Apple plans to use its advanced M2 Ultra chips in data centers to run complex AI queries, according to a report by Bloomberg. Simple queries will continue to be processed on-device. The move is expected to happen before the release of next generation M4 series chips.
It was earlier reported by The Wall Street Journal that the iPhone maker is planning to develop custom chips to power data centers. The move is to bring higher security and privacy through the project internally code-named as ACDC (Apple Chips in Data Center). However, the company is now planning to use its existing M-series chips for data centers, as it believes the chips meet the expected security requirements.
Existing M-series chips to be used for data centres
As a result, Apple is not planning to develop custom chips for data centers but instead use the already powerful M-series chips for powering its cloud operations. Initially it will just be Apple’s data centers that will be powered by M2 Ultra chips. In the near future, third-party servers will also run Apple-designed chips.
Currently, Apple has servers across the United States and is developing a new center in Waukee, Iowa. The company is trying to catch-up with Google and Microsoft in terms of bringing in generative AI technologies to its services and operations. Last year, the iPhone maker released a machine learning framework named MLX, which improved the efficiency of AI models running on Apple Silicon.
Siri is set to receive major improvements, as Apple has been advancing its AI initiatives and has made major internal team restructures in the recent past. Recently, Apple released its M4 chip which it called “an outrageously powerful chip for AI.” Since the release of ChatGPT, the tech industry and the users have changed their perspective towards AI. It is to be seen how Apple plays catch-up or makes a surprise leapfroging release.
