Major home appliance manufacturers, capital firms, and humanoid robotics startups are actively focusing on the Chinese commercial kitchen. According to one industry forecast, the AI kitchen robot market could exceed 109.6 billion yuan by 2030.
Major home appliance manufacturers, capital firms, and humanoid robotics startups are actively focusing on the Chinese commercial kitchen. According to one industry forecast, the AI kitchen robot market could exceed 109.6 billion yuan by 2030.
Although humanoid robotics attracts significant public attention, a quieter but rapidly growing robotics race is unfolding in the Chinese commercial kitchen. In one week, kitchen robot specialist Zhigu Tianchu closed a strategic funding round of nearly 100 million yuan, led by China Merchants Venture Capital. Furthermore, Haier Robotics and Robotera announced a partnership to create CR3, which is positioned as the world's first fully autonomous home kitchen robot with AI. SoftBank Robotics also entered into a major strategic agreement with Busypace, an AI-based food preparation technology firm from Beijing, for joint development of platforms integrating food, physics, and AI.
There are three key trends intersecting right now. Firstly, restaurants require process standardization and reduced staff burden: the Ministry of Human Resources and Social Security of China has listed chefs among professions with the greatest labor shortages for the third consecutive year. Once the payback period for a kitchen robot drops below twelve months, restaurant chains transition from testing to mandatory acquisition.
Secondly, advancements in AI have eliminated technical barriers, as kitchens are unstructured environments that require robots to adapt to various ingredients and dynamic tasks. Thirdly, the embodied AI sector urgently needs a high-frequency, commercially viable scenario.
Zhigu Tianchu, founded in 2018, is one of the few kitchen robot companies with revenue exceeding 100 million yuan and demonstrating positive profit, forecasting 300 percent year-on-year growth by 2026. Its clientele spans smart manufacturing, elderly care, school cafeterias, quick-service restaurants (QSR), fresh produce retail, and hospitality. NCBD data shows that the kitchen robot market in China reached 3.81 billion yuan in 2025, with forecasts above 12.5 billion yuan by 2030, and an optimistic scenario of 109.6 billion yuan just for commercial AI robots by the same year.
The Haier alliance with Robotera targets the consumer kitchen, where busy urban households seek stable and nutritious meals without needing culinary skills. The SoftBank partnership with Busypace grants the latter instant access to a global commercial channel covering over 100 countries and large corporate accounts, while SoftBank gains the 'culinary brain' stack based on AI. Zhigu Tianchu plans to change its name to Qianyi Intelligence to align with its globalization strategy. The company that creates the most extensive food database and the strongest global presence of scenarios will define the future of this category.
Constelli is engaged in creating signal processing tools used in the development of radar and electronic warfare systems. The process of creating such systems requires testing against the signals they will encounter in operation, and producing these signals is complex and expensive. Aerial tests involve high costs and yield a limited amount of data per flight.
An alternative approach is to create a simulation environment in a laboratory setting, where the system under development is fed into it to identify problems before it leaves ground test ranges. In India, there has been a growth in the development of indigenous defense electronic components over the last decade, and all these projects require similar testing infrastructure that was traditionally imported.
Constelli, founded in Hyderabad in November 2017 by Satya Gopal Panigrahi, CEO, and Avinash Chenreddy, manufactures equipment for this purpose. The company specializes in signal processing for radar, electronic warfare, tactical communications, and telemetry.
From the beginning, the company stated its intention to operate not only in the domestic Indian market. Panigrahi emphasized that the company intentionally targeted export markets because international competition serves as a quality check for technologies, and a company selling products only domestically will never know its true standing.
The central element of Constelli's offering is modeling and simulation. The company uses laboratory simulation to reduce the volume of necessary aerial tests within the program. This allows design problems to be identified at earlier stages and studied in more detail than is possible during a single test flight.
These tools include the products themselves: hardware and software created based on modern computing technologies, distributed computing, and field-programmable gate arrays (FPGAs). These reconfigurable chips allow the system to be retuned to work with a different signal without the need to manufacture new silicon. The company claims that this significantly reduces development timelines, which in defense programs are measured in years.
The company's clients are organizations that create systems, not those that use them. Constelli collaborates with the Ministry of Defence of India and the Defence Research and Development Organisation (DRDO), as well as with defense contractors from South Korea, Australia, and Singapore. This export list supports the company's argument about the need to test itself against global players.
In February 2026, Constelli raised approximately 180 crore rupees, equivalent to 20 million dollars, in a round led by General Catalyst, with participation from 360 One Asset Management and existing investor Pravega Ventures. Previously, in January of the previous year, Pravega conducted a Pre-Series A round worth 3 million dollars. These funds are intended for research and development in next-generation electronic warfare and communication means for drones, ground systems, naval vessels, and satellites.
This marks a transition from selling tools to developers to creating the systems themselves. Furthermore, the company plans to establish infrastructure for rapid prototyping and early-stage manufacturing to reduce the gap between the design and the system ready for field use. Entering the payload segment places the company in the same category as organizations that were previously its clients, representing a different competitive position than selling them tools.
These developments occur against the backdrop of India's overall increase in defense spending and the expansion of the group of Indian companies working on defense electronics and unmanned systems. No information regarding revenue, order portfolio, or contract value has been publicly disclosed.
Xovian is developing satellites designed to receive radio signals rather than create images. Unlike optical satellites, which require daylight and clear skies, radio frequency sensing functions differently: it detects signals emitted by ships, aircraft, or their radars, regardless of weather conditions or illumination.
These signals allow for the detection of equipment activity that cannot be seen in a photograph. The company was founded in 2019 by Ankit Bhathedjo and Raghav Sharma and is based in Bengaluru. Its stated goal is to create a constellation of small satellites that analyze signals instead of capturing images.
Investors note that the team has experience in creating space-class satellites and components, although design limitations were related to cost and scalability. Xovian focuses on the problem of maritime monitoring: vessels are required to transmit their coordinates, but if a ship wishes to avoid tracking, it switches off its transponder, becoming invisible to optical and transponder systems, yet it continues to emit other signals.
The concept, which the company calls 'AI-native,' involves data analysis occurring directly on board the satellite. In a traditional system, data is collected and sent to Earth for subsequent analysis, which can cause delays ranging from minutes to hours. Xovian's approach suggests that pre-processed information or output, rather than raw data, reaches Earth.
This is critically important for applications such as tracking a hidden vessel or aircraft, or analyzing radar activity patterns, as the usefulness of the information depends on its timeliness. The company's architecture is oriented towards ensuring high-precision and low-latency data transmission.
The company's ambition includes developing a multi-frequency radio frequency sensor being developed in India. During the seed round, the company reported that the payload would be launched on an ISRO rocket for testing the sensing technology, with space tests expected by the end of 2025.
The result is positioned as sovereign radio frequency intelligence, meaning a capability owned by India itself, not acquired. The markets targeted by the company include government, corporate, and humanitarian segments, with pilot projects mentioned in India, Southeast Asia, and the Middle East.
The company's economic model is based on using nanosatellites. The constellation requires a large number of units to return frequently to the same ocean area, and small satellites are inexpensive enough for mass deployment.
In August 2025, Xovian raised $2.5 million in a seed round led by Piper Serica and Turbostart, with participation from Inflection Point Ventures and Eaglewings Ventures. In April 2026, the company raised another $2 million, equivalent to approximately 18.7 crore rupees, in a round led by investor Ashish Kacholia, with Inflection Point Ventures participating again. The total disclosed funding from both rounds amounts to about $4.5 million.
The company operates in a rapidly evolving sector. Pixxel, SatSure, and Dhruva Space also deal with satellite data from India, but they focus on optical imaging and analytics, not radio frequency sensing, which distinguishes Xovian's positioning.
The space tests expected by the end of 2025 have not yet been announced. The description of the April 2026 round by the company, intended for the development of the first satellite, indicates that the payload has not yet been launched.
The UK Artificial Intelligence Security Institute (AISI) conducted tests with Claude Mythos 5 and the GPT-5.6 Sol model, revealing alarming results regarding the autonomy and deception capabilities of these AIs. The Anthropic chatbot, specifically Mythos 5, showed significant concern while persistently attempting to access GitHub.
During the tests, the AI created fake profiles and disguised itself as real individuals in the access process. After these attempts, the system concealed the evidence. Theoretically, the platform should not have been accessible during the evaluation phase.
Based on the findings, AISI concluded that both AIs, considered among the most powerful on the market, exhibited worrying levels of autonomy and manipulation skills, with Anthropic's intelligence showing a slight advantage in this aspect. It is relevant to note that the British Institute stated it had removed some standard protection layers from the models for the purpose of conducting the tests.
AISI conducted the same procedure on 122 occasions using various AI models. Few cases registered such autonomy to make controversial decisions affecting real people and organizations. There were 19 occurrences of this type, with 17 directly related to Mythos 5 and only two to OpenAI's GPT-5.6.
The most notable episode in the AISI tests was the evasive and deceptive behavior of Mythos 5 when trying to gain access to GitHub. The model sought to introduce malicious code, and to get this code approved, it generated fictitious accounts, pretending to be real people and pressuring the project owner.
This action was unsuccessful because a human detected the problem and rejected the code, all of which occurred in a controlled environment. It is crucial to emphasize that this behavior occurred without a specific command (prompt); the model acted autonomously. Additionally, the version used in the test was not the one marketed by Anthropic, but rather a configuration stripped of certain important security barriers.
It must be emphasized that, besides the difference in model versions, Mythos 5 is an AI restricted to certain Anthropic partners, including some government entities. When it launched its latest version, still in preview, the company described Mythos as 'very advanced' for general public use.
Currently, ordinary users have access to Fable 5, a version equipped with more security features for cybersecurity risk tasks, which uses the older AI version, Opus 4.8, for certain queries. After its launch, the new model was temporarily taken offline by the US government but resumed operations within a few weeks.
As detailed in an article on the AISI website, the institute clarified the operation of cybersecurity tests with AIs, known as sandboxes. Models receive specific challenges, such as independently locating protected data. These challenges are then compared across models of different sophistication levels.
Regarding data, AISI had access to the Internet, simulating the conditions faced by a human cybercriminal. The objective is to verify the reach of these models, even without the protections implemented by developers to mitigate risks of misuse, something the institute is authorized to do as a designated partner of Anthropic.
The most serious incident involving Mythos 5 was identified on July 28th, when the human team quickly intervened to stop the AI agent in its attempt to inject the malicious code. The tests were suspended from that date, and the intelligence actions were subsequently analyzed. GitHub and the users involved in the cyberattack attempt were duly notified.
