09/11 2026
353
On September 10, the Qinghai Provincial Education Department and iFLYTEK entered into a strategic cooperation agreement in Hefei. The collaboration is set to focus on smart education initiatives, AI-driven general education courses, digital teacher training, and the promotion of standard Chinese language instruction.
While the news itself may not be revolutionary, it takes on a more intriguing aspect when considering iFLYTEK's current position.
What is Qinghai's Rationale Behind This Move?
According to the agreement, both parties are committed to integrating AI applications across several core educational domains: teaching, learning, examination, evaluation, and management. Specific plans include the launch of smart education pilot programs, the development of localized AI general education courses, and the utilization of AI assessment technology to bolster reforms in foreign language listening and speaking exams.
A notable aspect of the agreement is its emphasis on leveraging Qinghai's multi-ethnic regional characteristics to reinforce the teaching of standard Chinese language and to pioneer a model for AI-enhanced education in ethnic minority areas.
It's no secret that educational resources in Qinghai are relatively scarce, with significant disparities between urban and rural areas. While it remains uncertain whether AI can fully address these challenges, it at least offers a promising avenue for experimentation. For iFLYTEK, this collaboration not only aligns with policy directives but also secures an early foothold in the western market.
Since the start of the year, nine departments, including the Ministry of Education, have issued documents explicitly endorsing the digitalization of education in rural, ethnic, and previously impoverished regions. iFLYTEK's partnership with Qinghai is in line with this focus. Earlier, in April, the Liaoning Provincial Education Department signed a similar agreement with iFLYTEK. The educational footprint is thus extending from Liaoning to Qinghai.
The Real Picture from the Half-Year Report
Just prior to the announcement of the signing, iFLYTEK released its 2026 half-year report. The financial results present a mixed bag—or more candidly, continue to show losses.
Revenue for the first half amounted to 11.623 billion yuan, marking a 6.52% year-on-year increase. The open platform business surpassed smart education for the first time, becoming the primary revenue generator at 3.705 billion yuan, a 36% jump. Revenue from large model APIs and MaaS platform services saw a roughly 70% year-on-year increase, with the developer base expanding to 11.53 million.
However, net profit attributable to shareholders reported a loss of 204 million yuan. Although this represents a 14.68% year-on-year narrowing, the loss from non-recurring profits and losses widened to 637 million yuan, a 74.88% year-on-year decline. This suggests that the core business's profitability remains fragile.
During the earnings call, management addressed these concerns head-on, acknowledging that "many investors have voiced dissatisfaction." Chairman Liu Qingfeng explained that the company is ramping up investment, facing pressures on C-end consumption, rapid technological shifts in AI leading to extended customer decision-making cycles, and issues with the pace of government project execution—all contributing factors.
Is the 3 Billion Yuan H1 R&D Investment Justified?
A substantial portion of the escalating losses can be attributed to increased R&D investment.
R&D spending in the first half reached 3.007 billion yuan, a 25.73% year-on-year increase, accounting for over 25% of revenue. Funds were predominantly allocated to projects such as the Spark large model, AI glasses, and intelligent grading machines.
Liu Qingfeng indicated during the earnings call that R&D investment in the second half is expected to more than double that of the first half, with roughly half earmarked for large language models, speech, and multimodal research. By this calculation, full-year R&D investment could surpass 9 billion yuan.
This strategy of "sacrificing short-term profits for long-term gain" is not unique among tech companies. Nevertheless, the question persists: with the large model arena now crowded and the commercialization window uncertain, whether iFLYTEK can strike a balance between investment and returns remains to be seen.
On a more positive note, sales collections in the first half reached 11.896 billion yuan, a 1.535 billion yuan year-on-year increase, with a collection rate of 102%. Operating cash flow turned positive in the second quarter alone. At least the company's ability to generate cash is improving.
Scaling Back on G-End, Advancing on B-End and C-End
iFLYTEK's current strategy is to "strengthen C-end, deepen B-end, and selectively target G-end."
The data indicates that these adjustments are yielding results. G-end revenue declined by 2.65% year-on-year in the first half, while B-end and C-end combined grew by 10%, accounting for 76% of total revenue. Contract value increased by 27% year-on-year, signaling improved business quality.
However, C-end hardware faces challenges. Smart hardware revenue in the first half was 732 million yuan, a 15.96% year-on-year decrease; sales of AI learning machines dipped slightly by 1% year-on-year. The company attributed this to rising chip and storage costs, compounded by temporary supply shortages. Liu Qingfeng mentioned that the company anticipated price hikes early in the year and stockpiled inventory in advance, stating, "Without stockpiling, we would have had to spend several hundred million more in the second half."
Learning machine sales rebounded in July, with cumulative sales up 3% year-to-date as of July and single-month sales up over 30% year-on-year. Whether this recovery can be sustained will be evident in the third-quarter report.
The AI Education Market Is No Walk in the Park
Returning to education—iFLYTEK's long-standing sector—smart education revenue in the first half was 3.490 billion yuan, a slight 1.16% year-on-year decrease, surpassed for the first time by the open platform business. This shift underscores two issues: intensifying competition in the education informatization market and a strategic shift in the company's business focus.
Policy support is favorable. The Ministry of Education has emphasized the need to promote the regular application of "special delivery classrooms," "famous teacher classrooms," and "famous school online classrooms," confirming the direction of AI-driven educational transformation. However, AI applications in education require more than just technological prowess—they demand a nuanced understanding of teaching scenarios. Moreover, the execution pace and payment cycles of government projects pose real challenges to corporate cash flow.
Final Thoughts
iFLYTEK's agreement with Qinghai represents a routine step in its educational sector strategy. From Liaoning to Qinghai, its reach is expanding, but the signing is merely the first step. The real questions are whether implementation can succeed and what the outcomes will be.
iFLYTEK is currently at a critical juncture: R&D investment continues to rise, losses persist, but business structure adjustments are underway, cash flow is improving, and growth momentum in large models and open platforms remains robust.
While short-term performance may be lackluster, the long-term direction is relatively clear. Whether the company can navigate this transition period and convert investments into tangible revenue will be revealed in the coming quarters' financial reports.