150亿。这是逐际动力（LimX Dynamics）最新一轮融资后的投后估值——单位是人民币。
7月14日，深圳通用人形机器人公司逐际动力宣布完成近2亿美元Pre-IPO轮融资。投资方阵容豪华：IDG资本领投，全球AI硬件智造龙头蓝思科技、以及多家美元基金和产业资本跟投。这不是逐际动力今年第一次融资——2月份它刚完成2亿美元B轮，半年内累计融资4亿美元。
更值得注意的信号是：逐际动力已确认进入港股IPO保密审核阶段。一家成立仅四年多的创业公司，在人形机器人还没有大规模量产、还没有大规模商业化收入的情况下，就要冲刺IPO，而且估值给到了150亿。资本的急迫感，比任何行业报告都更说明问题。
为什么资本在这个时间点疯狂卡位？
人形机器人赛道喊了好几年了，为什么2026年夏天突然成了资本的香饽饽？三个信号叠加在一起。
第一个信号是工业场景的真实落地。就在逐际动力融资消息出来的同一天，雷军在微博晒出了小米人形机器人在汽车工厂"实习"4个月的成绩单：自攻螺母装配成功率从最初的90.2%提升到98%，距离熟练工人99%的合格率仅差1个百分点，并且新增了中涂作业、质量检测等多个工位能力。
98%是什么概念？这意味着在特定工位上，机器人已经可以替代人工了。不是实验室里的demo，不是展会上的表演，是在真实的汽车生产线上，4个月的持续迭代，把成功率从"勉强能用"做到了"接近人工"。当工业场景的ROI算得过来了，资本的想象力就被彻底打开了。
第二个信号是供应链的成熟。2026年以前，人形机器人的核心零部件——高扭矩密度电机、精密减速器、力矩传感器——基本靠进口，一台机器人成本动辄几十万上百万。但今年以来，国产供应链快速成熟，蓝思科技这类消费电子供应链龙头的切入，意味着人形机器人有望复制消费电子的降本路径：从几十万降到十几万，再降到几万。
第三个信号是大模型给机器人"装大脑"。以前的人形机器人是"预编程动作回放"，只能做预设好的动作；现在有了端到端的具身智能大模型，机器人可以理解自然语言指令、可以在陌生环境里自主决策、可以通过试错学习新技能。这才是真正的"智能"机器人，不是遥控机器。
IPO窗口期抢跑：不上市就出局？
逐际动力在这个时间点冲刺IPO，多少有点"抢窗口期"的意味。
人形机器人赛道现在的局面是：宇树科技已经上市，优必选在港股表现不错，波士顿动力还在现代汽车体系里，特斯拉Optimus进展时快时慢，国内一堆创业公司（逐际动力、智元、傅利叶、星动纪元等）都在抢"人形机器人第一梯队"的位置。谁先上市，谁就能拿到最多的资金、最多的人才、最多的媒体关注，在接下来的量产竞赛中占据先机。
但风险也是明摆着的：人形机器人的大规模商业化，可能还需要2-3年时间。现在上市，资本市场买的是"预期"，不是"营收"。如果2-3年后量产不及预期，估值泡沫随时可能破裂。
"人形机器人正在重复电动车的故事：先画饼融资→再烧钱量产→最后洗牌活下来3-5家。但这个过程中，上游供应链先赚钱。"—— 一位硬科技投资人
不过对逐际动力这类公司来说，"不上市可能风险更大"。因为这个赛道的烧钱速度太快了——研发、供应链、工厂、人才，每一项都是巨额支出。如果不借助资本市场的力量，很可能在量产前夜弹尽粮绝，被竞争对手反超。
资本不是傻子。半年4亿美元、150亿估值、Pre-IPO轮——这些数字背后是聪明钱在押注一个判断：人形机器人的量产奇点，比大多数人想象的要近。
当小米机器人在产线上做到98%合格率的时候，你就知道：这个奇点，可能真的不远了。
明天见。
$2 billion. That's LimX Dynamics' post-money valuation after its latest round — in USD equivalent, roughly $2.1 billion.
On July 14, Shenzhen-based general-purpose humanoid robotics company LimX Dynamics announced nearly $200 million in Pre-IPO funding. The investor lineup is blue-chip: IDG Capital led the round, joined by Lens Technology (the global AI hardware manufacturing leader), multiple dollar funds, and strategic investors. This isn't LimX's first rodeo this year — it closed a $200M Series B back in February, bringing its six-month total to $400 million.
The more notable signal: LimX has confirmed it's entered confidential Hong Kong IPO review. A startup barely four years old, sprinting for a public listing before humanoid robots have reached mass production or meaningful commercial revenue — at a $2B+ valuation. The urgency in capital markets tells you more than any industry report.
Why Is Capital Frenzying Right Now?
The humanoid robot track has been hyped for years. Why summer 2026? Three signals converging.
First signal: real deployment in industrial scenarios. The same day LimX's funding news dropped, Lei Jun took to Weibo to share Xiaomi's humanoid robot's four-month "internship" report card at an auto factory: self-tapping nut assembly success rate climbed from an initial 90.2% to 98% — just one percentage point shy of the 99% standard for skilled human workers — with new capabilities added in painting, quality inspection, and more.
What does 98% mean? It means at specific workstations, robots can already replace humans. Not a lab demo, not a trade show performance — four months of continuous iteration on a real auto production line, taking success rates from "barely usable" to "near human." When the ROI math works for industrial scenarios, capital's imagination is fully unlocked.
Second signal: supply chain maturation. Before 2026, core humanoid components — high-torque-density motors, precision reducers, torque sensors — were mostly imported, with a single robot costing hundreds of thousands of dollars. But this year, Chinese domestic supply chains have matured rapidly; Lens Technology and other consumer electronics supply chain leaders entering the game means humanoid robots could follow the consumer electronics cost-down path: from hundreds of thousands to tens of thousands, then down to thousands.
Third signal: large models giving robots "brains." Previous humanoid robots were "pre-programmed motion playback" — only doing preset movements. Now, with end-to-end embodied AI foundation models, robots can understand natural language instructions, make autonomous decisions in unfamiliar environments, and learn new skills through trial and error. That's a truly "intelligent" robot, not a remote-controlled machine.
Rushing the IPO Window: List or Die?
LimX sprinting for an IPO at this juncture has a distinct "window-grabbing" flavor.
The humanoid race looks like this right now: Unitree is already public, UBTECH is performing decently in Hong Kong, Boston Dynamics is still inside Hyundai, Tesla Optimus moves in fits and starts, and a pack of Chinese startups (LimX, Agibot, Fourier, Star Era, etc.) are jostling for "first tier" position. Whoever lists first captures the most capital, the most talent, the most media attention — and seizes the initiative in the coming mass-production race.
But the risks are obvious: large-scale commercialization of humanoid robots is likely still 2-3 years away. Going public now means capital markets are buying "expectations," not "revenue." If mass production disappoints in 2-3 years, the valuation bubble could pop at any time.
"Humanoid robots are repeating the EV playbook: hype and fundraise → burn cash on production → shakeout leaves 3-5 survivors. But along the way, the upstream supply chain makes money first."— A hard-tech investor
But for companies like LimX, "not listing might be riskier." This track burns cash at an extraordinary rate — R&D, supply chain, factories, talent, every line item is massive expenditure. Without tapping public capital markets, you could run out of ammunition on the eve of mass production and get overtaken by competitors.
Capital isn't stupid. $400M in six months, $2B valuation, Pre-IPO round — behind these numbers is smart money betting on one thesis: the mass-production inflection point for humanoid robots is closer than most people think.
When Xiaomi's robot hits 98% on a production line, you know: that inflection point might really be around the corner.
See you tomorrow.
"人形机器人正在重复电动车的故事：先画饼融资→再烧钱量产→最后洗牌活下来3-5家。但这个过程中，上游供应链先赚钱。"
—— 一位硬科技投资人
"Humanoid robots are repeating the EV playbook: hype and fundraise → burn cash on production → shakeout leaves 3-5 survivors. But along the way, the upstream supply chain makes money first."
— A hard-tech investor
逐际动力 · LimX Dynamics · 人形机器人 · Pre-IPO · 2亿美元 · 150亿估值 · 港股IPO · 具身智能 · 小米机器人 · 98%合格率 · 汽车产线 · 量产奇点
LimX Dynamics · humanoid robots · Pre-IPO · $200M · $2B valuation · Hong Kong IPO · embodied AI · Xiaomi robot · 98% success rate · auto production · mass production inflection
