Jensen Huang plays down OpenAI’s “Jalapeño” chip, saying Nvidia still has the supply and technology customers need

Nvidia CEO Jensen Huang on Wednesday brushed off questions about OpenAI’s newly discussed “Jalapeño” chip, saying he is not taking the project personally and that he remains confident in Nvidia’s ability to support the artificial intelligence industry with its technology and supply.

In remarks carried by Yahoo Finance, Huang suggested that early-stage hardware efforts and new chip announcements do not necessarily translate into long-term deployments. “Lots of projects get started. Lots of projects get canceled,” he said, framing the moment as part of a broader cycle of experimentation in AI infrastructure.

The exchange underscores how seriously Nvidia still treats demand planning and engineering momentum across the AI stack. Even as model makers and large technology customers explore custom silicon and alternative accelerators, Nvidia’s business model depends on keeping its GPUs and related software ecosystem positioned as the default path for training and deploying AI systems.

Huang’s comments also highlight the commercial reality that customers often evaluate multiple routes for compute. While some companies aim to diversify beyond vendor GPUs, the ecosystem built around Nvidia’s CUDA software platform, developer tooling, and data-center integration is designed to reduce friction for teams scaling AI workloads.

For Nvidia, the key issue is not whether competitors can prototype chips, but whether those chips reach production at scale and integrate smoothly into the workflows that enterprises and cloud providers run. Nvidia’s argument, as conveyed in the interview, is that the industry’s hardware landscape is dynamic and that many alternatives do not clear the threshold from concept to sustained volume.

The name “Jalapeño” refers to OpenAI’s chip effort discussed publicly in recent coverage, but Nvidia’s CEO did not describe the chip’s specifications, timeline, or which customers would use it in the near term in the information available here. The same limitation applies to any impact on Nvidia’s order book, as no contract terms or allocation changes were disclosed.

Nvidia also did not provide additional detail on how it evaluates competitive custom silicon internally, or whether it is seeing changes in customer roadmaps as a result of OpenAI’s announcement. That means investors and industry watchers will need more data, such as customer deployment indicates, supply commitments, or guidance changes, to judge whether any new chip initiatives affect Nvidia’s trajectory.

Looking ahead, the market will likely watch for signs that custom AI chips move from announcements to broader deployment, and for any evidence that major customers shift meaningful compute volumes away from Nvidia GPUs. Nvidia’s next earnings communications and commentary from large AI infrastructure customers could offer the most direct indications of whether “Jalapeño” changes the competitive balance.