Each venture in the SCI portfolio starts with a granted patent position and a defined product architecture โ not a market hypothesis. We do not announce companies before the IP is secured and the engineering path is clear.
SciPhAI is not a LiDAR company that added AI. It is not an AI company that added a sensor. It is a Physical AI platform in which the intelligence and the sensing systems were designed together from first principles โ because that is the only way to build a system where the sensor produces exactly what the intelligence needs, and the intelligence is built to reason about exactly what the sensor knows.
The insight is simple to state and took thirty years of preparation to engineer. If the sensing system is designed from the start to produce causally-structured data โ data that preserves the relationship between each illumination event and each detected return โ then the AI downstream has something real to reason from. Not statistics. Causation.In 2019 we filed the theory. In 2024 we filed the hardware. The theory specifies what the sensing data must look like. The hardware produces exactly that data. One unified architecture, two patent families, one team that built both.
The intelligence layer is State Navigation โ a causal AI framework that reasons from participation structure rather than statistical pattern matching. It requires PMkl as its native input: a matrix encoding which illumination events caused which detection events, with full causal history preserved.
The sensing layer is the ISS Platform โ the only sensing system designed to produce PMkl natively. Coded Propagation-Steered Illumination (CPSI): no moving parts, no motor, no MEMS. Dual mode: Temporal Direction Encoding (TDE) and Continuous Frequency-Shifting Radiating Pulse (CFSRP) โ direction encoded in timing and in carrier frequency. One detector for 360ยฐ. 100 kHz+ frame rate. Embeddable in vehicle glass. No commercial competitor for the CFSRP mode.
Jensen Huang named the category "Physical AI" at CES in January 2025. physical-ai.com was registered in June 2022 โ two and a half years earlier.
The original conversational knowledge discovery service โ publicly operating since 2011, nine years before Retrieval-Augmented Generation was named as a category by Facebook AI Research in 2020, eleven years before ChatGPT launched in 2022.
For any query, the system assembles a body of knowledge in real time โ building the participation matrix, computing association strengths and conditional occurrence probabilities โ and returns structured answers across multiple modes: novel content, consensus content, informative content, causal content. Users navigate a visual interactive knowledge graph (OSM) and continue the conversation recursively. The system generates structured articles automatically from each assembled body of knowledge.
The Ontological Subject Map โ a navigable knowledge graph derived analytically from the ASM structure โ was deployed in this service from 2011. Google launched its Knowledge Graph in May 2012.
Patent analytics, claim mapping, and IP landscape intelligence for technology companies, patent counsel, and corporate development teams. ATTVC provides the analytical tools to understand what IP terrain looks like โ who owns what, what is covered, where the freedom to operate exists and where it does not.
ATTVC also serves as SCI's external licensing interface โ the commercial arm through which the SCI patent portfolio is made available to companies operating in domains where SCI is not building an operating company. Healthcare AI, genomics, legal intelligence, and other verticals where the State Navigation framework applies but SCI has not yet formed a venture.
For organisations that need to understand the CPSI, TDE, or State Navigation IP landscape before making sensing or intelligence architecture decisions โ ATTVC provides the analysis and facilitates the licensing conversation.
The SCI IP framework โ causal intelligence, participation matrices, State Navigation โ
applies directly to healthcare AI and genomic modelling. In medicine, wrong answers cost
lives: statistical correlation is not sufficient for treatment decisions that require
understanding causal mechanisms. In genomics, genetic codes are an information universe
where causation matters more than pattern frequency.
We do not announce ventures before the patent position is secured and the product
architecture is defined. Licensing is available now through ATTVC for organisations
working in these domains who want to build on the SCI framework rather than wait.