I really wish these wannabe journalists (and the "tech" sites supporting them) would do their homework, or at least have their articles fact-checked by genuine experts before posting. He really mis-characterizes the effects of heat when he claims things like, Instead of "crunching numbers!" Bull-hockey! Yes, much energy is wasted in the form of heat, but if the device was more efficient, that energy would simply be saved, not taken away from number crunching. Yes, if the device is forced to throttle back in speed to prevent overheating, that is a loss in performance. But that indicates a lack of cooling - a different issue. If you delve into this blogger's sources (which, thankfully, he links to), there you see the real story here and that is the use of new materials that waste less energy. The big news is not really about less heat, but about consuming less energy to do the same amount of work. As one of the source articles points out, information and communications technologies already consume 8% of the world's generated electricity, and that doubles every decade. These new materials would allow devices to do the same amount of work per transistor while using less energy. Not only do they use less energy, but data centers would then need less air conditioning - which again means using less energy. And that is big news! It is important to understand that heat is not the limiting factor when it comes to transistor density. It is about the physical size of the gap in the transistor gates becoming so tiny, it is impossible to prevent the voltages from jumping (arcing) across those open gates. That is a whole different technological challenge from heat generation and energy consumption/loss that these new materials do NOT address. In portable devices (notebooks, smart phones, etc.) it means longer battery run times. Yes, it can mean they can pack more power in those devices because less space will be needed for heatsinks, fans, and ventilation. But each transistor is still doing the same amount of work - and there is nothing in those source articles about packing more transistors in the same amount of space to increase performance. It is all about saving energy. This is still incredibly important and a major breakthrough in advancing the state-of-the-art. I just wish this article put the emphasis on the key points, as the actual researchers and source articles have done.
