China stages first-ever humanoid robot kickboxing match
In a clash of metal, might and artificial intelligence, Unitree Robotics' humanoid robots threw punches and kicks in the world's first kickboxing fight of its kind in China.
The China Media Group (CMG) World Robot Competition – Mecha Fighting Series, the first-ever humanoid robot fighting tournament, kicked off in Hangzhou, Zhejiang province, on May 25.
State-owned Chinese Central Television broadcast the match, which included four Unitree G1 robots and their human trainers. Each fighter robot weighs about 35 kilograms and is 132 centimeters tall.
Each match had two sessions. The first required the robots to perform different actions, while the second involved three rounds of combat, each lasting two minutes.
A robot could score only if it hits its opponent's head or body. A punch scored one point, while a kick scored three. The fighting machines lost five points if they fell. If it couldn't get up within eight seconds, it lost 10 points, and the round ended immediately.
Fight footage showed the robots could punch and kick smoothly, but their attacks were soft and more like pushes than strikes. Some robots lost their balance after kicking or moving backward, but most could stand back up after a few seconds. Trainers controlled their robots remotely with joysticks.
'It is not easy to teach robots different movements,' Wang Qixin, a director at Unitree, told the CCTV. 'We used artificial intelligence (AI) technology to train them.'
'First, we captured the data of the movements of some professional kick-boxing athletes, and then the robots can learn these movements in the virtual world,' he said.
Li Gaofeng, a researcher at Zhejiang University's College of Control Science and Engineering, told the National Business Daily that the match demonstrated a leap in China's humanoid robotic technologies in the past six to 12 months.
'Six months ago, people could not imagine that humanoid robots would be able to perform all these movements smoothly,' Li said.
'Combat fight is a difficult task for humanoid robots due to the intensive confrontation during the fight. Robots need to mind their movements and react to their opponent's moves,' he said. 'All these requirements significantly challenge the robots' algorithms, electronic parts and speed reducers.'
A speed reducer is a gear between the motor and the machinery that reduces the rate at which power is transmitted. Better speed reducers can achieve faster and more precise movements.
Industry experts told the Global Times that the kickboxing fight created a 'train-through-competition' talent pipeline for China's fast-growing robotics sector and will drive innovation in perception, control and execution technology.
Tian Feng, former dean of Chinese AI software firm SenseTime's Intelligence Industry Research Institute, said combat sports demand full-body coordination, upper limb capabilities, battery endurance and material durability.
EngineAI, a Shenzhen-based robotics company, will host a large-scale combat match for full-sized humanoid robots in December in Shenzhen, Guangdong province. The event aims to showcase cutting-edge AI and robotics technologies while promoting industrial upgrades and adoption.
From dancing to fighting
Only four months ago, Unitree's H1 humanoid robots surprised the audience by dancing with people at the 2025 CCTV Spring Festival Gala on January 28. Then, on March 19, Unitree's G1 robots performed side-flips and kick-ups in a video, showing a significant improvement from its H1's backflip performed a year ago.
Some Chinese commentators said Unitree had surpassed America's Boston Dynamics, as the latter's Atlas could only perform cartwheels. However, others said Boston Dynamics' robots are more advanced because of their precise and steady movements.
On April 10, Unitree released footage showing a kickboxing fight between G1 and a male boxer and another between two G1 robots. The human boxer could easily knock down the robot, which was only half his weight.
The company requested its users to refrain from making dangerous modifications or using the robot in a hazardous manner.
A columnist at Mydrivers.com, a Chinese consumer technology website, said G1's reaction was slower than humans and that its attacks were not precise.
Meanwhile, a security camera video showed that a Unitree H1 had tried to attack the developers during a test at a factory in China. The robot was initially sitting dormant, but suddenly began flailing its limbs violently.
Attached to a crane, the 1.8-meter-tall robot did not hit anyone before an engineer shut it down.
The video went viral on the Internet as some grow concerned about the safety of humanoid robots, especially when Chinese robot makers plan to deploy their products to people's homes within the next three to five years.
Read: Chinese humanoid robots get reality check in half-marathon debut
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Asia Times
11 hours ago
- Asia Times
Marco Rubio's and Miles Yu's war on Chinese students is misguided
In an age of escalating geopolitical rivalry, democracy's strongest foundations — press freedom, civic trust and public accountability — are being eroded by a perfect storm of surveillance, suspicion, and systemic misinformation. This is especially visible in US-China relations, where bipartisan hawkishness has led to sweeping proposals like Senator Marco Rubio's latest effort to revoke visas from Chinese students and researchers — treating them as national security risks by default. Joining the chorus is Miles Yu, a former Chinese international student who became a top China policy adviser in the first Trump administration. In his widely cited essay, 'Enabling the Dragon,' published in November 2024 the week after Donald Trump had won the election, Yu argues that US universities have become naive enablers of the Chinese Communist Party, serving as academic outposts vulnerable to intellectual theft and ideological infiltration. Yu urges that the United States should sharply restrict academic engagement with China, calling such cooperation a national security threat. His claim is sweeping: that China has 'outsourced' its academic system to exploit American openness, and that the US must respond by severing intellectual ties. Both Rubio and Yu are also ignoring the data: Chinese nationals make up the largest share of foreign students in STEM fields — computer science, engineering, math and the physical sciences. According to the National Science Foundation, more than 80% of Chinese PhD recipients in these fields stay and work in the US after graduation, contributing directly to American innovation, entrepreneurship, and research leadership. Many have founded startups, filed patents and worked in cutting-edge labs at US universities and tech companies. The idea that they are 'outsourcing' American prosperity to China is not only false — it's self-destructive. If these students are forced out, the US will not only lose a competitive advantage in global talent — it will damage its innovation ecosystem at its roots. Immigration-driven innovation has been one of the few consistent engines of American prosperity in a polarized and gridlocked political climate. Treating every foreign-born talent as a potential spy will only drive them into the arms of competitors. Moreover, this zero-sum framing misrepresents how education actually works. American universities are not ideological weaklings — they are spaces where critical thinking, civic inquiry and pluralistic values are cultivated. Chinese students are not arriving with monolithic loyalties — they are shaped by their experiences here, often becoming some of the most perceptive critics of authoritarianism and some of the strongest defenders of democratic ideals. Diaspora students and scholars, such as the founders of China Labor Watch and Human Rights in China, have often been at the forefront of documenting abuses, challenging both Chinese state narratives and the overreach of US suspicion. They are not security liabilities — they are civic actors. And yet, they are increasingly caught in the middle. Media outlets rush to publish stories about alleged espionage long before there's due process. Federal task forces pressure universities to cut off collaborations without context. On social media, platforms like X — once Twitter — amplify xenophobic paranoia while silencing legitimate voices. The result is a digital public sphere poisoned by fear and disinformation, where nuance disappears and policy becomes a blunt instrument of exclusion. In my research — China's Emerging Inter-network Society — I explore how diaspora communities and digital platforms are reshaping political consciousness. Platforms like WeChat and TikTok are indeed double-edged: they can be used for surveillance, but also for storytelling, mutual aid, and grassroots advocacy. What Yu fails to mention is this: He was once 'the dragon' he now seeks to shut out. To presume otherwise is to vastly underestimate the power of American education — something Yu himself should know firsthand. Yet there's a glaring irony: Yu himself is living proof that American education works — not just as a system of knowledge transmission, but as a transformative force of values, perspective and civic engagement. Yu came to the US in the 1980s as an international student from China. He benefited from the very system he now decries — one that welcomed global talent, nurtured individual potential and allowed a Chinese-born scholar to rise to the highest levels of US policymaking. If America had treated him then the way he now proposes treating others, Miles Yu might still be teaching Maoist doctrine in Anhui, not advising presidents in Washington. If Miles Yu truly believed Chinese students couldn't be trusted, one wonders why he chose to stay and serve in the US government rather than return to China after pursuing his PhD degree. Doesn't his own life prove the power of American education to transform, inspire, and integrate? If we now assume every Chinese student is a CCP foot soldier, does that include him too? Or is he the exception who proves the value — not the danger — of keeping the door open? He chose to stay in the United States not because he was coerced but because the openness and meritocracy of American institutions resonated with him. If we now claim that every Chinese student is a sleeper agent for Beijing, then Yu's own journey becomes an inconvenient contradiction. Isn't he the evidence that America's democratic model can win hearts and minds? That contradiction isn't just ironic. It's emblematic of a dangerous drift in US national security thinking in which suspicion has replaced strategy and identity has replaced evidence. If the US blocks Chinese students while maintaining that it wants to 'compete' with China, Beijing will likely frame the move as hypocritical — claiming it reveals American insecurity rather than confidence in its democratic model. The retaliatory measures may not just hurt bilateral relations but also signal to other countries the risks of aligning too closely with US policy on China. Yu's central claim is that Chinese students and scholars serve as covert extensions of the Chinese Communist Party (CCP), sent not to learn but to spy, steal, and subvert. This argument has gained traction in parts of Washington, where fears of intellectual property theft and technological competition are real and justified. But let's be clear: there is a vast difference between targeted counterintelligence and collective suspicion. To reduce an entire population of students — numbering over 270,000 annually — to latent threats is both empirically unfounded and strategically foolish. Chinese students are not a monolith. Many come precisely because they seek an alternative to the CCP's control. Some become critics of the regime. Others stay, contribute to US innovation, or build bridges that serve American interests abroad. Treating them as presumed agents of espionage doesn't protect US security — it undercuts America's greatest soft power asset: its openness. We are now witnessing the consequences of this worldview hardening into law. In May 2025, Secretary of State Marco Rubio, citing security risks, announced that his department would move to revoke or block Chinese student visas in 'sensitive' research fields outright, citing national security risks. The proposal would give broad authority to federal agencies to deny or cancel visas without due process, based not on individual conduct, but on nationality and field of study. This is not strategic caution — it's blanket exclusion. And it mirrors the logic of Yu's essay: that anyone Chinese by origin or association is inherently suspect. Such policies are dangerously close to the racialized fearmongering of the Chinese Exclusion Act era, now dressed in tech-sector clothing. They undermine US universities, punish innocent scholars, and hand the CCP a propaganda victory. If carried out, this policy won't stop espionage — it will cripple American research labs, isolate Chinese dissidents, and accelerate talent flight to competitor nations like Canada, the UK, and Australia. The Trump administration's aggressive stance on Chinese espionage is haunted by the very intelligence failures it now seeks to prevent. As Sue Miller, the CIA's former chief mole hunter, has pointed out, the collapse of US spy networks in China more than a decade ago — a debacle that saw scores of informants arrested or executed — remains unresolved. That strategic humiliation not only decimated on-the-ground intelligence, it also created a culture of institutional paranoia in Washington. Now, instead of rebuilding trust and refining intelligence practices, the Trump-era approach has leaned heavily on suspicion and overreach — particularly targeting ethnic Chinese scientists, scholars, and students. But blunt tools don't fix complex failures. The overcorrection has led to high-profile wrongful prosecutions, deteriorating academic collaboration and growing mistrust within diaspora communities. The United States' inability to root out past internal breaches has fueled a form of policy scapegoating — one that risks trading precision for profiling. Without credible reform of intelligence capabilities and transparent accountability for past missteps, the crackdown will remain reactive, politically charged and ultimately self-defeating. Yu frames UA-China academic collaboration as 'outsourcing,' suggesting the US has ceded control of its intellectual infrastructure to a hostile power. But this misunderstands both how American academia works and why it thrives. Academic exchange is not a one-way transaction. It's a competitive ecosystem, where ideas are tested, refined and challenged through global participation. Chinese students and researchers don't dilute US education — they elevate it. They help fill STEM classrooms, contribute to breakthroughs in AI and biomedical research, and keep US universities globally dominant. Cutting them off would hurt America far more than it would hurt China. Yes, vigilance is necessary. Research security protocols should be strong. Federal funding should come with guardrails. But throwing out the entire system of engagement, as Yu and now Rubio suggest, would be self-sabotage. If enforced, Rubio's proposal to ban Chinese students will not only undercut America's higher education system — it could also trigger swift retaliation from Beijing. China may impose reciprocal visa restrictions on US students, scholars and education programs, halt joint research initiatives or tighten controls on American academic access to Chinese data and field sites. More strategically, it could restrict elite talent from going to the US, incentivize a reverse brain drain or escalate a global narrative campaign accusing the US of racial discrimination. Such moves wouldn't just harm bilateral ties — they would damage America's soft power, alienate diaspora communities and send a troubling signal to other nations about the risks of engaging with US institutions. Ironically, by closing the door on Chinese students, Rubio and his allies may be doing more to weaken America's global leadership than to defend The U.S.-China contest is not just about chips, jets, and rare earths. It's about the future of global norms — openness versus control, pluralism versus authoritarianism. In this battle, academic freedom is not a vulnerability. It's a weapon. It is what makes the US different from — and stronger than — the system the CCP promotes. If we start mimicking Beijing's paranoia, walling off knowledge, and excluding people based on their passport, we risk becoming what we claim to oppose. Yu himself is living proof of that freedom's power. He came to the US seeking truth, found it in an open society and used it to shape national strategy. That's a success story, not a turn around now and advocate for closing the gates behind him is not only short-sighted — it's a betrayal of the very ideals that made his own story possible. A call for strategic openness Miles Yu transferred himself from Chinese student to gatekeeper by pulling up the ladder behind him. What we need is not blanket restriction but smart engagement, clearer funding rules, targeted export controls and honest dialogue with university leaders – and, yes, a robust national security posture. But we must resist fear-driven policies that punish potential allies and weaken our intellectual base. The best way to 'outcompete' China is not to become more like it — but to double down on what made the US the envy of the world. If we follow Yu's and Rubio's advice, we may win a battle of suspicion — but lose the war for global leadership. If the US wants to outcompete authoritarian regimes, it must stop mimicking their logic. Surveillance, guilt by association and ideological profiling are not strategies for innovation — they are symptoms of decline. Democracy's strength lies in openness, in attracting talent, and in offering a system that can inspire — not coerce — loyalty. Rather than banning students, the US should reinvest in the institutions that make it a magnet for global minds: its universities, its press, and its civic infrastructure. Journalists must be more careful not to amplify racialized suspicion. Lawmakers must recognize that brainpower, not fear, drives prosperity. Scholars like Miles Yu must reckon with the contradiction between their personal journeys and the policies they now advocate. Democracy does not win by closing its doors. It wins by proving it is worth entering. Yujing Shentu, PhD, is an independent scholar and writer on digital politics, international political economy and US-China strategic competition.
RTHK
3 days ago
- RTHK
Chip design firm Synopsys puts stop to Chinese sales
Chip design firm Synopsys puts stop to Chinese sales Synopsys says it is blocking sales and fulfillment in China and halting new orders until it receives further clarification. File photo: AFP American semiconductor design software firm Synopsys has told staff in China to halt services and sales in the country and stop taking new orders to comply with new US export restrictions, according to an internal letter. The United States has ordered a broad swathe of companies to stop shipping goods to China without a license and revoked licenses already granted to certain suppliers, people familiar with the matter have said. Products affected include design software and chemicals for semiconductors, they said. Synopsys on Thursday suspended its annual and quarterly forecasts after it received a letter from the Bureau of Industry and Security of the US Department of Commerce, informing it of new export restrictions related to China. The internal letter sent to staff in China on Friday said "based on our initial interpretation, these new restrictions broadly prohibit the sales of our products and services in China and are effective as of May 29." To ensure compliance, Synopsys said it was blocking sales and fulfillment in China and halting new orders until it receives further clarification. The measures affect all customers in China, including employees of global customers working at sites in China and Chinese military users wherever they are located, the letter added. The steps Synopsys is taking in light of the new restrictions have not been previously reported. Synopsys did not immediately reply to a request for comment. Alongside Cadence and Siemens EDA, Synopsys is among the top three companies that dominate electronic design automation (EDA) software that chipmakers can use to design semiconductors used in everything from smartphones to computers and cars. Restricting Chinese firms' access to EDA tools would be a big blow to the industry as Chinese chip design customers heavily rely on top-of-the-line US software. Synopsys, Cadence and Siemens' Mentor Graphics control more than 70 percent of China's EDA market, Xinhua reported in April. Chinese companies that have said they use Synopsys and Cadence software include design firm Brite Semiconductor, Zhuhai Jieli and semiconductor IP portfolio provider VeriSilicon. The letter sent to staff in China on Friday also said that Chinese customers' access to its customer support portal SolvNetPlus had been disabled. (Reuters)
RTHK
4 days ago
- RTHK
Xiaomi set to let the good times roll with YU7 launch
Xiaomi set to let the good times roll with YU7 launch Xiaomi's new electric sports utility vehicle YU7 draws interest at a Xiaomi store in Beijing. Photo: Reuters Xiaomi rolled out its new sports utility vehicle in Beijing on Thursday, as the firm best known for smartphones and consumer electronics gears up to further challenge Tesla in the world's largest auto market. Xiaomi launched the YU7 at 13 of its Beijing showrooms and will start taking orders for the vehicle in July. It is keen to repeat the success of the sporty SU7 sedan, which launched last year and has outsold Tesla's Model 3 on a monthly basis since December. Analysts have said the YU7 could pose a major threat to Tesla's best-selling Model Y but its launch comes at a time when Xiaomi, a relative newcomer to China's highly-competitive EV market, has seen new EV orders fall after a series of controversies. Chinese authorities have been investigating a fatal highway crash at the end of March involving an SU7 in driving-assistance mode and Xiaomi has apologised for "not clear enough" marketing after customer complaints of false advertising. Liu Jiaxing, a 34-year-old tech worker, was among the first visitors to Xiaomi's flagship showroom in Beijing Oriental Plaza on Thursday, eager to catch a glimpse of the emerald green YU7. Liu said he was fond of the styling and colour as well as the fact that Xiaomi vehicles connect with the firm's personal gadgets and smart home products, which he felt pointed to how local brands understood Chinese consumers better than their foreign counterparts. "I used to be more prone to US, German and French car marques, but the fast progress of China's EV sector prompts me to focus more on the products rather than brands," he said. Another visitor was Tom van Dillen, managing partner at German management consultancy Greenkern in Beijing, who said he was not a fan of some of the YU7's intelligent features, which he described as "unnecessary", but said the YU7 was a formidable challenger to the Model Y. He cited a "physical ecosystem advantage in the showroom where there is a dedicated area with accessories that only fit into Xiaomi cars" and their competitive price. Xiaomi has said that it will only announce the YU7's pricing in July. HSBC Qianhai estimated in a note last week that the new SUV will be priced between 230,000 yuan and 330,000 yuan and that Xiaomi could ship 100,000 YU7 units this year and 249,000 units in 2026. The Model Y is priced from 263,500 yuan in China. (Reuters)
