Two transformative AI classes are reshaping how organizations operate: digital AI, which reasons over enterprise knowledge to power chatbots and workflow agents; and physical AI, which learns to control robots and autonomous systems from video, gameplay, and sensor telemetry. Both face the same fundamental bottleneck – raw knowledge at scale resides in heterogeneous, private corpora that existing infrastructure cannot access reliably or efficiently.\
We propose Universal Knowledge Perception (UniK) as a common platform covering the full knowledge lifecycle – ingestion, enrichment, indexing, retrieval, and continuous evaluation – across modalities from rich text and video to molecular data and sensor telemetry. UniK is built on Polymath Retrieval, a multi‑index fusion over automatically enriched indices, requiring no task‑specific fine‑tuning.\
Across five digital AI domains (medical literature, open‑domain QA, chemistry, legal video proceedings, and government open data), UniK combined with an open‑source 70‑billion‑parameter model consistently matches or outperforms proprietary frontier LLMs that are orders of magnitude larger: 76% RAG accuracy on government data versus 47% for GPT‑5; 77.9% on medical QA without fine‑tuning; leading all open‑source chemistry pipelines.\
The same infrastructure directly addresses the data curation, indexing, and retrieval challenges of physical AI world‑model training, where the knowledge problem is harder but structurally identical.\
UniK demonstrates a unified knowledge processing path for both digital and physical AI, offering a practical route for enterprises to build reliable, scalable AI capabilities over multi‑modal, large‑scale knowledge.\
Review