Rumor
Apple plans mass production of proprietary AI server chips in late 2026 and data center build-out in 2027
Tuesday, January 13, 2026 at 12:43 PM
Apple is reportedly planning to mass-produce its proprietary AI server chips starting in the second half of 2026. This initiative is expected to coincide with the establishment and operation of Apple's own data centers beginning in 2027, signaling a projected increase in demand for on-device AI integration.
Context
Apple’s custom AI server chips, codenamed Baltra, are set to enter mass production in the second half of 2026 through a strategic partnership with Broadcom. This move follows the accelerated launch of a Houston, Texas facility that began shipping server hardware in late 2025. These internal chips, utilizing TSMC’s advanced 3nm process, are designed to power the Private Cloud Compute infrastructure, eventually replacing the current reliance on high-end M5-series silicon to support the "Apple Intelligence" rollout.
The strategy is part of a $600 billion U.S. investment plan, with a massive expansion of proprietary data centers slated to come online in 2027. By reducing dependency on third-party hardware from Nvidia, Apple gains greater control over AI performance and user privacy. This vertical integration is expected to unlock more sophisticated server-side AI features by 2027, potentially triggering a significant upgrade cycle for the iPhone and Mac as advanced intelligence becomes a core consumer offering.
Related Companies
Apple
AAPL