Recently, Hangzhou Lianhui Technology Co., Ltd. (referred to as “Om AI”) announced the completion of a new funding round worth hundreds of millions of yuan. This round was led by Qianhai FOF (Fund of Funds), with participation from Hangzhou government industrial funds and various other capital partners. Concurrently, the company officially open-sourced the VLX-Seek 1.5 on-device native, fine-grained perceptual multimodal model. Delivering powerful performance that surpasses NVIDIA’s core models at the same parameter scale, this model further solidifies the company’s industry-leading position. The funds raised in this round will be used for the technical iteration of the VLX model series, deepening R&D barriers, and accelerating the commercialization of physical AI scenarios.
Currently, the AI industry is shifting from a cloud computing power race to on-device deployment, with physical AI emerging as the next big frontier. The strategic investment by Qianhai FOF fully demonstrates the capital market’s long-term confidence in Om AI’s technical strength, business model, and industry prospects. As a pioneer in the global physical AI sector, Om AI was the first in the industry to propose the concept of an “on-device native” streaming multimodal model for physical AI. Based on its self-developed VLX foundation model, the company has built a full-link closed-loop technology encompassing “continuous perception, precise localization, and action decision-making.” The newly open-sourced VLX-Seek 1.5 model boasts impressive highlights: its 3B parameter version outperforms NVIDIA’s LocateAnything-3B across multiple core metrics, boosting accuracy in drone embodied AI scenarios by 62.9% and reducing target false alarm rates by 74.8%.
On the commercialization front, the company is completing the full-link industrial closed-loop for on-device AI. The “one brain, multiple forms” capability based on the VLX core foundation has achieved large-scale deployment across multiple scenarios. In collaboration with Lenovo and Apple, the company launched “OttoBox AI Studio” tailored for AI PC hardware. It has also partnered deeply with leading embodied AI enterprises to empower various terminal devices with perception, memory, and decision-making capabilities. Additionally, its self-developed wearable AI visual hub, Homer AI, has served nearly 100,000 visually impaired users nationwide, with the platform’s monthly average AI calls reaching tens of millions, placing its commercialization progress at the forefront of the industry.
Zhao Tiancheng, CEO of Om AI, stated: “Looking ahead, the company will continue to drive on-device native technological innovation, remain committed to open source, and empower the developer ecosystem. We aim to collaborate with the entire industry chain to build an open and collaborative physical AI industrial ecosystem, promote the inclusive application of on-device AI scenarios, and inject continuous momentum into the high-quality development of China’s artificial intelligence industry.”