A：OM Robot Company was listed on the main board of Hong Kong Stock Exchange on December 30 last year and was incorporated into the Shanghai-Hong Kong Stock Connect on March 9 this year. Our company focuses on AI parasitic home mechanical systems, and our main business is to develop and sell robot products for the home environment, which is different from most other industrial robot companies.

A：Our vision is to solve the complicated and trivial tasks in the family through robot technology, so that people can free their hands to do what they like, realize a "lying flat" lifestyle, and let robots do some housework that they don't want to do.

A：At present, our company's main products include the universal humanoid nanny robot H1H1, which was launched in January this year. It has been sold and won favorable comments at CES in North America. It is expected to move towards ordinary terminal families within three years. The world's first tennis robot that can be played by real people; As well as a companion robot that performs well in the Japanese market, the robot provides emotional companionship, similar to electronic cats and dogs.

A：The "brain" plays a central role in robotics, determining its intelligence and ability to adapt to the environment. OM Robotics divides the brain iteration phase into three phases and focuses on the cutting-edge technology of VLA (Visual Language Action Model) combined with the world model, enabling robots to output actions directly from visual and semantic information without human intervention. At present, we are exploring a multi-line strategy of one brain, that is, the same brain can be applied to robots of different shapes, and the overall intelligence level can be improved through data training of multiple actuators.

A：At the perceptual level, OM Robotics has a significant technological lead. We use machine vision control technology to enable the robot to achieve three-dimensional positioning with ordinary RGB cameras, which is beyond the technical level that requires lidar or 3D depth of field cameras. So our technical strength is very strong when it comes to perception algorithms.

A：The second flywheel is the distributed neural network, which treats all robots as interconnected individuals, forming a whole through a bionic communication protocol and role architecture. The third flywheel is robot positioning technology, which has the ability to sense the environment, just like the human cerebellum, which can realize three-dimensional digital twins and accurately identify various parts of the home environment.

A：The One Model 1.7 front stra bar RL model is published on our official website and is one of the most detailed disclosures of brain models among listed companies. The model uses the event model to superimpose upper-level understanding, and helps the robot to make decisions and plans through high-dimensional understanding. The "predicted policy later" channel is uniquely designed, and it does not need to generate global images or videos to predict the future, but compresses key information and transmits it to the lower-level understanding module in the form of video representation. Compared with past models such as physical intelligence and NVIDIA growth system, our model has made a lot of innovations and optimizations in film and television characterization, pre-model training, micro-action structure and successful memory module, which not only significantly improves the accuracy, but also outperforms competitors in resource consumption, robustness and simulation tests, reaching the industry-leading level (SOTA).

A：The role of the understand expert in the middle layer is to convert the information from the upper brain into instruction documents similar to machine instructions, which are then transmitted to the executive layer.

A：The execution layer will output the displacement, vector, angle and other information that the robot can identify, and convert it into specific execution actions such as motor rotation instructions.

A：Our robotic arm can perform a series of daily operation tasks in real machine data, such as carrying clothes, folding clothes, washing dishes, and delicate hand movements, such as folding paper cups and pouring coffee from test tubes. The success rate is higher, the lowest is more than 96%.

A：In a highly dynamic environment, such as table tennis, our nanny robot can achieve a success rate of 300 consecutive returns, reaching 72%. This suggests that our model also outperforms under high dynamic tasks.

A：Our model combines the world model and the BLA full-stack word, creating original film and television representation techniques that are stronger than the North American model. At present, the model has reached the level of twin towers, and the success rate of the four items is very high.

A：At this stage, the company has not launched products directly for the home market because there are several major problems: first, the required data is not available or the amount of data is too small; second, the safety and cost of the robot body, such as high safety requirements to prevent accidents, and the need to control the cost to make the product price moderate, so as to find a suitable positioning between the mobile phone and the automobile market. In order to ensure safety, we first designed the robot to be wheeled, so that it will not fall even if the power is cut off, avoiding the safety hazard caused by the traditional vertical robot that may fall when encountering obstacles. In addition, we independently developed all parts to reduce material costs, and through accurate algorithms to make up for the lack of precision on the hardware, to achieve hand-eye combined operation, to ensure the stability and safety of the robot.

A：We design and manufacture our own planetary reducer and other key components through full stack self-operated components. Although this brings the challenge of precision difference, we have successfully reduced the material cost to $3000 by optimizing the software algorithm to make up for the lack of hardware, thus realizing the cost leadership.

A：Data is crucial for future competitive advantage. In ten years, data will become the most important competitive factor. Therefore, our company has sufficient and high-quality data resources as nourishment, which will provide the company with a significant advantage in the future competition.

A：Our company adheres to the principle of real machine mining, collecting data from the real world to train robots. 90% of the data comes from the collection of real scenes, and 10% of the data comes from scenes that cannot be seen everyday. We hire local workers near the data factory and assembly plant in Huizhou for man-machine collaborative collection. This high-quality data collection method is costly, but can ensure data quality. In addition, we also use the gripper acquisition method without the robot body and the head-mounted device under the first-person perspective to collect heterogeneous data to achieve multi-scene, multi-quality data coverage.

A：We build a data pyramid structure from high to low. First, high-quality data is obtained through man-machine collaborative collection, secondly, the middle layer data collection is carried out using jaws without robot body, and finally, the bottom layer adopts the first-person perspective, and employees wear head-mounted equipment to collect real data in daily life around the clock to ensure that the data required for model training is comprehensive and rich.

A：In the data collection phase, we will filter the data according to the application scenario and commercialization goal. For example, for nanny robots, we collect data in four scenarios: cooking, washing dishes, washing clothes and folding clothes, and think that these scenarios are sufficient to meet the needs of families. Therefore, in the collection, we will focus on collecting high-quality data in these areas.

A：Our advantages mainly have three points: first, reduce costs and increase efficiency, drive hardware through algorithms, design parts by self-research in the whole stack, realize cost control, and already have a large-scale market foundation; Second, it has the advantage of data. It has a large amount of high-quality data collected by real machines, and cooperates with Shenzhen municipal government to build a data collection factory second to none in the country. It will collect a large amount of data by the end of 1 million hours; the third is excellent management and team, composed of founders and executives with rich industry experience. The team comes from top research institutions and large factories, and is supported by well-known external resources.

A：we have three product series, among which s mate tennis robot is our main product, which can realize independent fighting without special field support, and is two years ahead of competitors in technology. At present, the products have been welcomed by the market, and they have performed well in independent stations and Jingdong channels. It is expected that sales will exceed hundreds of millions of yuan this year. In addition, we have also set up shared robot experience points in Shenzhen, Guangzhou and other places to provide scanning and renting playing services. The charging standard is cheap, so that more people can experience professional tennis sparring at a lower price.

A：Yes, our future online teaching software AI coach can replace coaches and sparring partners, analyze the user's power chain through cameras, and use Bluetooth headphones to directly give feedback and guidance.

A：In addition to tennis products, we will also launch a peek ball product because it is suitable for more people to participate, because of its slower ball speed and strong social attributes. In addition, next year we will launch a robotic arm version of the product, mainly for printing Federer's tennis balls, with the goal of beating him and creating a memorable moment.

A：We will also launch a high-end universal humanoid robot with autonomous movement, eye tracking, emotion recognition and other functions, which can provide assistance according to the user's living habits, such as delivering goods and controlling household equipment. At the same time, the original family of special robots Switch Pot will continue to grow, with autonomous learning capabilities, and can perform various tasks through the perception layer.

A：Pet companion robots are currently pre-sold in Jingdong at a price of about 4300 yuan. They have the functions of autonomous movement, emotion recognition and handprint echo. They are suitable for family use, especially in scenes with high pet replacement value.

A：About 0.2 billion families worldwide need babysitting services. According to the penetration rate of 5%, the market size can reach 100 billion US dollars. Although the main force is still nanny products, with the development of the market, pet companion robots are also important directions.

A：The company's total revenue last year 0.9 billion, up about 48% year-on-year, gross profit 54%, gold bar profit 12.77 million. In the next few years, we will increase investment in research and development to attract talent, while controlling costs and ensuring healthy cash flow. In terms of market splitting, the share of the Japanese market is likely to decline, while the North American market will add more channels. In addition, the company plans to use existing funds to build an asset-light operating model, build valuable services around the property security chain, and consider investing in upstream and downstream related areas.

A：Woan is actually a leading brain world company, committed to becoming the leading robotics company in the country. Its future ecological vision is similar to Apple's IOS system, which can not only be developed for third parties, but also serve as the "Apple" of the robot industry, that is, to achieve great value through open platforms and hardware sales.

A：Our product range mainly includes three new products, which are expected to contribute about 0.3 billion of the revenue this year. The company's total revenue target for this year is 1.5 billion, which was 0.9 billion in the same period last year. Of the new 0.6 billion increment, we estimate that half (I. e. 0.3 billion) comes from the sales superposition of these three new products, and the other half comes from the natural growth of switch board's old products.

A：Yes, we have showrooms and videos showing this sports robot, as well as humanoid housekeeping robots. The product has been launched on Jingdong in China for more than a month, and the price is cheaper than that in overseas markets. Moreover, the government has a state subsidy policy, and the final price is less than 10000 yuan. At the same time, overseas independent stations are priced at US $2000, offering two modes of lease or purchase.

A：At present, 95% of our channels are in overseas markets, of which 60% are in Japan, more than 20% are in Europe and more than 10% are in North America. The North American market will also expand in the future, but the proportion of the Chinese market is relatively small. The company attaches importance to cash flow and chooses the target group market with high willingness to pay. In Japan, we have more than 400 dealers and stores on display. In the United States, through the acquisition of a company, we were able to access a large number of offline channels such as Apple Store, costco, Wal-Mart and best buy, which only started to happen this year. In Europe, the German market is growing rapidly, doubling last year.

A：All products are priced around gross margin cost-plus, maintaining a gross margin of 50-55%. At the same time, it will follow the sales model of competitors. For example, in Japan, facing the expensive product of vote, our product price is relatively low, and the monthly cost of the rental model is only about 880 RMB, which is far lower than the price of vote. While ensuring sufficient gross profit margin, the sales model of competitors should also be considered.

A：At present, in the domestic and overseas markets, the leasing model refers to the user who buys the product and pays a monthly fee to obtain the right to use the product, rather than leasing in the traditional sense. If you stop paying the monthly fee, the product will no longer update the iteration. The lease of the stadium, on the other hand, requires a separate negotiation with the stadium and payment of a share.

A：According to estimates, if 0.1 billion people around the world play tennis, the penetration rate reaches 1%, and the market size can reach 2.5 billion US dollars. If it grows to 5%, the market size will reach 10 billion US dollars in five years. At present, the product has been sold worldwide, not just in the Japanese market, and is expected to be on the Amazon platform next year, is expected to achieve a wider market harvest.