As of May 10, 2026, over 1,500 AI toy companies are registered in China, with Huawei’s Smart HanHan plush toy selling 10,000 units in China in its first week. But despite the seemingly innocent nature of these toys, the AI kids’ toy market is still largely unregulated, with little oversight or accountability. This raises concerns for parents, developers, and anyone invested in the future of AI-assisted education.
Key Takeaways
- The AI kids’ toy market is a rapidly growing industry, with over 1,500 companies registered in China as of October 2025.
- The market is largely unregulated, with little oversight or accountability.
- Companies like Huawei, Sharp, and FoloToy are already selling AI-powered toys that claim to have sold thousands of units.
- The AI kids’ toy market is poised to become a major trend in cheap trinkets at trade shows like CES, MWC, and Hong Kong’s Toys & Games Fair.
AI’s growth Kids’ Toys
The AI kids’ toy market is a relatively new phenomenon, but it’s growing fast. According to the source, over 1,500 AI toy companies are registered in China as of October 2025. This is a staggering number, especially considering that the market is still largely unregulated.
AI’s growth in toys didn’t happen overnight. It’s rooted in broader shifts in consumer tech and AI accessibility. By 2020, voice recognition and natural language processing had improved enough to be embedded in low-cost hardware. That same year, Amazon’s Alexa-powered children’s devices hit global shelves, proving there was appetite for interactive, responsive toys. The pandemic accelerated demand as parents sought digital tools to keep kids engaged during lockdowns. By 2023, startups in Shenzhen began reverse-engineering similar AI functions using cheaper chips and open-source models, slashing production costs.
China’s manufacturing infrastructure enabled rapid scaling. A single factory in Dongguan can now produce 30,000 AI toy units per day, integrating microphones, Wi-Fi modules, and pre-trained language models for under $15 per unit. This cost efficiency has turned AI toys into disposable consumer items—more akin to seasonal gadgets staffing booths at trade shows lower prices lower barriers to entry. The result? An explosion of brands, many operating under shell companies with no public track record.
The timeline is telling. In 2022, fewer than 200 AI toy firms were registered in China. By 2024, that number had jumped to 700. The spike to 1,500 by late 2025 reflects not just growth but a land rush mentality—companies flooding the market before regulations can catch up.
The Unregulated Market
The lack of oversight and accountability in the AI kids’ toy market is concerning. With little regulation, it’s difficult to ensure that these toys are safe and effective. This raises questions about the quality and safety of these products.
There are no mandatory standards for data handling, voice recording storage, or model training sources. Some toys record every word a child says, uploading it to unsecured cloud servers. Others use AI models trained on unverified datasets, which can lead to unpredictable responses. In early 2025, a FoloToy unit was reported to have repeated a racial slur after a child mispronounced a word—an incident that sparked public outcry but prompted no formal recall.
The absence of certification bodies compounds the issue. In the U.S. the Federal Trade Commission enforces COPPA, which limits how children’s data can be collected. The EU has GDPR-K, a stricter framework for minors. But China has no equivalent nationwide rule for AI toys. Local consumer protection agencies have issued advisories, but enforcement is inconsistent. A probe by Guangdong provincial authorities in December 2025 found that 62% of sampled AI toys failed basic cybersecurity protocols, yet none were pulled from shelves.
The hardware itself is also a risk. Many toys use third-party firmware with known vulnerabilities. Some rely on outdated AI chips that can’t support over-the-air security updates. That means once a toy is sold, it can’t be patched—leaving children exposed to potential hacking. Researchers at Tsinghua University demonstrated in early 2026 that one popular model could be remotely accessed to activate its microphone, turning it into a surveillance device.
Even basic durability standards are ignored. Drop tests conducted by an independent lab in Shenzhen showed that 40% of AI toys cracked after a one-meter fall, exposing internal circuitry. For young children, that’s a choking hazard. Yet these products carry no warning labels.
The Players
Companies like Huawei, Sharp, and FoloToy are already selling AI-powered toys that claim to have sold thousands of units. Sharp’s PokeTomo talking AI toy, for example, went on sale in Japan this April. Huawei’s Smart HanHan plush toy sold 10,000 units in China in its first week.
Huawei’s entry is particularly notable. Known for telecom infrastructure and high-end smartphones, the company launched Smart HanHan as a “learning companion” for children aged 3 to 8. It responds to voice commands, tells stories, and quizzes kids on basic math and vocabulary. Its first-week sales were boosted by integration with Huawei’s existing ecosystem—parents can monitor interactions through the Huawei Home app. That connectivity, while convenient, also means the toy collects voice data linked to family accounts.
Sharp’s PokeTomo, released in Japan, takes a different approach. Marketed as a “digital pet,” it uses AI to simulate emotions, reacting to tone and repetition in a child’s voice. It purrs when spoken to gently, “cries” if ignored, and learns to recognize its owner’s voice over time. Sharp claims it helps develop empathy in children. But critics point out that it reinforces dependency on artificial responses, potentially skewing emotional development.
FoloToy operates on the lower end. Based in Shenzhen, it sells a range of AI dinosaurs, robots, and animal figures for under $30. Its marketing highlights “smart conversation” and “adaptive learning,” but teardowns show it runs on a generic NLP model with scripted responses. Still, it’s popular in tier-2 and tier-3 Chinese cities, where price matters more than features.
Other players include startups like Xiaozhi Tech and MiaoMiao Interactive, both of which launched AI storytelling toys in 2025. These devices allow kids to co-create stories with the AI, which responds in real time. While creative, they’ve drawn scrutiny for generating narratives with violent or inappropriate themes when prompted indirectly.
The Potential Implications
The AI kids’ toy market is poised to become a major trend in cheap trinkets at trade shows like CES, MWC, and Hong Kong’s Toys & Games Fair. But as the market grows, so do the concerns about safety and effectiveness. Parents and developers will need to be vigilant in ensuring that these toys meet certain standards.
At Hong Kong’s 2026 Toys & Games Fair, over 80% of new product booths featured AI toys. Many were nearly identical—plush animals with glowing eyes and voice sensors—differing only in branding. Some vendors offered “white label” AI toys, allowing retailers to slap on their own logos. One distributor claimed they could deliver 10,000 units in three weeks, complete with custom voice responses.
This commoditization risks turning AI toys into fads, not tools. When novelty wears off, many will end up discarded. Environmental groups estimate that by 2027, discarded AI toys could contribute over 5,000 tons of e-waste annually in China alone. Recycling is rare—most units contain glued components that can’t be disassembled.
Beyond waste, there’s a cognitive cost. Some educators warn that constant AI interaction may shorten attention spans. A 2025 pilot study in Guangzhou found that children who played daily with AI toys spent 30% less time in unstructured imaginative play compared to peers with traditional toys. The AI’s immediate feedback loop—answer correct? praise! answer wrong? correction!—creates a dependency on external validation.
At the same time, these toys could widen the digital divide. High-end models like Smart HanHan are priced at $99, putting them out of reach for many families. Lower-cost alternatives, while accessible, often cut corners on safety and performance. That means wealthier children get smarter, safer AI tools—early advantages in language and logic—while others are left with glitchy, insecure devices.
What This Means For You
As a parent or developer, it’s essential to be aware of the AI kids’ toy market and its implications. With the market growing fast, it’s crucial to ensure that these toys are safe and effective. Here are a few takeaways to consider:
* Research the companies behind these toys to ensure they have a good track record.
* Look for reviews and ratings from other parents and users to get a sense of the toy’s quality.
* Be cautious of toys that claim to have advanced AI capabilities without providing clear information about their safety and effectiveness.
For developers, the rush into AI toys presents both opportunity and risk. If you’re building a new product, you’re entering a crowded, low-margin space where differentiation is hard. But there’s room to stand out with transparency—open-sourcing your model, publishing data policies, or offering local-only processing. One indie developer in Chengdu gained traction in early 2026 by releasing an AI teddy bear that works entirely offline, appealing to privacy-conscious parents.
Founders should consider the long game. Regulatory pressure is building. China’s Ministry of Industry and Information Technology has signaled it may introduce AI toy safety standards by 2027. Companies that ignore data hygiene now could face bans or fines later. Building with compliance in mind—like encrypting voice data or allowing manual deletion—won’t just protect users. It could become a selling point.
For educators and schools, the question is whether to adopt these tools at all. Some private kindergartens in Shanghai have started using Smart HanHan in language labs, citing improved vocabulary retention. But public schools remain cautious. Without proof of long-term benefits, many administrators see AI toys as distractions, not aids.
Competitive Landscape
The AI kids’ toy market isn’t just crowded—it’s stratified. At the top are tech giants like Huawei, using brand trust and ecosystem integration. They offer polished products with app support, parental controls, and regular updates. Their challenge is maintaining quality while scaling.
In the middle are regional electronics brands like Sharp, targeting specific cultural niches. PokeTomo, for instance, fits Japan’s long-standing fascination with robotic companions. These companies often have stronger safety records but move slower than startups.
At the bottom is a flood of no-name manufacturers, mostly based in southern China. They dominate online marketplaces like Taobao and Pinduoduo, where price drives decisions. These toys rarely list technical specs or data policies. Some don’t even have customer service numbers.
The real competition isn’t just between brands—it’s between philosophies. Is an AI toy a learning tool, a playmate, or a data collector? Huawei markets Smart HanHan as educational. FoloToy leans into fun. But the business model for most is data accumulation. Voice samples from thousands of children can be used to improve speech recognition models, train emotion-detection algorithms, or even build behavioral profiles.
That data has value beyond toys. A 2025 report noted that several AI toy startups were acquired by larger AI firms—not for their hardware, but for their datasets. One company, BabyBrain Tech, was bought by a voice analytics firm after collecting over 500,000 hours of child speech. The acquisition wasn’t publicly disclosed until months later.
Forward-Looking Questions
As the AI kids’ toy market continues to grow, it’s essential to ask questions about its future implications. Will these toys become more advanced, potentially giving children an unfair advantage? Or will they help to level the playing field and provide equal opportunities for all?
Who owns a child’s voiceprint after it’s recorded by a toy? Can it be sold, shared, or used to train other models? What happens when a toy’s server goes offline—do kids lose access to their “friend”?
Will governments step in, or will self-regulation prevail? Can industry groups create a certification mark for safe AI toys, like UL for electronics? And if they do, will consumers care—or just chase the cheapest option?
The answers won’t come quickly. But one thing’s clear: the teddy bear your child hugs today might be the most powerful AI device in your home. And right now, no one’s watching what it hears.
Sources: Ars Technica
