According to market intelligence, Morgan Stanley released a research report on April 10th. By tracking unusual capacity movements deep within the supply chain, the report uncovers Apple's extensive plans in the private cloud computing (PCC) sector. It reveals a strategic shift for Apple, moving from reliance on third-party cloud services towards significant internal capital expenditure on heavy assets. This provides a crucial reference point for assessing Apple's future cost control and profit trajectory.
Massive wafer orders expose AI ambitions. Apple is quietly securing substantial advanced packaging capacity from Taiwan Semiconductor Manufacturing. Apple's orders for TSMC's SoIC (System on Integrated Chips) are projected to reach 36,000 wafers in 2026 and surge to 60,000 wafers by 2027. This figure is highly anomalous; for context, the current largest SoIC customer, AMD, has a demand of only 42,000 wafers for 2026. Based on IDC data, high-end Mac shipments constitute a very small portion of Apple's annual sales, with actual SoIC consumption estimated at no more than 1,600 wafers. Clearly, such a massive volume of wafer orders is not intended for selling computers.
Targeting the "Baltra" Self-Developed Compute Server. Having ruled out conventional consumer-grade chips, the destination for this expensive, advanced capacity becomes the focus. Morgan Stanley's analysis points directly to a covert internal Apple project—a self-developed AI server chip (internal codename "Baltra") specifically designed for private cloud computing. "Based on the absolute scale of Apple's orders... we believe the majority of the SoIC capacity is allocated for Apple's 3nm AI ASIC ('Baltra') for its private cloud computing (PCC). This will replace the currently used M-series Ultra processors to achieve better AI inference performance and efficiency."
Capital Expenditure Shift and Core Uncertainties. This aggressive bet on foundational hardware not only corroborates previous rumors of a multi-year collaboration between Apple and Broadcom but also sends a clear signal to Wall Street: Apple has a high degree of confidence in an imminent explosion of inference demand for Apple Intelligence. Most AI workloads will be locked into running on Apple's own chip-based servers, shifting its AI infrastructure expenditure from traditional operational expenses (Opex) towards longer-term capital expenditures (Capex). However, behind this grand narrative of vertical integration, Morgan Stanley maintains a note of caution: although the strategic intent to pursue cost control is clear, the actual performance and energy efficiency of Apple's self-developed AI ASIC at scale remain the biggest unknowns in this high-stakes computing gamble.
Comments