2026-11-14 · ← News
Jensen Huang’s exit costs open models both capital and a voice in Washington
Nathan Lambert argues that Jensen Huang’s departure from NVIDIA creates two pressures for open-weight models: weaker political backing and less time for builders to find profitable products. NVIDIA still has a commercial interest in an open ecosystem, but its most visible personal champion is gone.
The image could not be loaded.
Researcher Nathan Lambert identifies two immediate weaknesses for the open-weight ecosystem after Jensen Huang’s departure from NVIDIA. The first is political: allies of open models are retreating in Washington, in his view. The second is financial: without Huang’s backing, builders must prove sooner that their products can make money.
The open-weight ecosystem lost the person who would pick up the phone
Huang supported open models through investment, NVIDIA products and public opposition to overly broad restrictions. In 2026, an open letter on US policy was signed by 25 companies including Microsoft, Meta, IBM and Palantir. NVIDIA also developed the Nemotron family and invested in companies building open models.
Lambert’s argument is more personal than a standard reading of corporate strategy. The ecosystem has lost someone with capital and access who could intervene during a political crisis. NVIDIA still has reason to support customers beyond a few large labs because those customers buy its computing infrastructure. What is uncertain is whether new leadership will accept the same political cost.
Without a patient sponsor, open models must find revenue sooner
Training a leading model requires large amounts of compute, talent and time. An open release does not generate recurring revenue as directly as a paid API. Companies therefore look for income in hosted inference, post-training, enterprise deployment and specialized products.
Huang’s support gave that market more time to search for workable economics. His exit increases pressure for faster product-market fit just as Lambert points to scarce compute and strong competition from closed labs. The political and financial weaknesses reinforce each other: less revenue means less capacity to fund both development and advocacy.
NVIDIA’s commercial incentive may outlive its former chief
The harshest scenario remains an interpretation, not an announced change in NVIDIA policy. The company has an economic reason to preserve a diverse model market. If development concentrates in a few labs with custom chips, those customers may become less dependent on NVIDIA GPUs.
The political claim also needs care. Lambert describes the risk of future restrictions and a weaker coalition, not an enacted ban on open-weight models. His July prediction of action within 6 months was a warning based on unofficial discussions, not a published government plan.
NVIDIA’s budget and a new coalition will show who fills the gap
The first signal will be NVIDIA’s investment in Nemotron, open-weight startups and supporting infrastructure under new leadership. The second will be whether Meta, Microsoft, Amazon and US open-model companies can form a shared political platform.
The third test is commercial: how many projects build recurring revenue without permanent subsidy from a chipmaker. Paying customers would give the political case concrete employers and products. Without them, Washington will mostly hear a safety debate in which closed labs have the stronger machinery.
Lilith's verdict
Open models have lost the person who could bring a cheque and call Washington in the same afternoon. Now they need to show their own cash register and their own contact list.