‘GPUs Suck’: Former Intel CEO Slams Data Center Hardware Limitations
At Ai4, Pat Gelsinger says AI’s future hinges on boosting power efficiency, infrastructure, and economics – not simply deploying more GPUs.
Get smart on it
A former semiconductor industry leader argued at a recent conference that current AI hardware is fundamentally limited by power inefficiency and computational constraints, and that breakthrough progress requires improvements across the entire infrastructure stack rather than simply deploying more processors. The discussion highlighted that bottlenecks now exist throughout the supply chain, including chip manufacturing, memory technology, energy availability, data center construction, and networking infrastructure. Industry leaders stressed that economic models for AI infrastructure remain unviable and would need dramatic cost reductions, while energy availability has become as critical a limiting factor as semiconductor advancement. The central argument was that future AI capability depends less on individual hardware components and more on solving systemic challenges in efficiency, power delivery, and the economics of building sufficient infrastructure to support AI development.
Related news
An Amazon data center could have the worst polluting power plant in the country
To power its new West Texas data center, Amazon is investing in the construction of a new power plant that could be one of the largest single producers of greenhouse gases in the US, according to the New York Times. The new gas-burning plant in Pecos County, Texas has received significant investment from Amazon and, at least initially, would not be connected to the state's power grid. Instead, its 35 natural-gas turbines would primarily deliver its 7.65 gigawatts of electricity
Do Data Centers Really Drive Up Local Temperatures?
Do data centers raise local temperatures? Explore their impact, causes of heat islands, and solutions to minimize environmental effects.
