Key Event: Brilliance Tech Reports H1 2026 Financial Results
Brilliance Tech disclosed its unaudited semi-annual results on August 28, 2026, with the following core metrics:
- Reporting Period: Six months ended June 30, 2026
- Total Revenue: RMB 1.236 billion, up 1,997.6% year-on-year
- Gross Profit: RMB 527 million, up 2,708.5% year-on-year
- R&D Expenses: RMB 804.4 million, up 40.7% year-on-year
- Net Loss: RMB 377.2 million, down 76.4% year-on-year (significantly narrowed)
- Adjusted Loss (Non-IFRS): RMB 337.2 million, down 38.9% year-on-year
Commercial progress includes successful supplier certification from internet clients and commencement of bulk product deliveries, enabling operational leverage to amplify profitability gains.
Business Milestones: From Certification to Bulk Deliveries
The Brilliance™ series for training and inference has entered large-scale commercial deployment. Customer coverage has expanded notably, spanning top internet companies, AI large model developers, national AI compute platforms, AI data centers, telecom operators, and enterprises across AI solutions, manufacturing, energy, utilities, fintech, and education sectors.
Internet enterprises and cloud vendors represent high-barrier, high-volume demand segments. The company has now completed supplier onboarding for major internet clients and begun bulk deliveries, unlocking substantial room for future revenue growth.
Scenario-wise, Brilliance™ products have achieved规模化 deployment in large language model inference, autonomous driving, embodied intelligence, and multimodal generation (text-to-image, text-to-video, music generation). For training workloads, Brilliance Tech has partnered with several large model firms to deploy multi-thousand-chip clusters for multimodal model pre-training and reinforcement learning, completing end-to-end migration of training, fine-tuning, and inference workflows for commercial operations.
Critical Contradiction: Heavy R&D Investment Amid Explosive Revenue Growth
The财报 reveals a notable contradiction: despite ~20x revenue growth, R&D spending increased 40.7% to RMB 804.4 million—confirming sustained heavy investment in technological iteration.
Comparative growth rates:
- Revenue Growth: 1,997.6%
- Gross Profit Growth: 2,708.5% (gross margin reaches 42.6%)
- R&D Growth: 40.7% (absolute spend remains above RMB 800 million)
The rapid scale-up of high-margin business摊薄 fixed and部分 variable costs, amplified by R&D efficiency improvements, drove the 76.4% year-on-year net loss reduction. This confirms the company’s transition from “burning cash for R&D” toward “scalable profitability.”
Deployment Scenarios:千卡 Clusters Matching International Benchmarks
The reported千卡 cluster performance represents a pivotal milestone. For large model training, Brilliance Tech’s solution achieves equivalent accuracy to international alternatives while delivering significantly improved training speed.
This means the hardware ecosystem and software stack now demonstrably support mainstream global large model training tasks—laying the technical foundation for commercial revenue acceleration.
Practical Guidance: Who Should Engage with Brilliance™ Now?
- Well-suited for immediate evaluation: AI large model R&D teams, national/enterprise compute platform procurers, and cloud vendors seeking supply diversification or cost optimization—especially when existing solutions face supply risk or pricing pressure
- Recommended for continued observation: Teams with extreme edge inference latency requirements; smaller teams lacking hundreds-of-chips deployment experience (complexity assessment advised)
In Conclusion
Brilliance Tech’s H1 results showcase a clear commercialization pathway for Chinese AI chipmakers—from bulk delivery validation to revenue scale-up. With R&D and revenue growth accelerating in tandem, the true industry test lies in sustaining technical endurance and ecosystem conversion efficiency—the decisive factor for breaking foreign dominance.