On the evening of August 20, Iflytek Co., Ltd. released its interim report for the first half of 2026. The financial results showed the company generated revenue of RMB 11.623 billion during the period, representing a year-on-year increase of 6.52%, while net profit attributable to shareholders posted a loss of RMB 204 million, narrowing the deficit by 14.68% compared to the same period last year. Meanwhile, total sales collections reached RMB 11.896 billion in the first six months, with operating cash flow turning positive in the second quarter on a single-quarter basis.
On the surface, revenue growth remained positive, yet the profit side faced short-term pressure while cash flow showed marginal improvement. However, beneath these seemingly contradictory financial figures, the more noteworthy developments lie in the company's shifting revenue structure, evolving AI commercialization model, and improving operational quality. As the large model sector transitions from technological capability competition to industry deployment racing, Iflytek, leveraging years of deep cultivation in large model technology, has begun realizing performance along pathways including open platforms, MaaS, industry intelligent agents, and AI terminal devices.
Thus, the key takeaway from this interim report is not merely how much revenue was earned or profit achieved in the first half, but rather that Iflytek is advancing from the phase of intensive large model investment into the climbing stage of AI business value realization.
Business Structure Upgrade Drives Revenue Growth
One of the core drivers behind the company's first-half revenue growth stems from proactive business structure adjustments. From the perspective of the three segments—government, business, and consumer—the company has continued its core operating strategy of "optimizing government business, deepening enterprise business, and strengthening consumer business." Specifically, government-sector revenue declined 2.65% year-on-year in the first half, attributable to the deliberate contraction of certain low-margin, delivery-heavy traditional project-based businesses. Simultaneously, combined revenue from the enterprise and consumer segments has reached 76% of total revenue, growing approximately 10% year-on-year and serving as the primary driver of overall revenue growth.
The shift in revenue composition under this framework indicates that Iflytek is not merely pursuing revenue scale, but rather reducing dependence on certain project-based operations, with growth momentum increasingly tilting toward the enterprise and consumer segments, which offer greater performance elasticity. Over the long term, this structural transformation carries far greater significance than short-term revenue growth rates alone: the rising proportion of enterprise, consumer, and operational-type businesses is also expected to enhance revenue sustainability and operating cash flow quality.
This shift becomes even more evident when examining the business segments disclosed in the financial report. During the first half, the open platform generated revenue of RMB 3.705 billion, up 36.01% year-on-year, with its revenue share rising to 31.87%, surpassing smart education to become the company's largest revenue source. Meanwhile, the open platform's gross margin improved by 3.63 percentage points year-on-year to 20.21%. Additionally, smart healthcare and smart automotive revenue grew 58.56% and 20.46% year-on-year respectively, representing important marginal growth points for company performance.
In the past, the open platform's business value was primarily reflected in AI capability interfaces and the developer ecosystem. In the large model era, guided by MaaS construction and intelligent agent industrialization strategies, its core business model has gradually extended into areas such as MaaS platforms, model invocation, Token services, and enterprise AI infrastructure. In the first half of this year, Iflytek's large model API and MaaS platform service revenue grew approximately 70% year-on-year, with the platform forming a dual-engine model of "API economy plus large model Token economy." As of the end of June, the platform boasted over 11.5 million AI developers, including more than 3.2 million large model developers, with the developer ecosystem scale continuing to lead the industry.
This effectively signals that Iflytek's AI commercialization is entering a new phase: the business model is progressively shifting from traditional project-based approaches to a "platform plus operations plus AI services" model, with AI capabilities beginning to translate into performance outcomes, pointing toward improved growth expectations ahead. The expansion of operational-type business scale and continuous operational quality improvements also appear in traditional advantage areas such as education. Although education business revenue declined 1.16% year-on-year in the first half, contract value grew 45% year-on-year, with a substantial volume of contracts still in delivery and acceptance phases. Meanwhile, intelligent grading machines have been deployed across more than 5,000 schools, and with the large-scale rollout of these devices, ongoing homework services generated approximately RMB 320 million in revenue during the first half.
Additionally, Iflytek AI learning devices underperformed sales expectations in the first half due to factors including chip and memory price increases and periodic supply shortages. However, in July, sales of Iflytek AI learning devices rebounded rapidly, with month-over-month sales growing more than 30% compared to the same period.
Short-Term Profit Pressure While Solidifying the AI Technology Foundation
Contrasting with the structural improvement on the revenue side, the company's profit side faced short-term pressure. However, breaking down the income statement reveals that the primary reason for the widened loss is not deterioration in core business, but rather the company's continued high-intensity investment phase in large models and new businesses. Profit pressure is closely linked to further increased R&D investment intensity; in the first half of 2026, the company's R&D investment reached RMB 3.007 billion, up 25.73% year-on-year. Selling expenses also remained at elevated levels; the company's selling expenses in the first half totaled RMB 2.285 billion, up 9.52% year-on-year. Promotion expenses for intelligent grading machines, imaging cloud, AI glasses, and overseas business alone exceeded RMB 100 million.
During this critical phase of business transformation, maintaining high levels of R&D investment is fundamentally about strengthening the AI technology foundation to support future performance realization. In the first half of this year, key technical progress included the release of the Spark X2 large model with comprehensive upgrades to general capabilities; the launch of X2-Flash with an open API, which enhances intelligent agent and coding capabilities while further reducing Agent application costs, laying the groundwork for continued commercialization of MaaS and the Token economy; and the release of the multimodal large model X2-VL, further expanding the boundaries of model capabilities.
More importantly, the pathway between the company's technical capabilities and commercialization is becoming increasingly clear. Particularly as model capabilities continue to strengthen, the ability to embed models into real business scenarios and generate economic value is the core variable that will truly differentiate companies in the next phase. Currently, Iflytek has established a complete commercialization pathway spanning model-platform-intelligent agent-terminal applications. Moreover, under the dual-engine model of "API economy plus large model Token economy," AI business has begun delivering performance, indicating that the company's large model applications are progressing from the capability validation phase into the value realization stage.
At a deeper level, short-term profit pressure and AI business growth are actually two sides of the same coin: the former primarily reflects current R&D investment intensity, while the latter demonstrates that earlier R&D investments are beginning to generate commercial returns.
Cash Flow Improvement Reflects Enhanced Operational Quality
On the cash flow front, although the company's operating cash outflows increased in the first half, the positive operating cash flow in the second quarter released a more encouraging signal. According to information disclosed during the company's performance briefing, operating cash outflows increased by RMB 1.98 billion year-on-year in the first half. Among these, strategic inventory build-up for items such as storage chips increased by RMB 626 million, and concentrated maturity of acceptance bills increased by RMB 423 million year-on-year, with these two factors explaining a significant portion of the incremental cash outflow in the first half. Meanwhile, sales collections in the first half reached RMB 11.896 billion, an increase of RMB 1.535 billion year-on-year, with the sales collection rate further improving to 102%.
In other words, the current cash flow pressure more accurately reflects supply chain inventory preparation and strategic investment timing rather than deterioration in operational quality. The actual improvement in collection capability serves as an important support for the marginal improvement in operating cash flow during the second quarter. This finding cross-validates the revenue structure changes involving the deliberate contraction of certain inefficient government projects and the increased proportion of enterprise, consumer, and operational-type businesses. As the open platform's revenue share rises and new businesses such as MaaS and large model APIs continue to grow, the company's core business model is progressively shifting from project-driven to operations-driven, a structural change also conducive to future cash flow improvement.
Summary and Outlook
In summary, the aforementioned financial discrepancies essentially reflect that Iflytek is currently operating in a phase where business transformation and high investment run in parallel. On one hand, the business structure is accelerating its migration toward enterprise, consumer, and operational-type services, combined with continuously improving collection capabilities, which is expected to further enhance revenue quality and cash flow levels. On the other hand, behind the short-term profit pressure lies high-intensity R&D investment, with solidifying the technology foundation serving as the cornerstone for safeguarding future performance growth.
From an industry perspective, over the past three years, the core of competition among large model companies has been more about technical capability and model iteration, with companies needing to continuously invest in computing power, talent, and R&D to secure tickets into the next stage of competition. Entering 2026, the core question in the AI industry has gradually shifted from "whether models can be built" to "whether models can genuinely generate revenue and profits."
Iflytek's advantage lies in its years of accumulation and presence across industries including education, healthcare, automotive, and consumer-facing businesses, providing the company with a relatively rich portfolio of real business scenarios. The revenue growth from open platform large model APIs and MaaS, the revenue increases in smart healthcare and smart automotive, and the gradual scaling of enterprise intelligent agents all indicate that the company's large model technical capabilities have begun realizing commercial value, precisely aligning with current industry trends.
Given that the company's AI business revenue is still climbing and expense outlays remain at high intensity, short-term profits are simultaneously absorbing substantial upfront costs generated by R&D, infrastructure construction, and new business expansion. Therefore, a valuation approach centered solely on current-period performance using price-to-earnings ratios struggles to fully reflect the multiplier effect of future AI business, rendering such metrics relatively limited in explanatory power. What will ultimately determine the company's valuation in the next phase remains whether business transformation and R&D investment can continuously translate into revenue from MaaS, Tokens, industry intelligent agents, and AI terminals, ultimately forming verifiable profits and cash flow.
The corresponding key points to watch include: whether AI business can sustain high growth, whether the revenue share of operational-type businesses can continue to rise, and whether R&D investment can ultimately convert into higher returns on capital. From this perspective, the 2026 interim report more closely resembles a phased turning point—Iflytek remains within its investment cycle, but the AI business has already begun delivering performance. As the company's model capabilities further land in specific application scenarios, what will truly determine the company's value going forward will no longer be solely the technical capability of the Spark large model or its ranking on leaderboards, but rather its ability to continuously enter real business operations and convert technical advantages into scalable, sustainable commercial returns.