TREE NEWS reports: Apple is developing a Mac app called Harbor that turns Apple-silicon computers into private AI inference servers, allowing users to download, manage and run open-weight models without uploading data to the cloud. The app can link two or more compatible Macs via Thunderbolt 5 to form a cluster and pool unified memory, running larger models or handling more concurrent requests. It is in private preview at Apple’s Claris software unit, with no launch date announced.
Apple reportedly developing Harbor app for local AI model inference on Macs
Apple moving into local AI inference is a notable signal that on-device, privacy-preserving compute is being treated as a product category rather than a developer curiosity. The Thunderbolt 5 clustering detail matters most: pooling unified memory across Macs would let Apple-silicon hardware serve larger open-weight models without cloud dependency, which is relevant to teams weighing data-residency constraints. The open question is whether a Claris-side private preview becomes a shipped, supported product, and how open-weight model distribution and management are handled if it does.
Generated by AI for reference only.
Share on WeChat
Open WeChat → Scan → then tap "…" to send to a chat or Moments.
Tap "…" in the top-right corner to send to a chat or share to Moments.