{"entity":{"display_name":"Ryzen AI Embedded X100","kind":"series","slug":"ryzen-ai-embedded-x100"},"facts":[{"id":"f_258a70d4d98953c007c8b388","predicate":"announced","value":{"event":"ECS 2026","what":"AMD put a Unified Memory architecture at the center of its physical AI strategy, with the NPU and CPU sharing one memory space and the GPU sharing the same pool"},"as_of_date":"2026-09-21","first_seen_at":1789963160,"last_verified_at":1789963160,"confidence":0.9,"sources_count":1},{"id":"f_012063a3efeaf2e2a5dbaaa0","predicate":"mention","value":{"text":"AMD frames the physical AI problem as a bandwidth and latency problem rather than an inference throughput one, pitching its integrated memory design against external accelerator setups on control loop speed"},"as_of_date":"2026-09-21","first_seen_at":1789963160,"last_verified_at":1789963160,"confidence":0.9,"sources_count":1}],"relations_incoming":[{"id":"rel_b77ca9fef857e1ee90f926f1","predicate":"produces","subject_slug":"amd","value":{},"as_of_date":"2026-09-21","first_seen_at":1789963160,"confidence":0.9,"sources_count":1}],"relations_outgoing":[],"updated_at":"2026-09-21T06:10:08.000Z","verification_summary":{"contradicted":0,"disputed":0,"overall":"unverified","supported":0,"unverified":3}}