Jensen Huang on Why ASICs Can’t Challenge NVIDIA’s AI Reign: It’s a “Team Battle”
The ongoing rivalry between NVIDIA and ASIC manufacturers, such as Google and Amazon, has once again been highlighted by NVIDIA’s CEO. As these tech giants unveil their latest solutions, debates intensify over whether NVIDIA’s innovations can be replaced, especially as the industry shifts from training to inference workloads.
NVIDIA’s Unmatched Innovation
Reports suggest that NVIDIA’s CEO, Jensen Huang, addressed the issue of ASICs during a recent Q3 earnings call. When asked about the role of AI ASICs or dedicated XPUs in architecture build-outs, Huang emphasized that the true competition lies not with companies but with the engineering teams behind the technology. According to him, few teams globally can match the complexity and innovation of NVIDIA’s offerings.
Question: Jensen, the question for you. As you think about the Anthropic deal that was announced and just the overall breadth of your customers, I’m curious if your thoughts around the role that AI ASICs or dedicated XPUs play in these architecture build-outs has changed at all? Have you seen, I think you’ve been fairly adamant in the past that some of these programs never really see deployments.
NVIDIA’s CEO: Yes. Thank you very much, and I really appreciate the question. So first of all, you’re not competing against teams — excuse me, against a company, you’re competing against teams. And the — there just aren’t that many teams in the world who are built — who are extraordinary at building these incredibly complicated things.
The mention of the recent Anthropic deal, which included infrastructure centered around Blackwell and Rubin systems, highlights the growing interest in custom silicon solutions. At the same time, Anthropic’s agreement with Google for their latest Ironwood TPUs throws light on the potential competition ASICs could bring to NVIDIA. Huang asserts that while companies develop custom silicon, the challenge lies within their engineering prowess compared to NVIDIA’s.
NVIDIA’s Dominance in AI
Huang confidently claims that NVIDIA excels in every AI sector, including pre-training, post-training, and inferencing, reinforcing the company’s position as ‘irreplaceable’ across these domains. He argues that random ASICs in CSP data centers are a suboptimal choice when compared with NVIDIA’s diverse technology stack, further implying that the custom silicon developed by Big Tech cannot yet rival NVIDIA on an engineering level. Even if these companies manage to replicate NVIDIA’s computing capabilities, he points out that their robust CUDA software stack continues to command industry attention.