Xynova, the Chinese unicorn dexterous hand developer, is seeking a new round of funding, partially to scale up mass production capacity, founder and CEO Alex Xia told Mergermarket.

The round is already in the market, drawing strong interest from top-tier investors, Xia said, adding that the company remains open to both industrial partners and financial investors.

Proceeds will also be allocated across continued R&D and building out the company’s data ecosystem, he added.

The company has raised approximately CNY 1.5bn (USD 223.2m) across four rounds in less than 12 months, with existing shareholders including internet and new energy vehicle industry players such as CATL, Xiaomi, JD.com, Meituan and NIO, according to a July press release on company’s official WeChat account.

Founded in 2024 and headquartered in Hangzhou, Xynova develops dexterous hands and the algorithms that control them. The company positions itself as a full‑stack provider for embodied AI manipulation.

Its flagship bionic dexterous hands cover the Flex and Prima series. The Flex 2 employs a proprietary hybrid drive technology combining tendon and direct‑drive systems, whereas the Prima 1 features a purely direct‑drive architecture. Both enabling highly fine-grained operations such as drawing precise lines and graphics and assembly in factory manufacturing.

The company targets an annual production capacity of 10,000 dexterous hands by the end of 2026 and currently employs over 500 people, with close to 300 engineers.

Product and strategy

Xynova operates a dual-product line strategy to address two distinct market needs.

Prima 1, a purely direct-drive hand launched in August, is aimed at algorithm developers and researchers who need a simple, controllable platform for training and testing models. Flex 2, the company’s flagship product, uses a hybrid-driven approach combining tendon and direct-drive technologies and is built for real-world industrial and consumer deployment.

“They are not a migration path from one to the other,” said Xia. “They are different products for different needs, and they will coexist.”

Behind any capable dexterous hand lies a robust data ecosystem. “High-quality datasets are extremely scarce globally; the entire industry has maybe only 100,000 hours of high-quality dexterous hand data,” Xia noted.

Xynova is building a pipeline for dexterous manipulation datasets with partners to collect, clean, annotate and validate real-world data from factory deployments.

Becoming the Nvidia of embodied AI

Rather than building full humanoid robots, Xynova focuses exclusively on the dexterous hand, which Xia describes as “the interface between AI and the physical world.”

“You need balance across all product metrics — degrees of freedom, force control, reliability — just to get a barely passable product,” he said. Because robots must operate in real human environments, Xia believes the best model worth following is the human hand.

“We aim to be the Nvidia of dexterous manipulation — a player that is indispensable in our own sector,” Xia said. That ambition, he explained, requires a level of technical and manufacturing prowess that puts the company in a league of its own.