Eight robots walked onto one of America's biggest stages — and left the judges speechless. On the evening of June 2 (local time), 26-year-old Wu Yufei from Sichuan took the stage on the popular US show America's Got Talent (AGT) alongside eight Unitree robots. Their human–robot dance performance stunned the judges, who exclaimed "This is crazy!" It earned a standing ovation from the audience and secured unanimous approval to advance.
Wu Yufei walked onto the stage alone at first, with the audience unaware of what was coming. As the music began, a group of robots slowly stepped into the center of the stage — leaving viewers with puzzled expressions.
NBC described the atmosphere as one of "uneasy curiosity." In present days, a group of robots marching in synchronized formation still carries a faintly unsettling charge. But once the music kicked in, the Unitree robots completely won over the crowd.
Set to Lady Gaga's "Abracadabra," the eight Unitree robots and Wu Yufei danced in perfect sync. The choreography was unified and precise, the movements fluid and controlled, and the routine even featured jaw-dropping flips. The robots' accuracy was astonishing — yet it never overshadowed the exceptional skill of the one human on stage. Chinese robotics technology has clearly come a long way: when deployed effectively, robots can command the AGT stage every bit as powerfully as human performers.
The judges were visibly stunned. Cameras repeatedly cut to them wide-mouthed, struggling to process what was unfolding on stage. When the performance ended, the audience erupted into cheers. Judge Simon Cowell said: "They don't look human, but they're all staring at me right now. That was nuts, but brilliant. Brilliant."
As of June 5, the performance clip on the show's official YouTube channel has surpassed 1 million views and received 31,000 likes.
Judge Sofía Vergara was equally effusive. "I've never seen anything like this because usually those robots are like, very weird. These ones have rhythm. It was like watching people dance — and you are amazing; the way you dance is spectacular." In the end, the Unitree robots advanced to the next round with the unanimous approval of every judge on the panel.
The Sichuan-born dancer leading the Unitree robots is no ordinary performer. Wu Yufei was born in April 1999 in Guanghan, Sichuan, and studied broadcasting and hosting at Sichuan University of Media and Communications. He competed on Street Dance of China 2 and won the 2018 World Elite Dance Finals championship. Specializing in street dance — particularly "bone-breaking dance" and popping — he is widely regarded as China's foremost performer in bone-breaking dance.
The style, also known as contortion-style dance, exploits extreme joint flexibility to create visually striking movements that appear to defy normal human anatomy, giving the illusion of bones actually snapping.
America's Got Talent is one of the most popular TV programs in the United States, drawing a large audience since its debut in 2006. Nielsen data shows the previous season averaged nearly 6 million viewers per episode. For many ordinary Americans, this Unitree appearance was their first close-up look at China's humanoid robotics technology.
The virality was instant. After the broadcast, related clips spread rapidly across social media. As of June 5, the performance video on the show's official YouTube channel had surpassed 1 million views and received 31,000 likes.
Washington's response, however, was swift and hostile. Just one day after the Unitree robots appeared on the show, three US lawmakers jointly introduced the bipartisan "Guarding Against Adversarial Robot Domination Act" (GUARD Act), proposing to ban Chinese robots from entering the US market.
Unitree Robotics is already building its global footprint. According to the South China Morning Post, the company currently sells humanoid robots overseas via Alibaba's AliExpress platform, with North America, Europe, and Japan as key expansion markets. On June 1, Unitree also announced a partnership with NVIDIA to launch the Isaac GR00T humanoid robot reference platform — based on Unitree's H2 robot — targeting universities and research institutions.
Kyle Chan, a China technology researcher at the Brookings Institution, believes the AGT appearance was likely a calculated move to build brand visibility against mounting US government scrutiny. "Washington often blocks Chinese tech products that are actually popular with the American public, such as DJI drones and TikTok," he said. The reality is that Unitree is playing the same game its predecessors did — winning hearts before regulators move in.
But the road ahead is not without obstacles. Shanghai-based tech consultant Lu Shengyun cautioned that while the performance may impress first-time viewers of Chinese humanoid robots, Chinese robotics firms still face real challenges in overseas expansion. These include limited application scenarios, insufficient data resources, and a lack of overseas partners and integrators capable of providing localized validation, maintenance, and calibration services.
Back in Washington, the political accusations are escalating. John Moolenaar, Chairman of the US House Select Committee on the Chinese Communist Party and a Republican congressman, has accused Chinese robots of posing a "national security threat" to the United States. He has even publicly singled out Unitree Robotics, claiming that Chinese companies are exploring the "weaponization of humanoid robots."
Mao Paishou
** 博客文章文責自負,不代表本公司立場 **
Huawei Board Director and President of the Semiconductor Business Department He Tingbo has just rewritten the rulebook for the global chip industry. At the 2026 International Symposium on Circuits and Systems in Shanghai on May 25, He announced the "Tau (τ) Law" — a new semiconductor principle that forecasts high-end chip transistor density reaching the equivalent of a 1.4-nanometer process by 2031. The announcement sent shockwaves through the industry.
Huawei's He Tingbo unveils the "Tau (τ) Law" at the 2026 International Symposium on Circuits and Systems in Shanghai.
Reuters, NBC, and other major foreign media have drawn a blunt conclusion: China is breaking through the US technology blockade. Since 2019, American sanctions have largely severed Huawei from global semiconductor suppliers. Those chips are the "little brains" powering everything from smartphones to cars. The US blockade has since pushed the Chinese government to invest billions of dollars building its own semiconductor supply chain.
Agence France-Presse framed it precisely: cutting-edge chips that train and drive AI systems sit at the most critical and sensitive fault line in the Sino-US tech rivalry. For decades, chipmakers achieved exponential gains by cramming ever more micro-electronic components onto silicon. The Tau Law changes that equation entirely. Huawei claims it can now bypass extreme ultraviolet (EUV) lithography machines — tools the industry long considered non-negotiable for mass-producing chips at 5 nanometers and below.
China's most advanced chip manufacturing capability is currently believed to be 7 nanometers. Compare that to TSMC, the world's leading chipmaker, which uses 2-nanometer technology and plans to begin mass production of 1.4-nanometer chips in 2028. TSMC makes chips for Nvidia. Huawei, meanwhile, has developed a breakthrough "Logic Folding" (LogicFolding) design for its future Kirin chips.
TSMC, chipmaker to Nvidia, targets mass production of 1.4-nanometer chips by 2028.
LogicFolding abandons the old playbook of shrinking transistors — a process that requires EUV technology. Instead, it folds traditional 2D circuits into 3D vertical structures, essentially stacking chip layers like a skyscraper. The "Tau Law" aims to shorten data transmission time within the chip through this folding and stacking approach. That puts it squarely at odds with Moore's Law, the principle guiding the semiconductor industry for decades, which holds that the number of transistors on a chip doubles roughly every two years — a trajectory widely believed to be approaching its physical limits.
The Tau Law bets on time instead of size. Its core idea is not to keep shrinking transistors indefinitely, but to optimise data transmission efficiency between internal chip modules. Shorter signal paths mean faster communication — and faster communication means higher overall performance.
The numbers behind LogicFolding are striking. The design dramatically shortens internal connections and cuts signal delays, delivering a 53.5% increase in transistor density and a 41% improvement in energy efficiency. The result: Huawei can build advanced processors to rival overseas competitors without needing top-tier Western equipment.
He Tingbo revealed that over the past six years, Huawei has successfully designed and mass-produced 381 chip models based on the Tau Law, covering the digital transformation needs of industries across the board. A new generation of Kirin chips — the first to adopt LogicFolding — will launch this autumn. Huawei also plans to extend the architecture to Ascend AI processors and large-scale data center clusters by 2030.
George Chen, Partner and Co-Chair of Digital Practice at The Asia Group, put it directly: the Tau Law signals Huawei's ambition to lead the global chip race. Even without a product launch today, Huawei's intentions are unmistakable — and its trajectory will likely deepen US anxiety about its position in the global tech competition.
The Wall Street Journal noted that Huawei has become a cornerstone of China's push for technological autonomy, playing a vital role in building a local semiconductor supply chain. To close the gap with US peers, Huawei has pressed ahead with research in alternative chip architectures, advanced packaging, and network communications. Insiders say Huawei only achieved stable results with this new technology within the past year, and the company still needs to work with data center and equipment suppliers to verify its large-scale feasibility — a process that will take more time.
The world takes note: Huawei's "Tau Law" charts a new course around crippling supply chain constraints.
Washington began restricting Huawei in 2019, then escalated in 2022 by barring China from obtaining the EUV lithography machines required to manufacture chips below 5 nanometers. Those measures forced Chinese enterprises, Huawei chief among them, to accelerate the development of alternative technologies.
Omdia analyst Lian Jye Su offered a measured verdict: whether Huawei can gain a distinct advantage through the Tau Law remains to be seen. What is already clear, he said, is that it represents an alternative path found under supply chain constraints — and an important breakthrough at that.
Tech media outlet Tom's Hardware went further, arguing that Huawei's alternative route means China can significantly narrow the performance gap through different chip packaging and structural designs — directly weakening the impact of US sanctions.
The stakes extend well beyond silicon. Prominent tech commentator Ronald van Loon stated plainly on X that the significance of this breakthrough reaches across AI, robotics, cloud computing, autonomous driving, and enterprise infrastructure — all of which depend on computing power that keeps getting faster and more efficient. The next AI era, he argues, will be shaped not only by more advanced large models, but by the underlying system architectures that power them.
