Arm Holdings revealed new chip blueprints and software tools on Wednesday. The company also modified how it releases those blueprints, which may hasten their adoption.
Due to its energy efficiency, Arm’s technology has become more prevalent in data centers and PCs, where it has fueled the growth of smartphones.
Smartphones continue to be Arm’s largest single market, where the business provides intellectual property to fierce competitors like Apple and MediaTek and Qualcomm, makers of Android chips.
Arm introduced new graphics processing units (GPUs) and central processing unit (CPU) designs on Wednesday, claiming that they are better suited to AI work. Additionally, it will offer developers software tools to make the process of deploying chatbots and other AI code on Arm chips simpler.
But the way those things are sold has changed more significantly. Previously, Arm only provided its technology in the form of specifications or abstract designs, which chip producers had to translate into a real chip blueprint. This was no easy process, considering the arrangement of billions of transistors, which are tiny switches that make up semiconductors.
Arm collaborated with Taiwan Semiconductor Manufacturing Co. and Samsung Electronics Co. Ltd. to provide physical design blueprints for the new devices, which are prepared for production.
Arm is not attempting to compete with its clients, according to Chris Bergey, senior vice president and general manager of the company’s client line of business. Rather, it is attempting to expedite their time to market while concentrating on other components that are becoming more and more crucial to PC and phone chips, like neural processing units (NPUs), which offer the finest AI capabilities.
The importance of that chip component has grown to the point where Microsoft claims its most recent AI features cannot function without it. Currently, Arm does not provide NPU technology for smartphones or computers. Bergey, the company, wants to offer more “done and baked” designs so that semiconductor companies can use them to integrate their NPUs.
“We’re combining a platform where these accelerators can be very tightly coupled,” said Bergey.
