Apple is reportedly collaborating with semiconductor giant Broadcom to develop its first server chip tailored for AI applications. According to a report by The Information, which cites three sources familiar with the project, the chip is internally code-named “Baltra” and is expected to be completed within the next 12 months.
Apple has a long-standing reputation for designing its own processors — branded as Apple Silicon and manufactured primarily by TSMC — for devices like iPhones and Macs. However, these chips were not specifically optimized for handling the intensive demands of AI processing.
Recent developments highlight Apple’s growing focus on artificial intelligence. The company recently unveiled Apple Intelligence features for newer iPhones and Macs, including integration with ChatGPT. While much of Apple’s AI strategy emphasizes on-device processing, certain high-compute tasks, such as those for Siri and Maps, rely on cloud infrastructure. The increasing demand for these AI-driven features underlines Apple’s need for more robust backend support. 
This is where the Baltra server chip comes into play. The partnership with Broadcom is particularly focused on advancing the chip’s networking technology, which is essential for connecting devices to networks for efficient AI processing. Broadcom’s expertise in networking components is expected to play a key role in enabling Apple to handle more intensive AI workloads effectively.
Looking ahead, Apple plans to expand its generative AI offerings significantly in the coming years. The development of the Baltra chip marks a crucial step in meeting these ambitions while ensuring seamless performance for cloud-based AI functionalities.
