Within the present section the place competitors in giant fashions is shifting from parameter scale to real-world deployment functionality, a bunch of Chinese language firms centered on edge AI is gaining consideration, and Om AI Expertise is one among them.
Based in 2021, the corporate has chosen to not pursue extraordinarily giant cloud-based fashions, however as an alternative focuses on edge-side general-purpose multimodal imaginative and prescient fashions, aiming to carry AI into actual units similar to PCs, cameras, and robots.
On the BEYOND Expo 2026 media day, Om AI Expertise showcased its AI-native content material creation product OttoBox AI Studio. Designed for media professionals and content material creators, it leverages native AI computing energy to supply capabilities similar to video evaluation, asset matching, script technology, and fast video manufacturing.
The corporate positions it as a content material creation companion for the AI-native period, aiming to enhance inventive effectivity.
In contrast with many AI firms that transfer from general-purpose fashions into utility layers, Om AI begins from a extra industry-driven basis. The workforce has lengthy been deeply concerned within the media and audiovisual {industry}, and due to this fact emphasizes constructing fashions based mostly on real-world issues reasonably than searching for nails with a hammer.
Dr. Zhao Tiancheng, CEO of Om AI, famous that long-term {industry} expertise not solely helps the workforce deploy fashions quicker, but additionally offers entry to giant quantities of high-quality real-world information. Of their view, true multimodal functionality is not only about recognizing photographs and textual content, however about understanding video, audio, and textual content concurrently.
One of many firm’s key technical focuses is video understanding underneath low-parameter fashions. In contrast with conventional approaches that depend on extraordinarily giant parameter counts and cloud-based GPU assets, Om AI emphasizes a small, exact, and quick edge-model strategy.
By decreasing mannequin measurement, AI can run straight on native units, decreasing inference prices and decreasing information add necessities, whereas additionally addressing enterprise issues round information safety and privateness.
This edge deployment benefit is especially vital in large-scale video evaluation situations. The corporate states that its fashions can obtain millisecond-level inference pace, making them appropriate for real-time purposes similar to safety, industrial inspection, and AIoT analytics.
At present, Om AI’s AI enterprise spans three main areas: AI PCs, AIoT, and embodied intelligence. Along with collaborations with Apple, Lenovo, and HP, its fashions are additionally utilized to robots, robotic canines, and drones, enabling these units to realize autonomous decision-making and motion capabilities.
Om AI can be exploring inclusive AI purposes. For instance, its Homer App, designed for visually impaired customers, permits object search and assisted navigation by means of smartphones or AI glasses.
The flagship model of OttoBox AI Studio has already established deep partnerships with main PC producers similar to Apple, Lenovo, and HP, finishing its deployment within the AI PC house and offering skilled customers with an out-of-the-box, ready-to-use expertise.
This 12 months, the corporate’s key strategic precedence is the launch of its next-generation edge multimodal mannequin VLX, which goals to additional enhance video understanding and decision-making whereas repeatedly decreasing operational prices.
Because the AI {industry} shifts from cloud-based competitors towards on-device deployment, firms like Om AI have gotten key drivers of real-world multimodal AI adoption.
