Ion Video has outlined plans to change the way video is stored and reused by separating a file’s structure from its underlying audio and visual data. This will allow AI systems to assemble new video sequences without having to re-edit or re-render the source material.
The Melbourne-based company described a process in which the internal structure of a video becomes a small virtual representation while the original media data remains unchanged. This virtual file guides the software to rebuild the sequence of frames on demand, rather than creating a newly rendered file every time the user or the system requests a change.
This approach targets long-standing limitations in digital video. AI tools can already analyze footage and identify objects, scenes, or timestamps. Ion Video claims it still struggles to recombine existing clips into new viewing experiences without the traditional workflow of editing, exporting, and saving new versions.
virtual structure
This method treats the video as a single fixed asset, but separates its internal structure from the raw data. Ion Video says its patents cover this separation and the creation of virtual structures that can be referenced without changing the underlying video and audio components.
In reality, the system assembles a series of frames on demand and discards them when playback is finished. This differs from standard workflows where each change creates a new file, increasing storage and computing requirements.
Finbar O’Hanlon, lead innovator at Ion Video, identified a problem with the way video files were designed. He argued that advances in streaming and cloud services will not change the core format of video as a finished asset.
“Traditional video files are designed as finished, rendered assets. Once created, they are static objects intended purely for playback and distribution. Platforms can compress, stream, and analyze video files, but their internal components cannot be easily manipulated or recombined without creating an entirely new file,” O’Hanlon said.
From Lineas
O’Hanlon’s work with video virtualization dates back more than a decade. In 2009 he applied for a series of patents which led to the creation of Lineus Technologies, which was later listed on the ASX. Ion Video is now presenting its work as a redesigned and expanded version of its earlier concept, with additional patents reflecting how AI is used to query and generate media experiences.
He argued that AI systems work best when using data that can be dynamically reconfigured. In his view, text, images, and code fit that model, but video hardly fits.
“Once a video is rendered, it becomes a sealed object. Video is designed for playback and distribution, not for intelligent systems to rearrange, recombine, or compose it,” O’Hanlon said.
prompt video
Ion Video positioned this technology as a way for AI to assemble content according to instructions. In one example, we discussed a user’s request for “5 Asian recipes under $15.” The company says its AI system can scan multiple cooking videos, identify relevant scenes, and assemble a sequence with the necessary steps. For follow-up requests, you can delete comments and keep only cooking actions and completed dishes.
The key point, the company says, is that the system doesn’t create a new render file for each variant. Instead, it references existing data and places it through virtual structures.
“Instead of editing or generating new files, AI models can dynamically assemble content based on instructions and prompts,” O’Hanlon said.
Focus on infrastructure
Ion Video sought to differentiate its business model from a consumer platform. There are no plans to build a direct-to-user video destination. Instead, the company described its technology as an infrastructure that sits beneath the existing video ecosystem.
“Our IP and technology is an infrastructure layer that sits beneath the existing video ecosystem. Its purpose is to help hyperscale cloud providers, AI developers, and streaming platforms innovate and integrate intelligent programmable video capabilities,” O’Hanlon said.
The company tied its pitch to the scale of global video consumption. The report cited an estimate that video accounts for 82% of all internet traffic and pointed to the volume of new uploads to platforms such as YouTube. These numbers are often referenced in industry discussions about bandwidth and storage, but the underlying estimates vary by methodology and time period.
Expense billing
Ion Video says economics will depend on being able to reduce repeated transcoding, storage and computing tasks across large video libraries. O’Hanlon said the company believes it can significantly reduce these processing costs in some scenarios.
“We believe we can save more than 70% in transcoding, storage, and compute costs for video processing,” O’Hanlon said.
He also outlined a licensing approach for large cloud providers, with pricing tied to what he described as “realized value.” O’Hanlon pointed to Alphabet’s infrastructure spending projections and argued that even modest reductions in video spending make sense at hyperscale.
Ion Video said it has filed additional patents to expand its video virtualization intellectual property. O’Hanlon said timing will be critical as AI systems increasingly interpret user intent and generate personalized digital experiences. “For the first time, AI can interact with video the same way it interacts with text,” he said.
