Apple is set to launch its own AI‑server chips in the second half of 2026, according to TF Securities analyst Ming‑Chi Kuo. The move follows the company’s $500 billion, four‑year investment in U.S. manufacturing announced in February 2025, which includes a new plant in Houston that already ships American‑made servers for its Private Cloud Compute system. Kuo predicts that the chips will power a wave of new data centers built in 2027, a step that signals Apple’s intent to meet a projected surge in on‑device AI demand.

The company has already partnered with Broadcom on a processor called Baltra, slated for 2026, and has been developing an internal project codenamed ACDC to create Apple Silicon optimized for AI workloads. By shifting to purpose‑built server chips, Apple aims to boost performance, cut power consumption, and improve thermal management beyond what its current Apple Silicon offers. If the timeline holds, Apple could become a major player in the AI‑infrastructure market, reshaping how its services scale across the globe.