AI Large Models Land in Learning Devices, Reigniting the Battle for a Trillion-Dollar Market?

[baoshi.rao]

On February 20, reports emerged that ByteDance has launched the 'Hippo Love Learning' app, which integrates AI capabilities targeting elementary, middle, and high school students, offering services such as photo-based problem-solving, homework assistance, oral arithmetic, essay grading, and literary cultivation.

ByteDance's education business once suffered significant setbacks but is now attempting to leverage AI large models to continue exploring the feasibility of its education ventures.

An education industry insider close to ByteDance told DoNews, 'Software has always been ByteDance's strength, but its ambitions in the education sector go far beyond that. With some AI products gradually being implemented, it’s not out of the question for ByteDance to re-enter the hardware space, potentially launching AI learning devices and other educational smart hardware products. Previously, ByteDance faced numerous challenges with its educational smart hardware products, but AI learning devices could be a breakthrough point.' In fact, AI learning devices have become a battleground for many education companies and internet giants. Currently, manufacturers that have launched AI learning devices include iFlytek, Zuoyebang, Xueersi, BBK, NetEase Youdao, Xiaodu, and others.

Although the features and functions of AI learning devices vary among manufacturers, they all incorporate AI technology to provide targeted learning solutions and resource recommendations based on students' learning progress, thereby improving learning efficiency. Additionally, these AI learning devices are equipped with abundant learning resources, including textbook synchronization, expert lectures, and extracurricular extensions, catering to the learning needs of different stages and subjects.

"The industry consensus is that the current education environment is constantly evolving. With the trend toward online education and the rapid iteration of AI technology becoming the norm, learning devices are expected to become popular educational smart hardware products again," said an industry insider. The AI learning tablet market is continuously expanding. According to the 2022 China Educational Smart Hardware Industry Report, learning tablets have become a key competitive focus in the educational smart hardware sector due to their adaptability to home learning scenarios and high market penetration. The educational smart hardware market holds immense growth potential, with an expected market size nearing 100 billion yuan by 2024 and a compound annual growth rate of 26%.

The 2021-2026 Learning Tablet Industry Venture Capital Trends and Investment Strategy Guidance Report published by the China Research Institute of Industry also indicates that China's learning tablet market will grow steadily at an annual rate of 15% from 2021 to 2025. By 2025, the market is projected to reach 8.75 million units, doubling the 2020 figures.

With the rapid iteration of AI technology, the functionality of AI learning tablets continues to improve and optimize. They now offer comprehensive coverage for home, school, and outdoor scenarios, attracting numerous educational manufacturers and major internet companies to enter the market. iFLYTEK was among the first internet giants to capitalize on the AI learning tablet trend. In July 2021, the company hosted a product launch event themed 'Ushering in a New Era of AI Learning,' officially rebranding its smart learning tablets as iFLYTEK AI Learning Tablets and unveiling the T10 model. The T10 represents iFLYTEK's latest technological achievement in the context of China's 'education burden reduction' policy, featuring breakthroughs in AI-based learning diagnostics and personalized improvement technologies.

Since then, iFLYTEK has successively released multiple AI learning tablet models. Its current lineup includes the X2Pro, X3Pro, X3 5G, T20, T20Pro, as well as the LUMIE10 and LUMIE10 Pro launched in July 2023.

Taking the relatively detailed financial report for the first half of 2023 as an example, iFlytek stated in the report that after the release of the Spark cognitive model, from May to June 2023, the company's GMV for C-end hardware reached a record high, doubling year-over-year. Among them, the GMV of iFlytek AI learning devices equipped with the Spark cognitive model grew by 136% and 217% in May and June, respectively.

The consumer market has responded accordingly. In 2021, iFlytek's market share in the learning device sector was about 4.0%, ranking fifth. However, by the first five months of 2023, its market share on JD.com's e-commerce platform had reached 16.1%, showing a clear upward trend. Particularly in May 2023, iFlytek's single-month market share on JD.com reached 23.6%, jumping to first place in the industry, with a 55.89% month-over-month increase in sales of its learning device products.

Baidu and Alibaba have also promptly entered the AI learning device market. In October 2023, Baidu launched the Xiaodu Qinghe All-in-One Learning Device at the 2023 Baidu World Conference, combining large model technology, teaching methods, and personalized education to serve as a one-on-one AI family tutor for children. Meanwhile, Alibaba's Tmall Genie has introduced the Z20 learning tablet, featuring a dual-capability interaction mode that combines generative AI with multi-sensory AI. Unlike previous learning hardware that primarily focused on "problem-solving through photo recognition" or "providing online courses," users can now engage in free bilingual conversations with the Z20 tablet to create personalized learning environments.

NetEase Youdao has also joined the race. In January 2024, it launched the Youdao AI Learning Tablet X20, powered by its "Ziyue" large language model. As early as the second quarter of 2023, NetEase Youdao reported in its financial results that driven by the "Ziyue" model, its smart hardware segment achieved net revenues of 220 million yuan, with gross margins increasing from 30.6% in the same period of 2022 to 35.8%. The company aims to leverage large models to provide more personalized and customized services, hoping that AI learning tablets and dictionary pens will jointly drive smart hardware revenues.

The influx of competitors has propelled AI learning tablets into the fast lane, but it has also led to a market flooded with products of varying quality. Currently, the AI learning tablet sector still requires more intense competition and consolidation to separate the wheat from the chaff. For manufacturers, AI learning devices represent a new business avenue that can generate additional revenue and help establish a foothold in the education market, opening up more possibilities. However, it's important to acknowledge that the development of AI learning devices still faces numerous challenges.

First, AI learning devices are relatively expensive. For example, the Xiaodu Qinghe Learning All-in-One retails for as high as 9,999 yuan; the high-end version of iFLYTEK's LUMIE10 Pro, equipped with the Spark large model, is priced at 9,899 yuan, while the T20 model costs around 7,999 yuan and the T20 Pro around 8,999 yuan. NetEase Youdao's AI Learning Device X20 starts at 5,999 yuan, and the Xueersi xPad2 Pro series begins at 6,599 yuan.

The high prices make it difficult for AI learning devices to achieve rapid market penetration. Data from iResearch shows that in 2022, China's online education market was valued at 582.5 billion yuan, while the market size for general AI products was only 21.11 billion yuan, representing a penetration rate of just 3.6%. Secondly, the educational value of AI learning devices has not yet gained widespread recognition, leading to ongoing developmental obstacles. Currently, the implementation of large AI models is still in the exploratory phase, with uncertain outcomes, especially when applied to the education sector, which has a lower tolerance for errors. This requires time for refinement and validation.

An AI industry insider told DoNews, "The core of AI learning devices lies in AI technology, but current AI technology still faces challenges such as annotated training data, comprehensive data acquisition, interpretation of output results, and the universality of learning. While integrating AI technology with learning devices is not difficult, the real challenge lies in the content—such as courses, teaching methods, and question banks. This process cannot be achieved in a short time but requires long-term accumulation."

Indeed, this is the case. Feng Xin, a parent who purchased an AI learning device, told DoNews that the learning content provided by the device sometimes does not align with actual teaching needs, making it difficult for students to follow. In terms of accuracy, AI learning devices also have issues. For example, when using an AI learning device to solve a math problem, the same problem was solved three times, yielding three different answers. Finally, at this stage, AI learning devices still suffer from severe homogenization. As more players enter the AI learning device market, the intensifying industry competition has also exacerbated the phenomenon of product similarity.

From a product functionality perspective, most AI learning devices on the market offer similar features, such as adaptive learning, precision teaching, and AI interaction. Although different brands may vary in some details, they generally lack innovation and differentiation, making it difficult to gain recognition from students and parents.

It is evident that current AI learning devices cannot fully replace parents and teachers, which is the fundamental reason for their developmental challenges. To boost the penetration rate of AI learning devices, manufacturers need to: first, address user pain points by conducting more data training for large models; second, leverage their unique strengths to create highly distinctive AI learning device products. Whether it's internet giants or education companies, everyone is trying to dominate the AI learning device market, but this path seems far from easy.

On one hand, the development of AI learning devices requires significant investment, faces high technical barriers, and has relatively long R&D cycles, making immediate results unrealistic in the short term. On the other hand, the functionalities of current AI learning devices on the market are becoming increasingly similar, resulting in minimal differentiation between products.

"The path being difficult doesn't mean AI learning devices lack value," said the aforementioned education industry insider. "In fact, the core of AI learning devices lies in data accumulation. With a sufficiently large pool of learning and academic performance data, companies can provide users with diverse and personalized educational solutions." In the long run, the value of AI learning devices lies not only in supporting individual learning but also in their potential to become a vital data platform and solution provider for "multi-party co-education" involving families, schools, and governments.

For individuals, AI learning devices act as direct learning assistants, offering personalized content to improve learning methods and habits. For parents, they serve as powerful monitoring tools, enabling timely understanding of children's learning progress and providing necessary guidance. For teachers, the greatest value of AI learning devices is their ability to record students' learning activities in real-time, helping to refine teaching content and methods. For relevant government departments, the comprehensive use of AI learning devices allows for the widespread and effective collection of student learning data, offering critical support for educational system reforms.

"Amid strong policy encouragement for the development of smart education, there is a high demand for building smart classrooms in schools, and smart teaching methods urgently need upgrades. The purchase demand for educational smart hardware, represented by AI learning devices, is pressing. Related manufacturers are expected to expand their school-focused business through government funding, school procurement, and family-paid value-added services, achieving multi-device connectivity that covers both home and school environments," the analyst noted. They believe that AI learning devices remain a promising business, and more players are bound to enter this field in the future.