The disparity in per capita AI investment is significant. How can ABLE DIGITAL (02687) fill the gap?

How many students does an average graduate advisor in Chinese universities supervise? According to the Ministry of Education's "2025 National Graduate Education Quality Annual Report," the student-to-faculty ratio in ordinary universities at the undergraduate level is approximately 1:24, while at the graduate level, it is about 1:18.5, with some engineering programs exceeding 1:25. In comparison, during the same period, the student-to-faculty ratio at research universities in the United States at the graduate level is about 1:6.8, with institutions like MIT at approximately 3:1, Princeton at about 5:1, Stanford at roughly 6:1, and Harvard at around 7:1. When converted, the average number of students each graduate advisor in China is responsible for is about 3 to 4 times greater than that of similar institutions in the United States. The disparity at the undergraduate level is equally apparent. The average student-to-faculty ratio at American four-year institutions is roughly 1:17, about 30% lower than in China. Meanwhile, the annual enrollment scale of graduate students in China has surpassed 1.3 million, and the supply of faculty has not kept pace, leading to a continuously increasing student-to-faculty ratio. Chinese university faculty are nurturing a much larger student population under workloads several times that of their American counterparts. This presents a nearly unsolvable arithmetic problem in the short term. The large-scale hiring of teachers is constrained by the training period for faculty and administrative limits, while significantly increasing per-student funding is limited by fiscal capacity. The physical world's pace of catching up is slow, yet another set of numbers reveals a more urgent gap. By 2025, the total investment in information technology and AI in Chinese universities is expected to reach approximately 28 billion yuan, a year-on-year increase of about 33%. While this figure appears substantial, when distributed across 47.63 million enrolled students, the per-student AI investment is about 588 yuan, equivalent to less than 81 US dollars. Moreover, this 28 billion yuan is primarily concentrated in hardware procurement, with a low proportion dedicated to software, services, and ongoing operations, indicating that this is merely the starting line for AI in higher education. There is no unified official statistic for American universities, but estimates based on per-seat subscription models and IT budget structures suggest that per-student AI investment is around 200 to 500 dollars (EDUCAUSE 2025 IT Budget Benchmark), approximately 2.5 to 6 times higher than in China. Beyond the numerical disparity, the structural issues are even more glaring. A considerable proportion of AI expenditures in Chinese universities is directed toward hardware infrastructure such as servers, GPU clusters, and network upgrades. In contrast, American budgets are more focused on software, services, and ecosystem development, with the latter typically offering higher marginal returns. A white paper from the China Educational Equipment Industry Association disclosed a telling figure: the average construction cost of an AI laboratory in Chinese universities reaches 12.8 million yuan, yet the equipment idleness rate is close to 47.8%. Nearly half of the equipment fails to actual educational output. Money has been spent, but the classrooms remain unchanged. With a student-to-faculty ratio difference of 3 to 4 times and per-student AI investments differing by 2.5 to 6 times, these two sets of figures together represent the reality faced by higher education in China: limited resources, with the dual burdens of expansion and quality improvement weighing heavily on them. The traditional path is no longer feasible; catching up can only find breakthroughs through new technological dimensions. AI's impact on higher education goes far beyond merely adding a tool to existing teaching processes. It alters the very nature of educational production efficiency. On the teaching side, AI-assisted systems can generate personalized exercises and feedback based on each student's learning pace, making it technically feasible for one teacher to provide differentiated guidance to 200 students simultaneously. The same workload would require 4 to 5 teaching assistants in traditional models. In graduate training, AI research assistants handle repetitive tasks such as literature screening, data analysis, and initial assessments of experimental proposals, freeing advisors from inefficient labor to focus on high-value academic guidance. The experimental training area is also changing. Physical AI technologies simulate high-risk, high-cost, high-consumption experimental scenarios in digital twin environments, allowing students to operate repeatedly without the limitations imposed by scarce equipment. How far are these scenarios from universities? Policy-level deployments have already provided a timetable. Since 2025, a series of documents, including the "Opinions of the State Council on Deepening the Implementation of the 'Artificial Intelligence +' Action" and the Ministry of Education's "Action Plan for 'Artificial Intelligence + Education'," have been issued in rapid succession. The "14th Five-Year Plan" explicitly states "deepening the implementation of digital education strategies" as a subsector of higher education. The 2026 National Education Work Conference further proposed accelerating the practical application of artificial intelligence in the field of education, with general education in AI across all stages of education also needing to be expedited. From the overall arrangement of the State Council to the rigid indicators set by various ministries, policy signals are highly consistent. The introduction of AI into higher education has shifted from an encouraged initiative to a mandatory requirement. Demand-side data also illustrates the issue. By 2025, the market for AI solutions in Chinese universities has seen a year-on-year growth of over 30%, maintaining rapid expansion for three consecutive years, becoming the fastest-growing subfield of educational information technology investments. Driven by both stringent policy requirements and real resource gaps, this market continues to accelerate. Digging deeper, current AI investments in universities are mainly focused on hardware acquisitions, with software and services making up a smaller proportion. The 47.8% idleness rate of equipment highlights that what universities truly lack is the AI capability that can be integrated into teaching processes; server rooms and servers are merely preliminary conditions. Once AI evolves from an auxiliary tool to a fundamental teaching infrastructure, deeply embedded in course design, academic assessment, experimental training, and research assistance, the market potential will far exceed the limits defined by current informational budgets. Lets do some simple calculations. China has 47.63 million enrolled students with per-student AI investment currently below 81 US dollars. Even if, in the coming years, we only reach the lower end of the American standard of 200 dollars, the corresponding market size for university AI alone will be close to 70 billion yuan. If we hit the median of 350 dollars in the U.S., this figure will exceed 120 billion yuan. This figure pertains only to one segment of higher education, not accounting for the incremental replacement of hardware by software and services, nor the inclusion of K-12, vocational education, and lifelong learning sectors. Widening the perspective a bit further, digital education is a systemic project covering all education levels and scenarios. Higher education represents the segment most capable of paying, with the most rigid demand and the strongest policy support, and is also the most mature experimental ground for the application of AI technologies. By starting from higher education and extending to vocational education and basic education, moving from single applications to comprehensive teaching infrastructure, the long-term ceiling for this track is far above one hundred billion. ABLE DIGITAL (02687) has entered the spotlight at this juncture. The company's direction is clear: addressing the structural shortage of educational resources in Chinese higher education and assessing the extent to which AI can fill this gap. The company, based on its self-developed multimodal large model, drives AI capabilities from single-point teaching assistance into comprehensive teaching infrastructure, covering the entire chain from course design, personalized learning, academic assessment, research assistance, to realistic training. This capability, refined in university settings, naturally possesses the extensibility to be replicated in adjacent markets such as vocational education and corporate training. The multimodal large model constitutes the core technological asset of ABLE DIGITAL. According to public information, the company has connected to World Labs, forming a technical matrix covering text, audio, visual, and 3D generation, utilizing visual frameworks like VOM, VLM, among others. The entire chain of capability is controlled entirely in-house. For technology to land in the classroom, it relies on a service network. ABLE DIGITAL has established a national service system connecting technology development with classroom implementation. This is particularly crucial in the Chinese higher education market, where there are significant disparities in infrastructure, faculty level, and digitalization between different regions and tiers of universities, making a one-size-fits-all solution impossible. A nationwide service network allows the company to tailor local deployments and ongoing operations to the specific situations of different institutions, transforming cutting-edge capabilities into scalable teaching productivity. The actual results can already be seen. With ABLE DIGITAL's AI system, a professor can provide differentiated guidance to hundreds of students simultaneously. High-cost experiments, previously limited by equipment and space, are now transformed into digitally iterative experiences. AI has been integrated into daily teaching processes, no longer a dusty device in the lab. In May 2026, ABLE DIGITAL was selected as one of the first component stocks in the "Hong Kong Commercial Daily Technology Index." This recognition from the capital market reflects the acknowledgment of its technical positioning and growth logic. Returning to the initial arithmetic problem: The disparity in student-to-faculty ratios at Chinese universities will not be eliminated overnight, and fiscal investment cannot be infinitely expanded. However, digital efficiency provides a feasible path to bridging quality gaps under limited resources. One professor plus an AI system can replace the workload of five teaching assistants. A digital twin laboratory can allow students to operate expensive equipment they once could only observe. The 28 billion yuan is merely a subset of the current informational budget, with a year-on-year increase of 33% still on the rise, the 2.5 to 6 times gap in per-student investment yet to be fulfilled, and the issue of nearly half of equipment lying idle awaits resolution. These numbers together sketch a market that has just begun its journey and has a long-term ceiling far beyond one hundred billion. Policies are pushing, demand is rising, and technology is maturing. The turning point for universities to move from buying hardware to using software and providing services has arrived. What ABLE DIGITAL is doing is leveraging this lever. The technical breadth of the multimodal large model aligns with the depth of demand in higher education, and the elasticity of platform deployment addresses the unbalanced resource distribution in Chinese universities. While resource disparities objectively exist, efficiency is the only lever that can be used to lift them.