PaXini Tech’s $150M Series B Signals a Unicorn-Scale Push in China’s Embodied AI

PaXini Tech just raised over $150 million, lifting its valuation above $1.5 billion and cementing its place at the forefront of China’s embodied intelligence race.

In the Chinese AI ecosystem, PaXini’s win is more than a funding round. The company touts a globally unique, hundred-billion-scale real-world multi-modal data system that it says is capable of training VTLA embodied intelligence large models. Translation: the core moat isn’t just clever software, it’s a data and perception hardware flywheel. The Tianjin-based “Super EID Factory” reportedly produces nearly 10 billion high-quality data points each year, spanning modalities as rare as tactile input, a capability many rivals guard as a strategic asset. The dataset, branded as the OmniSharing DB, underpins PaXini’s claim to a scalable, real-world-training regime for embodied AI that can act in physical environments.

The financing round was led by a trio of capital firms—Huangpujiang Capital, Kaitai Capital, and Xin’an Capital—with participation from dual-currency funds and strategic players like Zhuhai Science & Technology Industrial Group, alongside other institutional investors. In a sign that China’s AI funding network remains deeply mixed, existing shareholder Yida Capital oversubscribed, and Gengxin Capital China served as exclusive strategic advisor. In total, PaXini has drawn 14 top-tier investors worldwide, including affiliates of Meta and leading Chinese industrial players such as BYD, JD.com, and TCL, underscoring a rare blend of domestic-market focus and global capital participation.

From a policy-to-factory perspective, PaXini’s story illustrates several ongoing dynamics in China’s manufacturing AI ecosystem. First, the “感知驱动的通用智能” (embodied perception-driven general intelligence) framework is increasingly treated as a production-ready architecture, not a mere research curiosity. The emphasis on high-precision perception hardware and real-world data collection aligns with national aims to build autonomous, resilient industrial platforms that can operate across factories, warehouses, and logistics hubs.

Second, the investor mix signals a hybrid ownership environment: private funds feed rapid growth while state-connected groups (for example, Zhuhai’s city-backed technology entity) participate to anchor scaled infrastructure, data pipelines, and potential industrial deployments. That blend matters on the factory floor because it can translate into faster pilot deployments, policy-aligned data-sharing practices, and smoother access to domestic supply chains for sensors, edge hardware, and autonomy software.

Third, PaXini’s scale story—about 10 billion data points per year and multi-modal data coverage including tactile modalities—speaks to a broader trend: data-first AI is becoming a factory asset. In China’s sourcing and manufacturing ecosystems, this means suppliers and partners increasingly win or lose based on who can feed the most diverse, high-quality real-world data into trained models that can guide robotics and automation on the line.

What this means for global manufacturers and suppliers: expect intensified competition around data ecosystems, perception hardware, and cross-vertical deployments. If a Chinese embotified AI company can translate data scale into reliable factory performance—faster cycle times, fewer breakdowns, more adaptive robotics—its supply-chain partners will want in. And as PaXini’s ecosystem expands, downstream manufacturers should watch for co-development opportunities, not just licensing deals—where industrial giants like BYD or TCL could co-design perception systems tuned to specific plant layouts or product lines.

Two practitioner takeaways:

Data scalability is a primary differentiator. For a supplier, aligning with a partner that can continuously feed the system with diverse, high-fidelity data may be as valuable as the robot hardware itself.

Ecosystem synergies matter. The mix of private capital and state-backed groups points to a future where scale requires cross-institution collaboration—data platforms, component makers, and system integrators all moving in concert.

As PaXini scales, the next milestones to watch include the expansion of the Tianjin data pipeline, the performance of VTLA models in live production environments, and the degree to which the OmniSharing DB becomes a de facto industry dataset for embodied intelligence across sectors.

Sources

PaXini Tech Secures Over $150 Million Series B Financing, Valuation Surpasses $1.5 Billion

Knowin Completes Angel+ Round Financing, Valued Over $300 Million