Intel's new Loihi 2 neuromorphic processor is one of the first it has created on a 4 nm node

https://swellower.blogspot.com/2021/10/Intels-new-Loihi-2-neuromorphic-processor-is-one-of-the-first-it-has-created-on-a-4-nm-node.html

Intel has uncovered the Loihi 2, the second-age of what it calls neuromorphic chips. The first of these exploratory chips was manufactured on Intel's 14 nm process in 2018 yielding a 'neuron' count of 131,000 for each 60 mm2. The chip is designed to be scalable and was joined with 768 chips in Intel's custom Pohoiki Springs rack system for a sum of 100 million neurons at 300 watts. As per Intel, this system has the processing force of a hamster cerebrum. 

https://swellower.blogspot.com/2021/10/Intels-new-Loihi-2-neuromorphic-processor-is-one-of-the-first-it-has-created-on-a-4-nm-node.html

The new Loihi 2 has been created on Intel's impending 4 nm EUV node and was an ideal chip design to help Intel test out its new manufacture innovation. It supports 1 million neurons for each 31 mm2 with neuron density increasing 15 times. While process nodes are significant, just as significant is the manner by which dense the transistors (or in this case neurons) are. This is one reason why Apple's A15 Bionic created on TSMC's latest 5 nm innovation is superior to both the Snapdragon 888 and Exynos 2100 which are manufactured on Samsung's less transistor dense 5 nm process. 

https://swellower.blogspot.com/2021/10/Intels-new-Loihi-2-neuromorphic-processor-is-one-of-the-first-it-has-created-on-a-4-nm-node.html

The Loihi 2 has been created by Intel Labs, and presently is just being used by different researchers. The organization, nonetheless, does ultimately see it advancing into business scenarios where artificial intelligence is required. It differs from the neural cores we see in versatile chips today which still process instructions sequentially. Instead, neuromorphic chips like Loihi 2 reevaluate this design and use a completely incorporated memory and register model inspired by the cerebrum. To this degree, the neurons work on the whole as a unique system similar to quantum processing in some ways, yet using a customary chip design model.