Apple Explores Chipmaker Acquisitions to Boost AI Server Competitiveness

Apple is reportedly considering acquiring semiconductor firms to enhance its in-house development of artificial intelligence server chips. This move appears to be an effort to reduce the company's dependence on Nvidia, as its own AI chips have not met performance expectations.

The Information, citing multiple sources, reported that Apple has approached some semiconductor startups about potential sales and has discussed acquisitions with investment banks. Apple is currently using its own 'M2 Ultra' chip for internal AI servers, but its performance is considered insufficient for demanding tasks. The next-generation server chip, codenamed 'Blackbird,' was initially slated for release this year but has been delayed. Consequently, Apple is renting Nvidia chips through Google Cloud for high-performance computing needs, creating a structure reliant on external infrastructure for large-scale AI models.

This potential shift comes as Apple has been perceived as lagging behind competitors like Google and OpenAI in the AI sector. While historically hesitant about large-scale mergers and acquisitions, the company's stance may be changing, evidenced by its $2 billion acquisition of Israeli audio AI firm Q.ai in January. This strategic pivot may also be linked to leadership changes, with hardware expert John Turner set to succeed Tim Cook as CEO in September.

Bloomberg has also reported that Apple is revising its 'M-series' chip development schedule. The upcoming M6 chip will reportedly only have a standard version, skipping the Pro and Max variants, while accelerating the development of the AI-focused M7 Ultra chip. However, the launch of AI server chips based on this new architecture is not anticipated until 2029.