Apple is evaluating PrismML's technology. According to the startup, the technology can deeply compress high-performance AI models so that they can run directly on the iPhone, reducing memory usage by up to 15 times. This technological breakthrough is expected to allow more artificial intelligence calculations to be performed locally on the phone without sending requests to the cloud, thereby improving Siri's speed of operation and strengthening user privacy protection. If the performance claimed by PrisMML is fully realized in actual tests, the technology may reshape market demand for memory and data center computing power; however, analysts said that the AI industry as a whole will continue to consume large amounts of chip resources.

Apple is evaluating PrismML's technology. According to the startup, the technology can deeply compress high-performance AI models so that they can run directly on the iPhone, reducing memory usage by up to 15 times. This technological breakthrough is expected to allow more artificial intelligence calculations to be performed locally on the phone without sending requests to the cloud, thereby improving Siri's speed of operation and strengthening user privacy protection. If the performance claimed by PrisMML is fully realized in actual tests, the technology may reshape market demand for memory and data center computing power; however, analysts said that the AI industry as a whole will continue to consume large amounts of chip resources.