Apple Reportedly Looking to Acquire AI Chip Companies

2 hours ago 1
Apple is looking to get AI spot companies arsenic portion of an effort to trim its dependence connected Nvidia for demanding AI workloads, according to The Information.


Apple presently handles immoderate AI processing successful its ain information centers utilizing its ain chips, but relies connected Nvidia hardware housed wrong Google Cloud for much demanding tasks, an statement that includes the Gemini exemplary powering the overhauled mentation of Siri. Apple's ain AI server chip, internally code-named Baltra, had been expected to vessel this twelvemonth but has been delayed, according to radical said to beryllium acquainted with the project.

Apple has historically constricted its acquisitions to deals successful the hundreds of millions of dollars and avoided ample purchases, but that attack appears to beryllium shifting. In January, Apple completed its acquisition of Q.ai, an Israeli institution specializing successful interpreting code done facial micromovements, paying adjacent to $2 billion, 2nd lone to the $3 cardinal Apple paid for Beats Electronics successful 2014.

Apple besides signaled a alteration successful fiscal strategy during its most caller quarterly net call, erstwhile CFO Kevan Parekh told analysts the institution would nary longer people "net currency neutral" status, a argumentation nether which it had kept its currency reserves astir successful enactment with its full debt. Apple did not explicate the reasoning down the change, though the added flexibility could escaped up superior for larger acquisitions.

Apple is already reportedly pursuing acquisitions of AI companies that could assistance shrink AI models for much businesslike usage connected iPhone.

It is besides worthy noting that Apple's ain in-house spot plan capabilities originated with an acquisition. In 2008, the institution acquired PA Semi for $278 million, a woody that laid the instauration for the customized processors present utilized crossed galore of its merchandise lines.

Impending enactment transitions could besides bring a much assertive attack to dealmaking, with hardware main John Ternus acceptable to win Tim Cook arsenic CEO successful September, and spot enforcement Johny Srouji being fixed expanded work implicit each of Apple's hardware engineering, successful summation to semiconductors.

Apple's spot plan squad has traditionally focused connected battery-powered mobile devices alternatively than the high-performance server chips required to vie with Nvidia, the ascendant supplier of AI server hardware. That regulation became evident during improvement of the revamped Siri, erstwhile engineers reportedly attempted to tally Google's Gemini models connected Apple's ain server infrastructure but recovered that the chips, designed with Mac workloads successful mind, could not grip a exemplary of that scale. As a result, Apple was required to process portions of the caller ‌Siri‌'s workload utilizing Nvidia chips wrong Google's unreality infrastructure.

Acquiring extracurricular spot expertise would complement enactment already underway internally. Apple is presently processing a server spot based connected the M5 Ultra chip, according to Bloomberg, portion a aboriginal M7 Ultra chip is reportedly intended to substantially amended AI show to a level that could statesman to rival Nvidia's Blackwell chip. The M7 Ultra is expected to enactment up to 1.5TB of memory, astir treble the capableness of M5 Ultra, though Bloomberg added that a server spot based connected M7 Ultra is improbable to beryllium acceptable earlier 2029.

Acquisitions correspond lone 1 avenue Apple is pursuing to trim its reliance connected Nvidia. The Information first reported Apple's collaboration with Broadcom connected an AI server spot successful 2024, and Broadcom confirmed successful a securities filing past week that the companies had extended that concern done 2031.
This article, "Apple Reportedly Looking to Acquire AI Chip Companies" archetypal appeared connected MacRumors.com

Discuss this article successful our forums

Read Entire Article