The Chinese humanoid-robotics cohort gained a quieter, more interesting entry on Monday, May 11, when Shanghai-based Vbot closed a Pre-A round worth roughly 500 million yuan (~US$73 million), per The AI Insider and Gasgoo. The round was led by Dongfang Jiafu, Huatai Zijin and Fosun Ruizheng, with participation from SAIC Capital, Minghui Zhiyuan, and existing investors Cathay Innovation, Capital Today, and Hillhouse.
The funding line item that distinguishes Vbot from the rest of the 2026 humanoid cohort is not on the cap table. It’s on the warehouse floor. Per Dealroom, Vbot is already shipping its first commercial product — the BoBo quadruped robot dog, debuted at CES 2026 as 「SuperDog」 — with an initial production run of 500 units underway and a stated plan to scale monthly capacity to more than 2,500 units in June.
That run rate — 2,500 robot dogs per month, or ~30,000/year — is in the same ballpark as Boston Dynamics/Hyundai’s stated Atlas factory target of 30,000 robots a year, and it is happening at a company that is fourteen months old.
The cap table reads like an auto-AI alumni network
Vbot was founded in late 2024 by a leadership group pulled almost entirely out of the Chinese autonomous-driving stack:
- Yu Yinan — CEO; formerly VP and an early founding member of Horizon Robotics, China’s leading auto-AI-chip and L2+ ADAS player.
- Song Wei — co-founder; formerly chief software architect at Horizon Robotics.
- Zhao Zhelun — co-founder; formerly director of smart-driving products at Li Auto.
The structural significance is not that Vbot exists, but that the team that built it migrated. Horizon Robotics’ Journey-series chips and BPU stack power a meaningful fraction of Chinese L2+ deployments, and Li Auto’s smart-driving product organization is one of the higher-velocity ones in the Chinese OEM cohort. Both have shipped at scale. Both have hardened a deployment loop on Chinese roads. That deployment loop — perception, planning, autonomy, on-device inference, OTA — is the same loop a humanoid robot runs.
Three of the named investors fit the same pattern: SAIC Capital (the investment arm of SAIC Motor, China’s largest automaker), Capital Today, and Hillhouse — China-auto and growth-stage capital, not Silicon Valley humanoid speculation money. The auto-AI alumni are funding their alumni.
What Vbot is actually selling, and what it’s planning to sell
The shipping product, per KrASIA’s coverage of the BoBo unveil, is a quadruped robot dog designed for in-home companionship — small, friendly silhouette, AI-driven interaction, marketed to live and play among humans rather than haul payloads on a factory floor. Vbot is also “expanding offline experience stores across China,” meaning the BoBo go-to-market is consumer-retail, not enterprise-pilot. The 2,500/month June target implies the company has committed to consumer SKU manufacturing volumes for a category where Chinese demand can move quickly.
The planned product, per the funding announcement, is a full-size humanoid body with both mobility and task-execution capabilities, built around two technical commitments:
- A 「hand-foot integrated world model」 covering general manipulation and locomotion in one stack.
- An 「Agentic OS architecture for embodied intelligence」 — the language is borrowed from the same vocabulary Figure’s Helix 02 release used in January and that GitLab’s 「agentic era」 letter is using this week to describe a developer-tools workforce reorg. The phrasing has migrated into Chinese humanoid pitch decks more or less verbatim.
In plain English: Vbot is using the cash flow and supply-chain learnings from a $1,500-ish home robot dog to fund the manufacturing build-out and the model training for a humanoid that is still inside the R&D building.
How this shape compares with the US humanoid cohort
A side-by-side, as of May 13:
| Company | Latest round | Valuation | Units shipped | Run-rate target |
|---|---|---|---|---|
| Figure AI (US) | ~$1B Sep 2025 | $39B post | ~350 Figure 03 cumulative | 1/hr = 12,000/yr (BotQ) |
| 1X (US, NO) | C-round | undisclosed | NEO preorders, pilot units | scaling |
| Apptronik (US) | $935M cumulative | $5B+ | low hundreds, BotQ-equivalent ramp | scaling |
| AGIBOT (CN) | growth | — | 10,000 humanoids cumulative (Mar 2026 milestone) | 100k target by EOY |
| Unitree (CN) | growth | — | 5,500 humanoids in 2025 | 10k–20k in 2026 |
| Vbot (CN) | $73M Pre-A May 11 | — | 500 robot dogs, scaling | 2,500/month in June |
Vbot is the smallest entry on this list by funding and by humanoid maturity. It is the most interesting because of where it sits in the lifecycle: a year and change after founding, with a commercial product on store shelves and a humanoid commitment funded by that product’s gross margin rather than by a $39B paper valuation.
That pattern — ship a dog, fund the humanoid — is by now the explicit Chinese humanoid playbook. Unitree shipped Go2 and Go1 robot dogs before the G1 humanoid. AGIBOT hit the 10,000-humanoid milestone in March 2026 after a multi-year dog-and-wheeled-platform business. Vbot is doing the same thing again, one cohort behind.
The 2,500/month number is the one to track
The funding number is not the news. $73M is small by US humanoid standards and small by AGIBOT’s growth-round standards. The number that matters in the press release is 2,500 BoBo units per month by June. That’s a Chinese consumer-retail commitment to a quadruped robot SKU at a volume previously associated with low-end small-appliance categories.
Two reads:
Optimistic: Chinese home-robotics demand is real, Vbot has presold or has visibility into channel demand, and the dog is the manufacturing-and-distribution proof of concept that justifies the humanoid build. If 2,500/month sticks for two consecutive quarters, Vbot has a self-funded humanoid pathway no US competitor has.
Skeptical: 2,500/month is a soft-commitment number from a startup with a Pre-A round and an unmarked stadium of inventory risk if Chinese consumer demand for novelty home robots stalls in the back half of 2026. The “experience stores expanding across China” line is the part that has historically been overpromised in the consumer companion-robot category.
The middle reading — which is the most likely — is that the 2,500/month figure is aspirational, the real number will be 30–60% of that, and even at the discounted number Vbot is doing something the US humanoid cohort is not: shipping a real consumer product while the humanoid is still in R&D.
What to watch
- June BoBo run-rate disclosure. Vbot has effectively committed to a 2,500/month number on the record. The disclosure or non-disclosure of the actual June number is the cleanest signal of whether the humanoid roadmap is funded.
- The Agentic OS architecture spec. Chinese humanoid pitch decks have started using the「agentic」 vocabulary identically with Figure’s Helix 02 stack and 1X’s NEO software story. Whether Vbot’s spec turns out to be a meaningful re-architecture or a marketing rebadge will read out in the first humanoid demo.
- Whether SAIC Motor pilots the humanoid in a plant. SAIC Capital is on the cap table. SAIC’s joint-venture vehicle plants are an obvious adjacent buyer, and a Chinese-OEM humanoid factory pilot would mirror the BMW Spartanburg → Leipzig play that Figure ran in the US.
- Vbot’s Series A. Pre-A rounds at this size in Chinese embodied AI are typically followed by a $150–300M A within 9–12 months if the BoBo number lands. If it doesn’t, the next round is a strategic from one of the SAIC-side investors at a flat or down valuation.
The cleanest read on the morning after is that the Chinese humanoid build has now produced a recognizable funding template: an auto-AI veteran founds an embodied-AI company in late 2024, ships a consumer robot dog within 12 months, scales the dog into a self-funding manufacturing line, and uses that to bankroll a humanoid that doesn’t have to convince a $39B valuation to keep going. Vbot is the latest instance. It will not be the last.