{"text":[[{"start":10.08,"text":"Robots are not new to investors. Instead, what’s new is the scale of checks being written for companies pairing fast-improving AI agents with machines that can perform practical tasks in the real world. That helps explain why embodied-AI startup PsiBot, also known as Lingchu Intelligence, was able to raise a hefty 2 billion yuan ($280 million) just two years after its founding, in its angel and Pre-A funding rounds."}],[{"start":44.379999999999995,"text":"The company didn’t disclosed a valuation in its announcement of the funding last week. But the headline number stands out for its size for such a young company, in a sector that generated just $440 million in global humanoid-robot sales in 2025, according to IDC."}],[{"start":65.22999999999999,"text":"Just as important as the amount is the timing. Symbolically, PsiBot’s March 10 announcement came during China’s annual “Two Sessions,” the highpoint of its annual political season where the government discusses priorities for the year ahead. This year’s meeting was especially significant as it featured the unveiling of China’s 15th Five Year Plan, which put fresh emphasis on AI, robotics, and their industrial deployment. In that context, PsiBot’s funding looks like more than a simple vote of confidence in a single startup. It fits nicely into a broader push by policymakers and investors to move AI beyond the drawing board and into the real economy."}],[{"start":111.07,"text":"That shift is a key to what PsiBot is selling. Rather than joining the crowded race to build ever more theatrical embodied models that can dance and do kung fu, it says the new funding will support the nuts-and-bolts behind what makes such robots work, including larger-scale logistics deployments and a large-scale data collection system."}],[{"start":134.60999999999999,"text":"PsiBot is positioning itself neither as pure software nor as a fully vertically integrated hardware maker. Instead, it describes an approach of what it calls providing a “small full-stack.” It controls key elements of hardware design, such as structure, motion range, and degrees of freedom, while outsourcing production of core components and manufacturing to specialist suppliers. The division of labor is clear: PsiBot retains tighter control over system creation and fine-tuning without paying the full cost of designing and building everything in-house."}],[{"start":172.75,"text":"In practice, the hardest problem facing makers of a future generation of practical robots may not be building impressive machines. Instead, it may be gathering enough high-quality data to make those machines reliable in messy, real-world settings. Large language models have the internet as a vast bank of information to get the job done."}],[{"start":196,"text":"But robots have no such resource. They learn through simulation, teleoperation, and field use, making the data they collect a strategic asset. Well-funded competitors are already blending simulation, human-action video, and feedback from deployed robots to build larger training loops. PsiBot’s system appears aimed at obtaining the same capabilities by creating scalable, repeatable learning pipelines."}],[{"start":225.49,"text":"Practical deployment"}],[{"start":228.62,"text":"This emphasis on practical deployment is a key differentiator that sets PsiBot apart from some of its more headline-grabbing embodied robotic startup rivals. Embodied robots have captured the spotlight lately, especially after their splashy appearance on this year’s “Spring Festival Gala” program during the Lunar New Year. But excitement and commercial reality rarely move in lockstep. Tesla’s Optimus remains the most recognizable symbol of the race, yet meaningful production scale still isn’t expected until later this year."}],[{"start":265.71000000000004,"text":"PsiBot, by contrast, seems to be betting that less glamorous work, especially in logistics and manipulation, offers a clearer path to early revenue. And it’s doubling down by betting not only on its own embodied hardware, but also on providing general-purpose embodied intelligence, large-scale vision-language-action (VLA) models, and dexterous manipulation algorithms for other robot developers as well."}],[{"start":295.91,"text":"Its investor roster reinforces that view. The angel round drew state-backed and industrial capital, including Guokai Finance, Guozhong Capital, and a CCTV-affiliated industry fund. The Pre-A round brought in Shanghai state-backed Xuhui Capital, local state funds, market funds, and larger commitments from existing backers. For a young company, that is a serious endorsement."}],[{"start":322.66,"text":"The founding team also has strong credentials. Founder Wang Qibin brings more than two decades of hardware and commercialization experience and previously led JD.com’s robotics business. The company also points to a joint lab it developed with Peking University focused on embodied dexterous manipulation, adding an academic layer to its commercial story."}],[{"start":347.29,"text":"The broader market backdrop is also becoming more favorable. IDC says global humanoid-robot shipments reached about 18,000 units in 2025, up 508% year over year, with Chinese manufacturers taking an early lead in scale delivery. That suggests the industry is beginning to find commercial buyers outside the lab, helping explain why investors are suddenly willing to throw big bucks at the sector. Overseas, companies such as Figure AI and Skild AI have also raised large funds around variations of the same humanoid and robot-intelligence thesis. PsiBot’s new financing fits squarely into that wider rush of capital."}],[{"start":391.62,"text":"Even so, the sector is still in an early phase. Much of today’s demand remains concentrated in entertainment performances, education and research, and in data collection rather than broad industrial deployment. At the same time, the business model is starting to edge beyond one-off hardware sales toward offerings that provide recurring revenues for services, maintenance and platform-style offerings. That matters for PsiBot because it’s betting that data and training infrastructure could become a recurring source of value – providing steady revenue as an alternate to one-time sales."}],[{"start":431.46000000000004,"text":"Still, big fundraising and a sustainable business model are not the same thing. PsiBot says it has moved beyond the lab, citing deployments of its products in semi-structured logistics settings such as clothing warehouses, where robots handle tasks like barcode scanning and sorting across large numbers of product types. These repetitive-but-variable jobs are the use cases investors tend to watch most closely because they offer a clearer route to commercialization than staged demos. Even so, the company still needs to prove that early pilots can turn into mass deployments that bring meaningful revenue."}],[{"start":470.1,"text":"Looking ahead, the bigger question is where the enduring value in embodied AI will settle. The obvious winners may be companies that build the actual bodies. But an equally important group could be third-party providers of the data, training, and deployment systems needed to train robots and help them keep improving over time."}],[{"start":492.1,"text":"That appears to be the layer that PsiBot is targeting. If it can turn deployments into repeat business, its large fundraising at such an early stage of its development may look like a smart early bet on where embodied AI’s real leverage sits. If it can’t, its case will serve as another reminder that the development of physical AI is a tricky business where investors must guess – often unsuccessfully – where the market is going."}],[{"start":529.85,"text":""}]],"url":"https://audio.ftcn.net.cn/album/a_1773822562_7499.mp3"}