Apple reportedly working on AI chips for data center servers
Apple, which has traditionally not been the front runner when it comes to the AI wave that is sweeping the tech realms right now, is looking to make some headway. A new report by The Wall Street Journal reveals that the company is reportedly delving into the development of its own homegrown chip designed specifically to power AI software and tools within data center servers.
According to reports, the homegrown chip will build upon the tech behemoth’s prior efforts to make in-house chips, which run in its iPhones, Macs and other devices. If Project ACDC – as it is internally called – goes through, then the Cupertino-headquartered Apple will be taking a page out of the books of tech giants like Google, Microsoft, and Meta, all of whom operate data centers that run on in-house designed semiconductors.
Over the past decade, Apple has established itself as a leader in designing chips that power its iPhones, iPads, Apple Watches, and Mac computers. Project ACDC (which has been in the works for years) leverages this expertise, aiming to translate it into the realm of data center servers. While specifics are limited, reports suggest that Apple’s server chip might prioritize running established AI models. This stands in contrast to training AI models, where chipmaker Nvidia currently holds a dominant position. By focusing on this, Apple could optimize its data centers for tasks like real-time analysis and decision-making powered by AI. Media reports suggest that Apple has reportedly engaged in discussions with industry heavyweights like Alphabet’s Google and OpenAI to fuel its AI ambitions.
The emergence of Project ACDC is hardly unexpected, and comes at a time when there is a surge in interest surrounding generative AI (the eye-catching technology underpinning popular chatbots and other AI-powered tools). Apple’s rivals, like Microsoft and Meta, have aggressively invested in this space over the past year, and speculation suggests a potential unveiling at Apple’s forthcoming Worldwide Developers Conference (WWDC). For its part, one potential advantage Apple possesses is its experience in designing custom silicon. Compared to its rivals, Apple boasts a longer history and greater success in building bespoke processing chips. This expertise could translate into a significant advantage in the AI chip race, allowing Apple to develop highly efficient and optimized solutions specifically tailored for its own data center needs.
Still, Project ACDC represents just one facet of Apple’s broader AI strategy. As mentioned earlier, the company is expected to unveil new AI features at its upcoming WWDC in June. These features could be powered by a combination of in-house developed server chips, existing infrastructure, and potentially even partnerships with external AI providers like OpenAI or Google.
For Apple, a successful Project ACDC chip would give Apple a significant edge in developing and deploying its own generative AI models, thus potentially closing the gap with competitors like Microsoft and Meta. By having its own custom AI chip, Apple would gain greater control over its data center infrastructure. This could lead to optimizations in power efficiency, performance, and potentially even cost savings. It could also mean decreased reliance on external AI service providers like OpenAI or Google, and be a major win for Apple in the eyes of investors. Apple’s investors have already grown impatient with the company’s perceived lack of progress in AI, particularly when compared to the high-profile announcements and stock price jumps witnessed at rival companies like Microsoft, and Apple CEO Tim Cook has sought to assure investors that it will have major AI announcements soon.
