Apple is starting to incorporate generative artificial intelligence into its chip design processes. During a recent speech in Belgium, Johny Srouji, Apple’s hardware chief, emphasized the company’s interest in AI to streamline chip development as semiconductor technology becomes increasingly complex. He highlighted the potential for generative AI to enhance productivity by enabling designers to accomplish more in less time. Srouji’s comments also shed light on Apple’s reliance on third-party electronic design automation (EDA) tools, which are crucial for chip development.
Major EDA firms like Synopsys and Cadence are actively seeking to integrate AI into their tools, which further supports Apple’s chip design efforts. In detailing Apple’s evolution in hardware, Srouji recounted the journey from the A4 chip in the iPhone 4, introduced in 2010, to current technologies such as those found in the Vision Pro headset. He pointed out that the real challenge lies in the design phase, emphasizing the importance of coordinating hardware and software. He believes AI could facilitate this coordination, making the chips both faster and more reliable.
Moreover, in late 2024, Apple began a collaboration with Broadcom to develop its first AI server chip, internally named “Baltra.” This initiative aligns with Apple’s strategy to support AI services, enhancing capabilities across its devices. Unlike chips that operate on-device, Baltra is expected to serve Apple’s private cloud infrastructure, managing heavier AI workloads. Apple is striving to balance privacy concerns with the need for robust AI features.
While some AI functions will operate directly on devices, others, powered by server-based chips like Baltra, will rely on Apple’s “Private Cloud Compute.” Srouji noted that Apple is no stranger to significant hardware risks, recalling the transition from Intel to Apple Silicon with no backup plan. This bold approach is similarly applied to AI chip development as Apple seeks to refine its design process. The integration of AI in chip design will necessitate new talent, particularly engineers proficient in both hardware and machine learning.
Although Apple will continue to partner with manufacturers like TSMC for production, the design process is increasingly moving in-house. Ultimately, Apple aims to gain tighter control over its entire technological ecosystem, encompassing hardware, software, and AI infrastructure.