Financial Performance - In the first half of 2025, the company's operating revenue reached 10.911 billion yuan, a 17.01% increase from 9.325 billion yuan in the same period last year [1] - Gross profit was 4.389 billion yuan, reflecting a year-on-year growth of 17.12% [1] - The net profit attributable to shareholders was -239 million yuan, an improvement of 40.37% compared to -401 million yuan in the previous year [1] - The net profit after deducting non-recurring gains and losses was -364 million yuan, showing a year-on-year increase of 24.62% [1] - The net cash flow from operating activities was -772 million yuan, narrowing by 49.73% from -1.536 billion yuan in the same period last year [1] Sales and Collection - The company achieved sales collections of 10.361 billion yuan in the first half of 2025, a year-on-year increase of 14.99%, marking the first time collections exceeded 10 billion yuan in a half-year period [1] - The collection rate has improved for three consecutive years, rising from 93.75% in 2022 to 97.20% in 2024 [1] Business Structure and R&D Investment - The revenue structure of the GBC segment continues to optimize, with C-end business revenue becoming the main growth driver. G-end accounts for 26%, C-end for 32%, and B-end for 42% [1] - R&D investment in the first half of 2025 grew by 9.2%, accounting for 21.92% of revenue, maintaining stable investment intensity [1] - R&D expenses increased by 610 million yuan, focusing on three main areas: core technology related to the Spark large model, industry application product development and market promotion, and C-end hardware product channel marketing [1] Technological Advancements - The company has achieved a breakthrough in deep reasoning models, being the only one to train a large model based on domestic computing power, with the 70B parameter Spark-X1-0420 model surpassing several international benchmarks [2] - The model supports over 130 languages, providing a fully autonomous and controllable large model base as an alternative for the global market [2] Industry Applications - In the education sector, the Spark X1 has been implemented in over 270 regions, with more than 200,000 teachers participating in teaching practices, achieving an 88.4% activity rate [2] - In the healthcare sector, the medical large model has shown superior performance in tasks such as general consultation and report interpretation, with a 52% year-on-year revenue growth in grassroots consultation [2] - The company collaborates with leading financial institutions and state-owned enterprises to develop large models for various industries [2] C-end Product Performance - Revenue from learning machines increased by 104% year-on-year, with a strong presence in high-end shopping malls and rapid expansion into lower-tier markets [3] - Other products, including translation machines and recording pens, continue to see iterative improvements, with significant user engagement in applications like the iFlytek input method [3]