Apple is reportedly working with Broadcom to develop the first server chip designed specifically for artificial intelligence (AI) processing. The new chip, internally codenamed “Baltra,” is expected to be in mass production by 2026, according to sources cited by The Information.
The move reportedly puts Apple in partnership with big tech companies such as Google, which are developing custom chips to power AI-driven services. In doing so, Apple aims to reduce its dependence on Nvidia’s expensive and rare processors.
To manufacture this chip, Apple plans to utilize one of Taiwan Semiconductor Manufacturing Co.’s (TSMC) advanced manufacturing processes known as N3P. The partnership marks a strategic shift as Apple continues to enhance its AI capabilities across its devices and services.
Broadcom stock rose 5% on the news. The company has already benefited greatly from the generative AI boom, with its stock increasing 54% in 2024 after nearly doubling from the previous year.
This isn’t the first partnership between Apple and Broadcom. Last year, the two companies signed a multibillion-dollar agreement to develop 5G radio frequency components. Earlier this year, Apple announced at the Worldwide Developers Conference that it intends to use its server chips to power the AI capabilities of its devices.
According to the report, Apple has had great success in chip development and is replacing Intel chips in Mac computers with M-series processors. But despite internal advances, companies like Apple and Microsoft are still finding it difficult to completely break away from Nvidia, which remains a major player in the AI chip market. Google, which is also working with Broadcom on AI chips, is a notable exception.
Marvell, Broadcom’s main competitor in this space, estimates that the custom chip market could grow to about $45 billion by 2028, with both companies potentially sharing the market. .
(Information provided by Reuters)
