ChinaBiz Briefing | CATL Mining Push, Huawei Regains Crown, Alibaba AI App Layer, MiniMax Agents
China’s tech and industrial ecosystem is moving deeper into a phase defined by resource control, AI commercialization and platform competition. CATL is expanding upstream to secure raw materials, while Alibaba and MiniMax are pushing into the AI application and agent layers. At the same time, smartphone competition is reshuffling, and efficiency breakthroughs continue to reshape the economics of AI development.
CATL Commits $4.35B to Build Global Mining Platform
CATL plans to invest roughly RMB 30 billion ($4.35 billion) to establish a global mining resource subsidiary, integrating its existing upstream assets and expanding into overseas projects. The new entity will cover exploration, extraction and processing of key battery materials such as lithium and nickel, forming a more vertically integrated supply chain.
Why it matters:
Battery competition is moving upstream into resource ownership. As EV and energy storage demand accelerates globally, securing stable and cost-effective raw materials is becoming a core strategic advantage rather than a supporting function.
Huawei Regains China Smartphone Lead as Market Rebalances
Huawei has returned to the top of China’s smartphone market, supported by product recovery and stronger domestic positioning, while Apple maintained momentum through stable pricing strategies. Xiaomi fell out of the top tier, reflecting increasing pressure in the mid-range segment and intensifying competition across price bands.
Why it matters:
China’s smartphone market is shifting from volume growth to competitive reshuffling. Brand strength, pricing discipline and ecosystem integration are becoming more important as overall demand stabilizes and differentiation becomes harder.
Alibaba Pushes Meoo No-Code Platform to Capture AI Application Layer
Alibaba’s ATH unit is promoting Meoo, a no-code platform that enables enterprises to build AI-powered applications without deep engineering resources. The tool focuses on simplifying deployment of generative AI into business workflows such as customer service, data analysis and internal automation.
Why it matters:
The center of competition in AI is moving from model development to application deployment. Platforms that lower the barrier to building AI applications are likely to capture more value as enterprises shift from experimentation to large-scale adoption.
RedNote Open-Sources RELAX to Improve Multimodal Training Efficiency
RedNote released its RELAX framework as open source, reporting up to 76% improvement in training efficiency for multimodal reinforcement learning models. The framework targets reductions in compute cost while maintaining performance across text, image and video tasks.
Why it matters:
Efficiency is becoming a defining factor in AI competition. Lower training and inference costs enable broader adoption and faster iteration, especially as multimodal models require significantly higher compute resources.
Insta360 Targets Niche Users With Limited Edition GO Ultra Launch
Insta360 launched a Pogačar-branded limited edition of its GO Ultra camera, priced at RMB 3,598 and aimed at sports and outdoor enthusiasts. The product combines compact design with high-performance imaging, leveraging athlete branding to reach a targeted user base.
Why it matters:
In a crowded hardware market, differentiation is increasingly driven by niche positioning and ecosystem alignment. Targeted products allow companies to maintain pricing power while building stronger connections with specific user segments.
MiniMax Launches MaxHermes to Advance Self-Evolving AI Agents
MiniMax introduced MaxHermes, a cloud-based AI agent designed to improve continuously through iterative learning and task feedback. The system aims to handle complex workflows autonomously, moving beyond static responses to adaptive execution.
Why it matters:
AI agents are evolving toward autonomous systems that can refine their own performance over time. This shift could redefine software models, as continuous learning replaces traditional update cycles.
What to watch next
Watch whether CATL’s upstream expansion triggers broader competition for global battery resources, and whether Alibaba can establish a leading position in the AI application layer. At the same time, improvements in training efficiency and agent autonomy will shape how quickly AI moves from experimentation to large-scale deployment.