NVIDIA's CEO apparently looks threatened by the massive adoption of ASICs by tech companies, as they would act as a counter to Team Green's robust AI infrastructure.
NVIDIA's CEO Claims ASICs Won't Be Able To Execute Scale Deployment, But Competition Still Remains Fierce
While Team Green does sort of have a monopoly over the AI training markets, history does show us that such situations don't last for very long, and it seems like NVIDIA's "kryptonite" would likely be the integration of ASICs by its partners. In a post by DigiTimes, it is claimed that Team Green's CEO was asked about whether ASICs (Application Specific Integrated Circuits) would threaten the firm's dominance over the AI markets, and in response, Jensen gave a rather uncertain remark, claiming that they are "non-competitive".
Well, if you are unaware of what ASICs are, they are basically chips designed for a specific workload, or in the case of AI, are made to cater to a functionality like AI inferencing. They have tremendous advantages which we'll discuss ahead, but before that, it is reported that firms like Google, Microsoft, Broadcom, OpenAI and many others are involved in developing their custom chips, not only to challenge NVIDIA's dominance, but to find an alternative to the monopoly in the markets created by Team Green's hardware.
Jensen claimed that even if mainstream CSPs manage to create custom ASICs, they won't have the capacity to deploy them at a large scale, given that it not only requires the right expertise, but apparently many of them have tried and failed to do so. However, this time, with AI becoming mainstream, the industry is indeed exploring options apart from getting NVIDIA's hardware, given that not only is the supply chain exhausted due to relying on a single source, but ASIC has several advantages.
Given that firms like Microsoft and Google develop their custom AI chips, it would allow them to operate a more financially viable approach, since ASICs are tailored for a particular workload, providing far more efficient performance than clusters from NVIDIA. Moreover, firms will have the option of having their own supply, which means that they won't be influenced by market demand and experience delays, ultimately putting them ahead of the curve.
While ASICs sound like the right approach, they actually come with high R&D costs, and due to this, many CSPs prefer Team Green's plug-and-play AI solutions. But, NVIDIA's dominance is still challenged by the emergence of ASIC solutions, and with companies like Broadcom leading the race, NVIDIA would definitely have to consider the competition they will see ahead.
About the author: Muhammad Zuhair is a hardware and technology reporter for Wccftech, specializing in the semiconductor industry and the complex interplay between technology, manufacturing, and geopolitics. His coverage focuses on the corporate strategies and technological roadmaps of industry giants like TSMC, NVIDIA, Samsung, and Intel. Zuhair's expertise lies in deconstructing complex topics such as fabrication nodes (e.g., 2nm process), the economic impact of policies like the CHIPS Act, and the strategic development of AI infrastructure from NVIDIA, AMD and Intel.
Follow Wccftech on Google to get more of our news coverage in your feeds.
