The Venture Codex

Participated · Series A · Aug 2026

SEALIEN ROBOTICS designs and sells work-class remotely operated vehicles (ROVs) and related subsea systems, including the 6,000-meter-rated Singularity work-class robot, the 3,000-meter-rated Phoenix cable-burial system and the 1,000-meter-rated Taurus heavy-duty robot. The company is also developing autonomous underwater vehicles and an AI-based decision system to coordinate multiple robots. Its commercial traction accelerated in 2025 after winning an order from a UAE telecommunications operator for a Phoenix 600 system valued at nearly US$10 million, which later expanded from one unit to three. By the end of July SEALIEN had signed several hundred million yuan in new contracts for complete deep-sea robot systems, expected to sign a similar amount soon, and reported a pipeline of about CNY 1 billion in potential contracts under negotiation. Customers now span the UK, the Netherlands, Norway, Brazil, Australia, the UAE, Egypt, Singapore, Malaysia and Indonesia. The company’s growing order book indicates a shift from trials to sales of complete machines for high-value offshore projects and an expanding overseas footprint.

Participated · Series A · May 2026

Vbot designs and manufactures consumer-facing companion robots and is concurrently developing humanoid robots and an Agentic OS for embodied intelligence. The company launched the Big Head BoBo smart dog in late 2025 and began deliveries on May 8, with the first 500 units shipped and plans to deliver over 1,500 units in May and exceed 2,500 units in June. Vbot plans to scale mass production and expand its sales network while investing R&D resources into full-size humanoid bodies and integrated manipulation-mobility models. Founded in late 2024, Vbot has a leadership team drawn from Horizon Robotics and Li Auto and a staff of over 100 covering product design, engineering, AI algorithms, and commercial operations. The company emphasizes iterative deployment in real environments to build data loops and accelerate the development of general-purpose service agents.

Participated · Equity · Feb 2026

Moonshot AI was founded in 2023 by Yang Zhilin, a former Meta AI and Google Brain researcher. The lab gained prominence with its open-weight Kimi models—Kimi K2.5 notably drove rapid adoption and Kimi K2.6 became one of the most-used LLMs on the OpenRouter distribution platform. Moonshot competes with models from OpenAI, Google, Anthropic, ByteDance, Alibaba, Zhipu and other Chinese AI labs. The company reported annual recurring revenue exceeding $200 million in April, fueled by paid subscriptions and API usage. Moonshot attracted significant investor support prior to this round, having raised about $3.9 billion over the prior six months and counting backers such as Alibaba, Tencent, HongShan, ZhenFund, IDG Capital and 5Y Capital. Its valuation progressed from about $4.3 billion at the end of 2025 to roughly $10 billion in early 2026 before the current $20 billion valuation.

Participated · Series B · Feb 2026

Founded in 2023 and based in Beijing, Galaxea AI develops embodied intelligence systems centered on Vision‑Language‑Action (VLA) models and World Action Models. The company emphasizes a tightly integrated approach where VLA and World Action Models co‑evolve using massive real‑world datasets collected from open environments. Galaxea has open‑sourced three VLA models to date and contrasts its real‑data‑driven approach with competitors that rely primarily on simulation. The firm acknowledges challenges of high data collection difficulty, cost, and slower scalability inherent to its chosen data strategy. Following a nearly CNY1 billion Series B in February, Galaxea completed a nearly CNY2 billion Series B+ at a CNY200 billion valuation to fund further R&D, ecosystem expansion, and commercial deployment efforts.

Participated · Series A · Jan 2026

DeepWisdom (深度赋智) develops Atoms, formerly known as MetaGPT, an AI programming platform that organizes multiple intelligent agents into a virtual startup team so a single user can design, code, deploy and operate real products. The system lowers the marginal cost from idea to live business by systematizing the entire entrepreneurial workflow and supports a proprietary “race mode” that lets several models compete on code quality, security and UX to deliver the best result at 60 % of competitors’ cost. In the Xbench deep-research benchmark, Atoms scored 73, outperforming OpenAI o3 (65+) and Gemini 2.5 Pro (60+), and the platform is SOC 2 certified, GDPR compliant and end-to-end encrypted. Market traction has been rapid: in its first month, the product generated US$1 million in annualized recurring revenue with zero marketing spend, signed up more than 500,000 global users and stayed atop Product Hunt’s global leaderboard for four consecutive weeks. With its new funding, DeepWisdom will iterate on core multi-agent capabilities such as memory management and intent recognition, scale product deployment and open a Silicon Valley office to attract international talent. The founding team includes alumni of Tencent, Huawei, Google, Anthropic and ByteDance, led by Xiamen University computer-science graduate Chenglin Wu. The company has now secured the largest financing to date in China’s Coding Agent sector.