NVIDIA's monopoly over the AI markets is now being questioned by several analysts, who claim that increasing GPU power consumption won't be sustainable in the future.
NVIDIA Might Have To Opt For More "Power Efficient" Solutions If They Want To Continue At Existing Pace, Says Analyst
We all seem well aware of how NVIDIA has played its cards out in the AI segment, whether through gaining popularity among clients through its hardware arsenal or even witnessing immense interest from startups.
Everyone out there eyes NVIDIA as the primary source of their hardware, but it looks like some are more worried about the power consumption, as not only has the firm's share pricing dropped the largest in its history, but many analysts see the current approach as unsustainable and here is why.
Korean media reports that while NVIDIA has managed to see widespread adoption of its cutting-edge AI accelerators, the fact that concerns analysts out there is the growing power consumption. While average consumers know a data center equipped with thousands of NVIDIA accelerators as something "amusing," it is quite the opposite regarding the environmental effects. Team Green's A100s and H100 AI GPUs are known to be power-hungry, and when combining millions of them, imagine the power they would require.
A semiconductor analyst believes global data center power consumption is expected to increase to 85 to 134 terawatt hours by 2027, equivalent to what nations such as the Netherlands, Argentina, and Sweden consume. The analysts believe that if NVIDIA were to prioritize such power-hungry GPUs into the future, low-power alternatives would eventually gain traction, potentially outrunning NVIDIA, which is alarming for the company regarding its established monopoly.
In previous coverage, we disclosed how NVIDIA's existing AI accelerators are rumored to have been consuming up to 13,000 GWh of electricity annually, and this number is expected to increase significantly moving into the future, especially with the debut of Blackwell GPUs which are rated at over 1000W (for specific configurations).
There have been talks about nuclear-powered data centers as well, prominently mentioned in a project being developed by Microsoft and OpenAI, but that is still years ahead from now, and reaching a temporary solution will become a "must" for NVIDIA if they continue at this pace.
There are several other reasons why NVIDIA's monopoly won't exist in the future, such as the increase in competition, the passiveness of companies involved in the supply chain, and much more, but outrageous power consumption would likely prevent the company from sustaining itself in the longer run, given that they do not work on a solution.
News Source: Chosun Biz
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.
