PHANCY (06682) announced its financial results for the first quarter of 2026. The group achieved operating revenue of approximately RMB 1.458 billion, representing a year-on-year increase of 35.4%, with a gross profit margin of 35.1%.
The company's performance this quarter was driven by rapidly growing market demand and a significant increase in client Token calls, with all business segments expanding steadily. Against the backdrop of a relatively tight global computing power supply and enterprises' rising requirements for the security, stability, and data protection of AI systems, there has been continuous growth in demand for enterprise-exclusive AI deployments. Concurrently, the group's enterprise-grade AI platform, designed with a unified core architecture, supports both API call scenarios and enterprise-exclusive deployment needs, thereby enhancing overall application efficiency and resource utilization.
Furthermore, the supply chain management system established by the group over the years has increased its available computing power resources by over 200%. This has effectively alleviated supply pressure caused by rising Token demand and consistently ensured service delivery capability for clients. The group's API Token call volume in Q1 2026 was nearly six times that of Q1 2025 and exceeded the total call volume for the full year 2025 by approximately 40%.
The Agentic AI business also achieved substantial development. By the end of Q1 2026, the number of Agentic AI orders in hand had increased by 99% compared to the end of 2025.
**Core Business Progress: AI Technology Iteration Drives Operational Efficiency** By continuously deepening the application of digital employees and empowering business units with AI, the group has significantly shortened the cycle from product R&D to launch and commercialization. This contributes to improved overall operational efficiency and customer satisfaction. As of the date of this announcement, the XIN CHUANG Model Box has launched over 70,000 AI models adapted and optimized for domestic chip environments, far exceeding the group's previous expectations.
In May, the group launched PhanthyMovie, a professional-grade AI video generation platform focused on enhancing the professionalism, controllability, and stability of video creation, aiding the industry in achieving standardized, large-scale content production. Leveraging its outstanding technical capabilities and mature application proficiency, the product broke conventional industry commercialization timelines. Within just a few days, it entered into a long-term cooperation agreement with Huanxi Media Group Limited, involving Token usage cooperation worth approximately USD 200 million. The two parties will also jointly develop a new AI film and television content production platform, deepening their engagement in the AI cultural creation sector.
**Core Products Align with Industry Policy Direction** Since May 2026, relevant AI policies in China have sent positive signals regarding computing power infrastructure development, data element supply, and open-source governance. The group's core products, such as HAMi vGPU and the XIN CHUANG Model Box, align with these policy directions and industry development trends.
Regarding computing power resource allocation, relevant policies emphasize cross-regional coordination and improving the accessibility of computing power. HAMi vGPU, with its capabilities for unified scheduling and fine-grained resource partitioning, helps improve computing power resource utilization and optimize data center energy efficiency.
In terms of data and model governance, policies promote the development of high-quality datasets and compliant management. The XIN CHUANG Model Box supports multi-model adaptation and tuning, and integrates features for data provenance and security authentication, helping to reduce corporate compliance risks.
Additionally, HAMi vGPU supports unified management of multiple chips, which can improve the utilization efficiency of individual cards. The XIN CHUANG Model Box utilizes the EngineX engine to achieve batch adaptation of domestic chips and models, helping to address compatibility issues. It also tunes key models to enhance the efficiency of Token output.
Overall, through coordinated deployment at the computing power and model layers, the group has constructed an integrated "computing power-model" solution. This provides technical support for key issues such as resource utilization, data supply, security compliance, and domestic substitution, while also capitalizing on development opportunities presented by relevant policies.
Comments