AI | Apple is reportedly considering introducing a technology to run compressed AI models on iPhone

robot
Abstract generation in progress

According to CNBC, Apple Inc. in the United States is in talks with a small company in Silicon Valley, which claims it can compress powerful artificial intelligence (AI) models to a size that allows them to run directly on iPhones.

On Tuesday, PrismML publicly released a compressed version of Alibaba (09988)’s open-source model “Tongyi Qianwen” (Qwen), shrinking the model from about 54GB to under 4GB, so that all 27 billion parameters can run on iPhone 15 or newer models.

PrismML CEO Babak Hassibi said Apple is evaluating PrismML’s technology, which could make the voice assistant Siri faster and further improve privacy protection, because it would allow more AI processing tasks to be completed directly on the device rather than sending instructions to the cloud.

If PrismML’s claims are validated in real-world tests, the technology could reshape market demand for memory chips and data center computing power. However, analysts believe that AI development still depends on large volumes of chip support.

Apple has not commented on this.

Apple has recently released an iOS 27 public beta, enabling iPhone users for the first time to fully experience the Siri system upgrade. Apple is trying to make Siri more competitive when facing comparable products from companies such as OpenAI and Anthropic.

AAPL3.79%
View Original
This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.
  • Reward
  • Comment
  • Repost
  • Share
Comment
Add a comment
Add a comment
No comments