China Investors Bet Big on AI Toys as Next Hardware Frontier
Capital investment in artificial intelligence-integrated toys has reached a fever pitch in China in late 2025, emerging as the most concentrated sector for venture funding as the market shifts focus from large language model infrastructure to consumer-facing applications. With general software applications largely dominated by technology giants including Tencent and ByteDance, investors have identified intelligent toys as a high-growth vertical with proven demand and accelerated commercialization paths.
The sector has seen a flurry of activity in the fourth quarter of 2025, highlighted by Ling.ai, a startup founded only two years ago, closing a Series Pre-A financing round worth RMB 200 million (US$27.6 million) in November. This marks the company’s third capital injection this year, pushing its valuation to roughly RMB 1.2 billion, a trend mirrored by competitor Haivivi, which also secured significant funding and surpassed the billion-yuan valuation mark. This surge in activity brought the total number of financing events in the domestic AI toy sector to 19 in 2025, up from 14 in 2024, attracting top-tier firms such as Sequoia China and GSR Ventures.
This capital influx is driven by the maturation of generative AI models and plummeting inference costs, which have made complex interactions economically viable in low-cost hardware. Unlike the speculative nature of early AI investments, investors note that AI toys demonstrate substantial token consumption, indicating genuine user engagement and "emotional value" rather than mere novelty. However, the industry faces teething issues, including high product return rates and technical latency, as startups race to reach the milestone of one million unit shipments.
Surge in Valuations and Deal Potential
The AI toy landscape in 2025 represents one of the few primary market sectors seeing aggressive valuations and oversubscribed rounds. Ling.ai’s recent raise attracted a diverse coalition of investors, including state-backed Guofang Innovation, Didi, and listed companies like Runjian. The company, known for its "Ling.ai Little Cube" terminal, exemplifies the rapid ascent of hardware startups, having completed three funding rounds within a single year.
Similarly, Haivivi, established in 2021 and pivoted to AI in 2023, has solidified its position as a market leader. In June 2024, the company launched "BubblePal," an AI device attached to plush toys to enable conversation. Since its release, the product has sold over 300,000 units, generating revenue exceeding RMB 100 million. Haivivi and Ling.ai stand as the only two companies in the sector to disclose financing exceeding RMB 200 million this year, with both securing cumulative capital across six rounds.
Other emerging players, including Robopoet, KidoPally, and Ropet, also successfully closed finding rounds in 2025. Data from IT Juzi indicates that most deals this year exceeded the ten-million-yuan threshold, signaling a maturation from seed-stage speculation to growth-stage support.
Strategic Backing from VCs and Tech Giants
Investment strategies reveal a bifurcated approach between financial heavyweights and industrial strategics. Sequoia China has adopted a track-focused strategy, investing in multiple competitors including Robopoet and Haivivi, betting on the broader "emotional interaction" category rather than a single technical solution. Meanwhile, MiraclePlus remains the most active angel investor, seeding early-stage projects like KIGLAND and Bobo Forest.
Allen Zhu, Managing Director at GSR Ventures, identified AI toys as his "boldest" investment of 2025. Zhu cited high token consumption rates by relatively small startups as proof of strong user stickiness and genuine demand. Consequently, GSR Ventures led Robopoet’s angel round and followed up with an additional investment within three months.
Corporate venture capital has also re-entered the arena to secure ecosystem synergies. Tech giants are not just investing but actively developing products. Huawei partnered with Robopoet to release the "Smart Hanhan," while Baidu collaborated with Shifeng Cultural on the "AI Magic Star." This participation from industrial players underscores the sector's shift toward becoming a central pillar of consumer electronics.
Cost Drivers and Supply Chain Advantage
The commercial viability of AI toys has been significantly enhanced by a collapse in technical costs. Fierce price wars among China’s large model providers have drastically reduced inference costs, allowing hardware manufacturers to offer long-duration, high-frequency AI conversations at competitive price points.
Furthermore, China’s established supply chain, particularly in the Pearl River Delta, provides a unique advantage. The ability to rapidly integrate chip modules, sensors, and traditional toy manufacturing allows for swift product iteration. Industry projections support this growth trajectory; Statista forecasts the global AI toy market will reach approximately US$35.1 billion by 2030, growing at a compound annual rate of 14-16%.
Consumer Retention and Operational Hurdles
Despite the optimism, the sector faces significant operational challenges centered on user experience. High return rates remain a persistent issue; media reports indicate return rates for some AI plush toys range between 30% and 40%. Haivivi founder Li Yong acknowledged that while early return rates exceeded 30%, they have arguably stabilized closer to 20% due to improvements in interaction quality.
Investors are increasingly scrutinizing "pseudo-AI" products that merely wrap open-source models in plastic shells without genuine innovation. Linear Capital, an investor in Ling.ai, emphasized that long-term value depends on software capability—specifically metrics like weekly retention and active sharing frequency—rather than just hardware sales.
The industry consensus views 1 million unit shipments as the critical threshold for establishing a sustainable brand. While several manufacturers have surpassed the 100,000-unit mark in 2025, reaching mass-market scale will require overcoming distinct technical bottlenecks regarding response latency and interaction depth.