AI & TechArtificial IntelligenceBigTech CompaniesBusinessNewswire

Apple pursues AI chip deals to boost its servers

▼ Summary

– Apple is seeking AI chip acquisitions, a rare move for the company, due to its own AI servers using M2 Ultra chips struggling with tasks like powering the new Siri.
– A powerful internal server chip codenamed Baltra has been delayed, with a rival to Nvidia potentially not arriving until 2029.
– Apple is preparing to spend more on acquisitions, signaled by its finance chief dropping a long-held goal to hold as much cash as debt.
– Apple has already spent about $2 billion on Israeli AI startup Q.ai, its second-largest deal ever, and is in talks with startup PrismML.
– The company is also pursuing partnerships, committing to $30 billion of chips from Broadcom, as it seeks alternatives to Nvidia and addresses industry hardware constraints.

Apple built a multi-trillion-dollar empire on custom-designed chips, but when it comes to the high-powered processors required for artificial intelligence, its own creations are falling short. Now, the iPhone maker is taking an unusual step: it’s going shopping.

According to The Information, Apple has recently engaged with bankers and approached AI chip startups about potential acquisitions. This marks a rare departure for a company that has historically shied away from major buyouts. The root of the problem is clear: Apple’s internal AI servers, powered by its own M2 Ultra chips, are struggling to keep up with demand. The heavy computational work behind the new, Gemini-enhanced Siri is being handled instead by Nvidia chips inside Google’s cloud. Apple tried to rely on its own hardware, and it simply wasn’t sufficient.

The company had a server chip of its own, codenamed Baltra, expected to debut this year. That timeline has slipped. Bloomberg reports that a chip capable of rivaling Nvidia may not arrive until 2029, with an intermediate M5 Ultra upgrade intended to bridge the gap. This delay forces Apple to explore alternatives it would rather avoid.

Buying its way out of this situation would break Apple’s traditional playbook. Its largest acquisition to date was the $3 billion Beats purchase in 2014. Its entire chip division was built on a far smaller deal: the $278 million acquisition of PA Semi in 2008. Even this year’s spending looks modest by comparison. Apple paid around $2 billion for the Israeli AI startup Q.ai, its second-largest deal ever.

Several signs suggest Apple is now ready to open its wallet further. CFO Kevan Parekh recently told analysts the company would abandon its long-standing goal of holding as much cash as debt. That shift frees up significant capital, and Apple had $45.6 billion in cash at the end of March. Additionally, Tim Cook is set to hand the CEO reins to hardware chief John Ternus in September, while chip chief Johny Srouji now oversees all of Apple’s hardware. Both are engineers, and both may be more inclined to acquire their way out of a bind rather than wait for internal solutions.

Acquisitions aren’t the only path forward. Apple is also in talks with PrismML, a startup that specializes in shrinking large AI models to run directly on an iPhone. The company has also committed to $30 billion in chips from Broadcom, a partnership now extended through 2031.

All of these moves point in the same direction: Apple wants to reduce its reliance on Nvidia and escape the broader memory and hardware constraints reshaping the industry. Designing its own silicon remains the long-term goal. Buying someone else’s technology might provide the shortcut. For a company that prides itself on independence, needing either option is the real story.

(Source: The Next Web)

Topics

apple ai chips 95% chip acquisitions 90% nvidia dependency 88% server chip struggles 85% financial strategy 82% leadership changes 80% ai model optimization 78% broadcom partnership 76% past acquisitions 74% gemini-powered siri 72%