The architecture behind ION's patent estate
Video was built for playback. The file formats, encoding standards, and delivery infrastructure developed over the past three decades were designed to display content on a screen, not to let intelligent systems work with video as data. ION’s video virtualisation patents protect the foundational infrastructure that changes that.
Six granted US patents
Granted across the United States, Australia, Canada, Europe, South Korea, China and Hong Kong.
Earliest effective filing date: 17 August 2007 No known prior art challenge.
Four protected infrastructure primitives
Structural separation
Video structure separated from media sample data. Addressable programmatically without duplicating the original asset.
Dynamic assembly
Orchestration
Segment-level governance
These are not defensive patents around a product feature. They are foundational architecture patents. Any organisation building video virtualisation infrastructure at scale will encounter this patent estate.
What this infrastructure is not
What ION is not
- A CDN or delivery layer. ION sits upstream as the data layer. CDNs deliver what ION virtualises.
- A transcoding service. ION's architecture eliminates transcoding, not provides it.
- A MAM or asset management system. ION integrates with MAM; it does not replace it.
- A metadata extraction or analysis layer alone.
- A synthetic video generation system. ION unlocks existing video; it does not create new content.
What ION is
- A data infrastructure layer that makes video structure programmable.
- A composition primitive enabling dynamic assembly from a singular, protected master source.
- A governance architecture that enforces rights at the segment level at runtime.
- The foundational layer for Video Superintelligence: infrastructure for intelligent systems to compose with existing video.
The next layer of the stack
What this means for infrastructure buyers
Hyperscalers
A protected infrastructure capability around video as data, dynamic assembly at scale, and runtime governance. These capabilities sit outside the reach of conventional video infrastructure.
AI Companies
The infrastructure layer that enables Video Superintelligence: intelligent systems working with existing video as programmable material, assembling outputs from real footage without derivative file proliferation.
Chip Vendors
Optimisation targets that do not yet exist in conventional silicon roadmaps: the data operations required for virtualisation, dynamic assembly, and governed resolution. This points to a new workload class for silicon optimisation.
A defensible infrastructure position
Six granted US patents cover the foundational architecture. A pending filing in Australia and the United States extends that architecture into runtime governance for the agentic AI era. Together, they form a defensible infrastructure position around the central unsolved problem in Video Superintelligence: how to make existing video programmable, composable, and governable at scale, without breaking ownership and control.
Build with ION
Our fastest-growing data type can now be searched, assembled, and composed as intelligent infrastructure.
The foundation exists. The category is defined.
Request Technical Documentation
Take a deep dive into our architecture, APIs, and integration requirements