Introduction
As AI begins to grapple with the physical world, a fundamental question quietly emerges: how does a robot actually perceive its own position in space? Vision tells it what it sees, encoders report how many degrees its joints have turned, and force and tactile sensors tell it what it has touched. Yet when a robotic hand grasps a soft object, when fingers occlude each other, or when it enters a narrow cavity, all these sensing modalities can fail. The inherent physical limitations of these approaches cause the robot to lose track of its end‑effector position.
This is precisely the problem that Ommo Technologies ("Ommo") – a spatial intelligence technology company – aims to solve.
Funding Round & Use of Proceeds
Recently, Ommo completed a Series A funding round of tens of millions of USD. The round was co‑led by VMS Group (Hong Kong) and a well‑known investment fund, with participation from Kangjun Capital. DianShi Capital acted as the long‑term exclusive financial advisor. The raised capital will be used for technology iteration and mass‑production system development of its core spatial positioning system, as well as to accelerate commercial deployment in embodied intelligence, advanced manufacturing, medical, and other scenarios.
The Origin: From 3D Interaction to a Hard Technical Wall
Ommo’s founder, Minjie Zheng, recounts that his entrepreneurial journey began in university with explorations of 3D interaction technology. He once developed an interactive e‑book that enabled digital interaction with paper books. After founding Ommo, the team initially continued along this technological direction, but encountered a critical hurdle during product implementation: existing mature positioning technologies could not continuously capture unconstrained, high‑precision full‑motion data in space.
"We surveyed various mainstream and niche positioning technologies – optics, millimeter‑wave radar, and others – but found it difficult for any of them to deliver the all‑scenario, high‑quality interactive experience required for consumer‑grade products. Take mainstream optical positioning: it suffers from an inherent physical flaw – once the line of sight is blocked by an object, tracking fails completely. For example, a flexible endoscope used in minimally invasive surgery becomes entirely occluded once inside the body, making optical positioning impossible," says Zheng.
After multiple rounds of论证, the Ommo team decided to develop its own underlying technology and set strict criteria: low hardware cost, support for mass production, and stable output of ultra‑high 3D interaction precision.
The Solution: Permanent‑Magnet Magnetic Field Positioning
The team ultimately chose the permanent‑magnet magnetic field positioning approach. In fact, magnetic positioning has long been an alternative solution in the industry to overcome optical occlusion – but conventional electromagnetic schemes rely on coils with alternating currents to generate magnetic fields. Although they offer "see‑through" capability, they suffer from prominent engineering drawbacks: susceptibility to metal interference and surrounding electromagnetic devices, and the need for massive coils in large‑space deployments, making lightweight and miniaturization difficult.
Permanent‑magnet positioning, while also in the magnetic‑tracking family, differs fundamentally in how the magnetic field is generated. It uses mechanical rotation of permanent magnets to create characteristic magnetic fields. This difference brings two key advantages:
-
It eliminates the large coils and magnetic amplification structures required by traditional electromagnetics, allowing the entire hardware to be extremely compact.
-
It solves, from the ground up, the metal‑interference bottleneck that has plagued conventional electromagnetic positioning.
Zheng explains: "We tackle this from two levels. First, we use low‑frequency quasi‑static permanent‑magnet signals to weaken eddy‑current effects from conductive metals and the resulting secondary magnetic interference. Second, we employ high‑complexity characteristic magnetic field coding to identify and correct environmental disturbances. To put it simply, traditional electromagnetic devices usually rely on regular periodic signals and preset field models – once the waveform distorts, they are completely lost. In contrast, the magnetic signals we design carry unique signatures – think of a 'five‑pointed star' or 'irregular ellipse' with multiple identifiable points. Even if interference occurs, it only deforms local parts of the waveform; the core features remain, and our algorithms can identify distortions in real time and perform error correction."
Essentially, Ommo builds a miniature "indoor BeiDou" in physical space. The host unit – the magnetic field source – contains a continuously rotating permanent magnet, while miniature magnetic sensors capture the magnetic field signals at their locations in real time. Algorithms then compute full 6DoF position and orientation data. Currently, the system achieves sub‑millimeter accuracy within the specified working range. The smallest magnetic sensor measures about 0.8 mm – smaller than a grain of rice – and can be embedded in robot end‑effectors, wearable devices, or medical instruments, making it suitable for narrow, occluded, and continuous‑interaction environments.
Full‑Stack Team & Technical Moat
Around this underlying technology, Ommo has built a full‑stack team covering magnetic field modeling, spatial computation, sensors, precision mechanics, electronic systems, production testing, and quality management. Core members come from Apple, Intel, Riot Games, TTI, Samsung, and other technology leaders, with several members having serial entrepreneurship experience and a proven track record of taking products from prototype to scalable manufacturing.
Although permanent‑magnet positioning appears to rely on mature magnets and MEMS sensors, true commercialization still requires solving multiple engineering challenges: complex magnetic signal generation, non‑metallic precision mechanics, sensor calibration, spatial computation, wireless synchronization, and production consistency. Ommo has spent years integrating these pieces into a repeatable, customer‑verifiable complete system. This cross‑disciplinary accumulation – spanning mechanics, electronics, algorithms, and manufacturing – forms the company's long‑term technical moat.
"In addition, from the very beginning we designed for low‑cost industrialization. Magnet materials leverage the mature supply chain of the new‑energy vehicle industry, and sensor chips come from the well‑established consumer electronics supply chain," Zheng adds.
First Application: Medical Surgical Navigation
Initially, Ommo deployed its permanent‑magnet positioning technology in medical surgical navigation. Zheng explains: "The operating room is a highly controlled environment with near‑zero tolerance for error – its stringent requirements are ideal for verifying the system's fundamental accuracy." The company has established a quality system compliant with ISO 13485 and has conducted years of technical validation with leading global surgical navigation companies.
To date, Ommo’s clients include over 100 medical device companies both domestically and internationally. Several supporting navigation devices are currently in the medical device registration phase, covering neurosurgery, orthopedics/sports medicine, dentistry, and other subspecialties.
Next Frontier: Embodied Intelligence
Beyond medical scenarios, embodied intelligence is Ommo’s current strategic focus. Zheng notes that over the past year, many clients in the embodied AI space have encountered technical bottlenecks during real‑world deployment, as their equipment cannot perform high‑precision flexible operations. For example, in grasping, insertion, and screwing actions, the finer the manipulation, the more easily small pose errors are amplified into slippage or failure upon contact.
"These clients proactively approached us, hoping to use our high‑precision sensors to compensate for operational errors," says Zheng.
At the same time, the embodied industry faces a common pain point: an extreme scarcity of 3D physical manipulation datasets. Large language models can be trained on the vast digital text available on the internet, but the 3D physical operation data needed for robot training has not yet been collected at scale. For instance, when a robotic hand picks up a cup and drops it – was it due to insufficient grip, wrong grasping position, or deformation of the object during grasping? Answering this requires multimodal information including end‑pose data and tactile data.
Currently, the industry uses video, optical motion capture, IMUs, encoders, and force/tactile sensors for data collection. However, during finger occlusion and continuous contact, it remains difficult to obtain continuous, absolute, high‑precision 6DoF trajectories at low cost. This is the new opportunity Ommo sees.
Having been thoroughly validated in the medical field, Ommo’s positioning system – in terms of accuracy and stability – can be migrated to embodied AI to build data‑acquisition platforms. The company's first data‑collection glove is already equipped with medical‑grade permanent‑magnet sensors to capture human hand manipulation postures.
Zheng admits that this market segment "has grown much faster than I initially expected." Since the beginning of this year, Ommo has engaged with multiple embodied‑AI clients and initiated collaborations, while deciding to accelerate the full product portfolio rollout.
Vision: The "Spatial Nervous System" for Physical AI
In Zheng’s view, spatial intelligence is the underlying neural hub for physical AI. Large language models act as the robot's brain – understanding instructions and making decisions. Robotic arms and dexterous hands are the limbs that perform physical tasks. Ommo’s permanent‑magnet positioning system, meanwhile, bridges the real‑time feedback loop between environmental perception and limb execution – "replicating the innate 3D spatial awareness that humans possess."
"We want to give robots a stable 'spatial intuition'. This may well be the indispensable infrastructure for AI to truly step from the digital world into the physical world," says Zheng.
Investor Perspectives
-
VMS Partner, Terence Ng, commented: "Precision, size, and environmental adaptability are often difficult to achieve simultaneously. Ommo’s spatial perception system brings all three into a single solution and has proven itself in extremely low‑tolerance scenarios like surgical navigation. The calibration algorithms and data accumulated over years of real‑world deployment constitute a moat even deeper than the hardware itself. And the occlusion, narrow spaces, and continuous contact faced by robot end‑effectors during grasping, insertion, and assembly are essentially the same challenge. VMS looks forward to a long‑term partnership with Ommo, bringing this capability to broader industries."
-
Kangjun Capital stated: "Transitioning from the stable, ultra‑precise medical domain to the evolving, infinitely generalizable embodied AI field, Ommo has leveraged its engineering and algorithmic prowess to push the boundaries of technology application, becoming a cornerstone for building physical AI. Kangjun Capital is optimistic about Ommo’s technology platform and looks forward to leveraging our industrial resources in life sciences and healthcare to support Ommo’s journey toward cutting‑edge applications."
-
Huang Jun, Partner at DianShi Capital, added: "As Ommo’s long‑term exclusive financial advisor, DianShi Capital has witnessed the entire journey – from original technology and product engineering to cross‑scenario expansion. Ommo’s rarity stems from its original underlying technology, its validation in demanding scenarios, and its potential to extend into the broader spatial intelligence market. We will continue to accompany the company, bridging industrial and capital resources, and firmly believe that Ommo will become a spatial intelligence platform with lasting competitiveness."
