Apple Vision Pro upgraded with the powerful M5 chip (www.apple.com)

Apple announced an upgraded Vision Pro powered by the new M5 chip — a third-generation 3 nm SoC with a 10‑core CPU, a next‑gen 10‑core GPU that supports hardware‑accelerated ray tracing and mesh shading, and a 16‑core Neural Engine. Compared with the prior model it renders ~10% more pixels, can boost refresh to 120 Hz, and pairs with the R1 sensor hub (12 cameras, five sensors, six mics) to stream visuals with ~12 ms latency. Apple claims up to 50% faster system AI features and up to 2× faster third‑party AI workloads; battery life is improved (≈2.5 hours general, ≈3 hours video). The device ships with a softer Dual Knit Band and starts at $3,499, available for pre‑order and demos now. For the AI/ML community this is meaningful on several fronts: M5’s Neural Engine and on‑device Foundation Models support lower‑latency, privacy‑preserving generative and perceptual features (Apple cites use cases like JigSpace’s natural‑language-driven 3D model exploration). Hardware ray tracing and mesh shading open richer, real‑time spatial rendering for games, simulations, and visualization pipelines, while visionOS 26 adds persistent spatial widgets, enhanced Personas, spatial scenes (generative depth for photos) and APIs that encourage native spatial apps and multimodal agents. Enterprises and creators — from pilot training to medical imaging and interactive product visualization — gain a higher‑performance, lower‑latency platform, but developers will need to target the M5/Neural Engine and new rendering primitives to fully exploit its capabilities.