An Interview with Li Zhifeng, CTO of MAAS:
The Next Five Years of AI Computing Infrastructure
A conversation between Gina Xu, CEO of MAGPIENEWS, and Li Zhifeng, CTO of MAAS—from the approval of the “Lingyan” large language model to distributed container-based computing centers, enterprise-level security governance, and deep integration with the real economy: an in-depth discussion on the technical roadmap and commercialization of AI computing infrastructure.
Guests for This Discussion
A technology leader with deep expertise in computing power and storage, and a Hong Kong-based consulting firm specializing in data-driven IR/PR services—a dialogue bridging technology and capital.
Zhifeng Li
CTO of MAAS (Maase Inc). Ph.D. in Physics from the University of Vienna; Deputy Director of the Computing Power and Storage Research Center at the China Investment Association. He has long focused on the technical architecture and engineering implementation of AI computing infrastructure, and has led the development of the “Lingyan Architecture” and the technical roadmap for distributed container-based computing centers.
Gina
CEO of MAGPIENEWS. Specializes in providing IR, PR, and IPO consulting services to publicly traded companies, employing data-driven strategies, engaging in long-term dialogue with leaders in the tech industry, and bridging the gap between technology and capital markets.
A Six-Minute Summary of This Conversation
Don't have time to watch the full video? Here are six key insights Li Zhifeng shared during the discussion.
The Technological Boom in Large Models
AI is currently experiencing a major technological boom. MAAS’s in-house “Lingyan” large language model has officially been approved, and its general-purpose language model is now ready for use.
Distributed Container-Based Computing Center
Traditional centralized data centers take at least a year to build and are very costly; distributed container-based computing nodes can be deployed quickly and more efficiently, enabling services to reach customers faster.
Enterprise-Level Security Governance
Treating “security” as a hallmark of AI—the synergy between enterprise-level security governance and enterprise-level implementation systems—is an unavoidable challenge for computing infrastructure.
Integration and Optimization of Vertical Models
Based on the “Lingyan Architecture,” model optimization will focus precisely on two key areas: the integration of general-purpose capabilities with vertical-specific models, enabling large models to truly adapt to industry-specific scenarios.
Deep Integration with the Real Economy
The true value of AI ultimately lies in its deep integration with the real economy—computing power is not the goal; rather, it is the genuine improvement in industrial efficiency.
Commercialization Is Accelerating
Three major orders in a row have validated the technical approach, and over the next three to five years, MAAS aims to become a leading force in the AI computing infrastructure sector.
Memorable Quotes
Excerpt from the original dialogue, in both Chinese and English.
“It’s the explosion of AI from technology.”
“就是 AI 从技术大爆发。”
—Zhifeng Li , on the Current Stage of the AI Industry“It is this enterprise-level security governance.”
“是这个企业级的安全治理。”
—Zhifeng Li , on the “Safe AI” Label“The high efficiency of distributed computing power nodes.”
“分布式算力节点的高效率。”
—— Zhifeng Li, on Distributed Container-Based Computing Centers“We hope that MAAS can become a …”
“我们希望 MAAS 能够成为一个……”
—— Zhifeng Li, on the Company’s Vision and Plans for the Next Three to Five YearsWant to join the next conversation?
MAGPIENEWS continues its dialogue with leaders in the tech industry, bridging the gap between technology and capital markets. If your company has a technology story worth sharing, please feel free to contact us.
Contact us