Apple is said to be working on a new AI server chip. According to popular Apple analyst Ming-Chi Kuo, Apple’s own AI server chips are said to enter mass production in the second half of 2026. Kuo also emphasized that the company plans to start building and operating its own data center in 2027. In a post shared on social media platformApple has long relied on its silicon team to differentiate its products. From the A-series chips in the iPhone to the M-series processors in the Mac, custom designs created by Apple have consistently delivered increased performance and efficiency. Recently, the tech giant also shipped its own cellular modems (C1 and CX) and wireless connectivity chips (N1). The move to AI server-class chips is seen as a natural extension of this strategy, giving Apple greater control over the infrastructure supporting its AI services.
Apple’s AI server chip deployment schedule
– 2026: Mass production of Apple’s AI server chips begins.– 2027: Apple’s new data centers are expected to come online and expand AI workloads.– Short-term: Prior to full deployment, the chip may be deployed on a small scale in existing data centers.
Apple and Google partnership
Separately, Kuo also shared his thoughts on the recently announced partnership between Apple and Google. The iPhone maker has confirmed it is partnering with Google to use its Gemini model as the basis for a smarter, more capable Siri, arriving later this year. “After careful evaluation, we determined that Google’s technology provides the most capable foundation for Apple Foundation Models,” Apple said in a joint statement with Google. Kuo writes about the partnership: “Two near-term challenges for Apple to develop its own AI will force it to work with Google.(1) This year’s WWDC AI demo cannot afford to fail again.(2) With the development of cloud AI, users’ expectations regarding the quality of AI services are increasing. Even if Apple delivered on its past promises regarding Apple Intelligence and Siri, it now seems inadequate and more powerful on-device AI models are needed. ”
