NVIDIA's CEO has once again commented on the battle with ASIC manufacturers, such as Google and Amazon, claiming that there are not many teams out there that can do what Team Green does.
NVIDIA's CEO Says That No Engineering Team Could Match His Company's Pace of Innovation
The debate around NVIDIA vs. ASICs has intensified ever since companies like Google have announced their latest solutions, with the idea that NVIDIA's tech stack is replaceable as the world transitions from training to inference workloads. At the latest Q3 earnings call, CEO Jensen Huang talked about the ASIC buildout within Big Tech, and when asked about whether these programs would lead to actual large-scale deployment, here's what NVIDIA's CEO had to say:
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.
If you're still confused about what Jensen means here, he was referring to the recent Anthropic deal, which included an infrastructure buildout centered around Blackwell and Rubin systems. This partnership came at the same time Anthropic also signed a deal for Google's newest Ironwood TPUs, raising fresh questions about whether ASICs could realistically compete with NVIDIA. Jensen commented on this inference made by the analyst, posing the question that when companies develop custom silicon, it isn't actually them versus NVIDIA; rather, the core competition lies within the engineering teams.
Jensen says that there aren't many teams out there that could put in the work NVIDIA does, which is one of the reasons why the firm denies the heating rivalry with ASICs. I recently discussed Google's TPUs as a competitive option in the world of inferencing, but Jensen claims that NVIDIA is the best in every AI segment, which indicates the firm's commitment to being completely 'irreplaceable' across all three fronts of the AI industry, being pre-training, post-training, and inferencing.
Interestingly, Jensen notes that when it comes to a CSP, placing a 'random ASIC' in data centers is a far less optimal choice compared to opting for NVIDIA's technology stack, as the firm's offtake is diverse. So ultimately, Jensen does believe the custom silicon from Big Tech still hasn't reached a level where it could compete with NVIDIA on an 'engineering' level. And, in an ideal world, even if they manage to replicate computing capabilities, Team Green has a robust software stack by the name of CUDA, which ultimately attracts all industry spotlight.
Follow Wccftech on Google to get more of our news coverage in your feeds.
