Nvidia announced on November 12, 2026, that cofounder Jensen Huang had resigned as president, CEO and a board member. Chief financial officer Colette Kress has taken over as interim CEO while the board searches for a permanent successor, according to the company’s official announcement.
The departure removes the executive who helped build Nvidia over 33 years. It also leaves an important strategic question unresolved: how strongly will the company continue backing downloadable AI models and the startups developing them?
Business Insider’s November 17 report connects Huang personally to that effort, including Nvidia’s Nemotron models, startup investments and an ecosystem intended to compete with closed-model providers such as Anthropic and OpenAI. Its argument that his departure reduces that competitive threat is an assessment of the stakes, not an announced change in Nvidia policy.
What Nvidia’s Leadership Announcement Confirms
Nvidia’s SEC filing gives the precise sequence: the board accepted Huang’s resignation on November 11, effective immediately. The company announced it publicly the following day.
The transition includes several concrete changes:
- Colette Kress becomes interim CEO while retaining her positions as executive vice president and CFO.
- The board will consider internal and external candidates for the permanent CEO position.
- Harvey C. Jones becomes Nvidia’s first board chair. The filing says he has served on the board since 1993.
Neither the announcement nor the filing states why Huang resigned. Although Business Insider describes his departure as an ouster, the official documents record a resignation without explaining its cause.
Kress has been Nvidia’s CFO since September 2013. The board’s public message emphasizes continuity: Jones said it expects the company’s “operations, customer commitments and product roadmap to continue without interruption,” according to the announcement.
That assurance matters for customers planning hardware purchases and deployments. It is less specific about discretionary investments, acquisitions or the long-term budget for open-model development. Those decisions are central to the competitive question now surrounding Nvidia.
Why Huang’s Open-Model Push Matters
Business Insider describes a strategy with several connected parts: Nvidia would develop models itself, finance other developers and help make their work broadly available. The goal was an ecosystem in which businesses could obtain and adapt capable AI without depending entirely on a closed frontier-model provider.
Its reporting identifies these components:
| Component | What Business Insider reports | Strategic significance |
|---|---|---|
| Nemotron | Nvidia developed its own family of open models | Gives businesses another source of models they can adapt and operate |
| Startup investment | Nvidia backed Reflection AI, Thinking Machines Lab and Poolside | Supports model developers outside the largest closed-model labs |
| Distribution | Nvidia was moving to acquire Hugging Face | Would connect the strategy more closely to a major model-sharing platform |
The Hugging Face item is a reported acquisition effort, not evidence of a completed purchase.
Replit CEO Amjad Masad underscored Huang’s personal role at a conference on November 16. “Jensen is the premier champion of open models,” he said, according to Business Insider. That is one named executive’s assessment, relayed through the publication, and helps explain why the leadership change has become a funding question as well as a management story.
What Open Weights Change for a Business
An open-weight model makes its learned parameters available for download. As Business Insider explains, these models can be modified and operated independently, giving companies more control over how they deploy AI.
A business might adapt a model for a specialized task and run it within infrastructure it controls. With a closed-model service, the provider generally operates the model and sells access to its capabilities. The customer gets a service without taking on the full deployment job.
That difference changes where costs and responsibilities sit. Self-hosting can replace some provider usage fees with spending on computing capacity, engineering, monitoring and maintenance. It can also give a company more control over deployment and customization.
I’d treat that as an operational choice before treating it as a guaranteed saving. A downloadable model still needs infrastructure and people to run it, and the cited reporting supplies no workload-by-workload cost comparison. “Open-weight” also should not be read as a promise that every release has identical licensing terms or includes every training artifact.
The Competitive Fight Runs in Both Directions
Nvidia’s interest in open models has a straightforward business rationale. If more companies can build useful AI independently, more organizations may need computing infrastructure to train, adapt and serve those models.
Business Insider makes that connection explicitly: a larger open-model ecosystem could create more potential customers for Nvidia’s chips. Under that strategy, Nvidia could benefit even when a model developer or application company captures the end-user relationship.
Closed-model labs face a different incentive. Their businesses depend in part on customers paying for capabilities that are difficult or inconvenient to obtain elsewhere. Strong downloadable alternatives can put pressure on that proposition, particularly for tasks where a business can adapt a model to meet its needs.
That is why Business Insider frames Huang’s departure as favorable to Anthropic as the startup prepares for a stock-market debut. Its November 17 report said Anthropic was preparing to go public that week. The competitive argument is that investors evaluating future AI-service profits must also consider how capable and accessible open alternatives might become.
But the challenge also runs in the other direction. Business Insider reports that Anthropic disclosed software in its IPO filing that lets it shift AI workloads among chips from Nvidia, Google, Amazon and AMD.
That capability would give Anthropic more flexibility over hardware selection. A lab able to move work between suppliers has more options when assessing capacity, price and performance.
The useful parallel is this: Nvidia’s open-model push could reduce customers’ dependence on a small number of AI labs, while the labs’ multi-chip software could reduce their dependence on Nvidia. That is an analysis of the incentives described in the reporting, not proof that either side has already secured an advantage.
A Market-Share Comparison Needs Care
Business Insider also cites a November 14 analysis by JPMorgan Asset Management’s Michael Cembalest estimating that Nvidia represented 39% of leading AI labs’ future computing commitments. The article contrasts that with Nvidia’s share of more than 80% of AI chip revenue in recent years.
Those figures measure different things. Future commitments describe planned spending; historical chip revenue describes sales already made. The figures also cover differently defined markets. They should not be read as a measured fall in Nvidia’s market share from above 80% to 39%.
Frequently Asked Questions
3 questions
1Who Is Leading Nvidia Now?
Colette Kress is interim CEO and remains executive vice president and CFO. Nvidia’s board is searching for a permanent CEO and will consider internal and external candidates.