Stay connected with BizTech Community—follow us on Instagram and Facebook for the latest news and reviews delivered straight to you.
Apple is exploring generative artificial intelligence (AI) to boost the speed and efficiency of designing its custom chips, according to private remarks made by Johny Srouji, Apple’s senior vice president of hardware technologies. Speaking in Belgium while receiving an award from semiconductor research firm Imec, Srouji said Apple sees generative AI as a powerful tool to compress chip design timelines and increase productivity.
“Generative AI techniques have a high potential in getting more design work in less time, and it can be a huge productivity boost,” Srouji said. He emphasized that Apple’s ability to stay ahead in silicon innovation depends on leveraging the most advanced tools available, especially from EDA (electronic design automation) companies like Synopsys and Cadence Design Systems, both of which are integrating AI into their platforms.
From A4 to Vision Pro
Srouji’s remarks offered a rare look into Apple’s internal chip development history, starting with the A4 chip used in the iPhone 4 back in 2010. Since then, Apple has expanded its custom silicon program across virtually all its product lines — iPads, Apple Watches, Macs, and most recently, the Vision Pro headset.
He highlighted the lessons Apple has learned over the years: moving fast, betting big, and not looking back. That was evident when Apple abandoned Intel chips in 2020 and transitioned its entire Mac line to Apple Silicon without a fallback plan — a move that has since proven highly successful.
Project Baltra: Building Apple’s AI Server Chip
In a move that signals Apple’s broader ambitions in cloud-based AI, the company is quietly developing a server-class AI chip known internally as “Baltra.” Built in partnership with chip supplier Broadcom, Baltra is designed to handle backend AI tasks too intensive for iPhones, iPads, or Macs.
Baltra will support Apple’s “Private Cloud Compute” system — a server-side architecture that processes AI queries securely and anonymously. It will underpin key services in Apple Intelligence, the company’s new AI framework launched across its devices in 2025. Unlike on-device chips that power real-time AI features, Baltra will be deployed in Apple’s data centers to handle heavier AI inference tasks.
Why AI and Hardware Go Hand-in-Hand for Apple
Apple’s approach to AI is rooted in its long-standing philosophy: own the full hardware-software stack. That’s why building its own AI chips — both for devices and servers — is essential. Server chips like Baltra allow Apple to control performance, energy efficiency, and most critically, privacy.
Srouji made clear that while AI may grab headlines, it’s the integration into hardware and design workflows that determines real success. With Apple betting heavily on on-device and cloud-based AI services, the ability to rapidly iterate custom silicon using generative AI could be the next major lever for maintaining its lead in both user experience and chip innovation.
