memory and networking suppliers to commit capacity well before the systems ship. Huang persuaded those partners to build ahead of demand and helped Nvidia move to an annual release cycle. Suppliers have every commercial reason to keep working closely with Nvidia. The first real test will be whether the new leadership can secure the same commitments and resolve the inevitable bottlenecks while keeping that cadence.

Nvidia’s management structure raises a similar question. Huang had unusually many direct reports, which gave technical leaders a direct line to the person who could make a decision. That helped the company move quickly and place big bets. It also concentrated decisions in one person. Kress brings continuity, especially around customer commitments and capital allocation. Over time, I’ll be looking for the board and the permanent CEO to show who owns product decisions, supplier relationships and export oversight, and whether the company can keep making decisions at the speed customers expect. A stronger board and clearer reporting lines could also make Nvidia more durable. The trade-off is whether additional oversight can catch problems early without slowing the engineering and supply decisions that have made the company so effective.

Washington adds a separate risk. For Nvidia, tighter export oversight could mean more scrutiny of distributors and end customers, slower approvals and a narrower set of markets it can serve. China also matters beyond the revenue Nvidia might lose there. If Chinese developers have to build on domestic chips, their usage gives local suppliers feedback and a reason to improve their hardware and software. Over time, that can help create an alternative ecosystem that competes for customers in other markets as well.

What it could mean for the stock

My starting point is that the leadership change raises the range of outcomes for the stock. Nvidia still has a strong position in AI infrastructure. Customers buying a complete system get chips, networking and software that already work together, which matters when they’re spending billions and want that capacity up and running quickly. That gives Nvidia a real advantage with enterprises, governments and cloud providers. Its engineering teams, installed base and relationships with suppliers also give the next CEO a substantial business to build on.

The pressure is likely to show up first among Nvidia’s largest and most technically capable customers. Google’s TPUs and Amazon’s Trainium show that large cloud providers can build viable alternatives for some workloads. As more developers access AI through a model’s API, many can build products without ever having to think about which chips are running underneath. The work of moving a model to new hardware remains significant, but large customers have the money and engineering talent to do it where the economics justify it. Nvidia can keep growing into a bigger AI market while capturing a smaller share of each incremental dollar. If customers gain bargaining power at the same time, earnings can grow more slowly than revenue as margins normalize.

I’m also watching who funds the next wave of demand. The biggest cloud companies have other businesses that generate cash to invest in AI. The newer, specialized cloud providers rely more heavily on outside financing and on customers actually renting the capacity they’ve built. Nvidia has used its balance sheet to help parts of that ecosystem grow. That can make sense if demand catches up and those customers can finance themselves. If AI revenue takes longer to develop and the new leadership is less willing to extend support, some buyers may order fewer systems. Financially stressed operators could also put used GPUs on the market, adding supply just as Nvidia and its partners are bringing more new capacity online.

The first checkpoint is Nvidia’s November earnings call. The reported quarter mostly reflects decisions made well before the leadership change, so I’ll pay more attention to forward guidance and what management says about the next product cycle, supplier capacity and demand from customers that depend on outside financing. After that, the evidence will come from execution over several quarters. Can Nvidia deliver the next generation on time, maintain pricing as alternatives improve, and grow sales to customers that can fund their own purchases? The board’s findings and its choice of a permanent CEO will also help clarify how much the company needs to change the way it operates. A clean product ramp and steady orders from well-funded buyers would make me more comfortable underwriting the growth ahead. Order delays or discounts would matter more if they appear across several customers, since that would point to pressure beyond the transition itself.

For shareholders, the distinction I’d draw is between a company’s competitive position and what investors are paying for it. Nvidia can remain the leader in AI infrastructure and still deliver a less attractive return if the stock assumes years of exceptional growth and margins. There are two ways that can play out. Lower margins can reduce future earnings even if sales stay strong, and greater uncertainty around those earnings can reduce the multiple investors are willing to pay. If Nvidia keeps hitting its milestones and demand broadens to customers that can pay their own way, the case for a premium remains intact. For now, I’d put more weight on whether Nvidia can sustain its pricing and turn that demand into cash flow than on any single quarter’s revenue beat.

Next
Next

The Midterms: What Changed & What it Means for Your Portfolios