新紫光集团与阿里云达成AI全栈合作

On September 21, New Tsinghua Unigroup and Alibaba Cloud signed a full-stack AI cooperation agreement, marking a major industry milestone of powerhouse partnership in China’s technology sector, with direct implications for future upgrades to intelligent services accessible to the general public. Key takeaways are as follows:

1. Core strengths of the two parties: Rooted in Tsinghua University and originally founded as Tsinghua Science & Technology Development Corporation, New Tsinghua Unigroup has operated in the technology sector for 38 years, with core businesses spanning semiconductor chips, ICT digital infrastructure, and artificial intelligence, serving over 100 countries and regions and generating annual revenue exceeding RMB 200 billion. Alibaba Cloud brings mature cloud computing services, Qwen large language model (LLM) capabilities, and a global service network.

2. Existing foundation for cooperation: Alibaba Cloud’s proprietary cloud AI Stack all-in-one training and inference platform has already been deployed in New Tsinghua Unigroup’s data centers in Shanghai, Shenzhen and other locations to support LLM training and inference. Combined with the Qoder platform and Qwen office productivity tools, the system has established new efficiency paradigms for intelligent office work and software R&D at New Tsinghua Unigroup.

3. Future cooperation will cover intelligent office upgrades, AI application deployment in key industries, and expansion of overseas technology services. The two parties will also explore cutting-edge technologies including 3D DRAM, near-memory computing, and on-device intelligence, with user-friendly AI services set to roll out to the market progressively.

The full-stack AI cooperation between New Tsinghua Unigroup and Alibaba Cloud provides a benchmark reference for technology brands seeking to build technological moats, expand user markets, and advance global expansion. Key actionable insights are as follows:

1. Approach to building technology-centric brand perception: Brands can fill capability gaps through deep partnerships with leading technology players. For example, New Tsinghua Unigroup leverages Alibaba Cloud’s stable, secure cloud computing infrastructure and Qwen LLM capabilities to reinforce its professional brand positioning in semiconductors, ICT infrastructure, and AI applications. The two parties will also jointly define requirements for AI cloud computing chips and advance R&D in cutting-edge areas such as 3D DRAM and near-memory computing to strengthen core technological barriers.

2. Reference pathways for product implementation: Brands can pursue Token operation, model fine-tuning, and toolset development to deeply integrate technological capabilities with internal data, business processes, and existing systems, building intelligent office and industry AI products aligned with user demand to capture opportunities in large-scale AI commercialization.

3. Global expansion methodology: Brands can leverage partners’ global node layouts and overseas compliance experience to co-develop intelligent computing and industry solutions for international markets, reducing costs associated with channel construction and compliance adaptation for overseas expansion.

The full-stack AI cooperation between New Tsinghua Unigroup and Alibaba Cloud signals multiple growth opportunities in the enterprise AI service and intelligent computing sectors, offering clear business prospects and actionable operational insights for sellers across all categories.

1. High-potential growth tracks: The partnership covers intelligent office tools, large-scale industry AI applications, overseas intelligent computing solutions, and on-device intelligent products—all high-growth segments in the technology service and smart hardware sectors. Sellers in related categories can position early on LLM-compatible office peripherals, smart hardware, and supporting industry service products to capture upcoming market demand.

2. Replicable cooperative operational models: Entities with different resource endowments can adopt deeply integrated, complementary partnership models. For example, sellers with physical industry resources and offline channels can collaborate with partners possessing cloud service and LLM capabilities to co-develop products, leveraging the latter’s global resources and compliance experience to reduce trial-and-error costs in overseas expansion.

3. Operational risk notes: When entering AI-related tracks, avoid hollow concept speculation. Prioritize practical products that deeply integrate with users’ existing business systems and actual usage needs, and reduce investment in concept-driven products with no tangible value.

The implementation pathway of the full-stack AI cooperation between New Tsinghua Unigroup and Alibaba Cloud provides clear guidance for manufacturing factories seeking to secure supporting orders, advance digital and intelligent transformation, and expand global business.

1. Accessible production order opportunities: The two parties will jointly advance the implementation of AI cloud computing chip solutions, and explore technology directions including 3D DRAM, near-memory computing, and on-device intelligence. Factories can proactively connect with demand for hardware processing, component manufacturing, and supporting equipment production across relevant industry chains to win orders for intelligent computing infrastructure and on-device smart hardware.

2. Actionable methods for digital-intelligent transformation: Factories can reference New Tsinghua Unigroup’s AI implementation experience by first deploying proprietary cloud AI training and inference platforms adapted to their own operations. Combined with LLM-powered office and R&D tools, this integration can deeply embed AI capabilities into production data, R&D processes, and existing manufacturing systems to improve efficiency across office work, R&D, and production.

3. Global business expansion opportunities: Factories can join the industrial cooperation ecosystem built by the two parties, relying on the ecosystem’s overseas channels and compliance experience to provide hardware production support for overseas intelligent computing projects, unlocking growth in B2B orders from international markets.

The full-stack AI cooperation between New Tsinghua Unigroup and Alibaba Cloud clearly demonstrates real customer demand, technology iteration directions, and high-potential service tracks in current industrial AI implementation, offering clear reference value for all types of B2B service providers.

1. Industry trend assessment: Enterprise AI applications have moved beyond the proof-of-concept stage to the phase of deep, integrated implementation. Core customer demand is no longer limited to simple LLM API access, but integrated AI solutions aligned with their own data assets, business processes, and existing IT systems. Enterprises with global operations also face common pain points related to overseas compliance and supporting services for cross-border intelligent computing.

2. High-priority service directions: Service providers can build capabilities around Token operation, model fine-tuning, AI toolset development, computing-power and power coordination, on-device intelligence adaptation, and supporting services for cutting-edge computing technologies, to meet customer demand for frontier technology exploration and AI application implementation.

3. Reference service implementation pathway: Service providers can collaborate with cloud vendors and real-economy industry players to build cooperative ecosystems, delivering full-stack services spanning underlying computing support, mid-tier model optimization, and upper-tier industry application deployment, to address customer pain points such as insufficient in-house technology and poor scenario fit during AI implementation.

The full-stack AI cooperation case between New Tsinghua Unigroup and Alibaba Cloud clarifies the core demands of large industrial clients for cloud service platforms and industrial service platforms, while providing a reference framework for platform ecosystem operation.

1. Core customer needs platforms must address: Demand from large industrial clients has evolved from simple cloud resource leasing to integrated delivery of stable, secure, and efficient cloud infrastructure, LLM capabilities, global node layouts, and overseas compliance experience. Platforms need to support clients across the full value chain: frontier technology exploration, internal efficiency improvement, industry application implementation, and overseas market expansion.

2. Actionable merchant recruitment and operation directions: Platforms can open ecosystem cooperation interfaces around their core capabilities, recruiting diverse ecosystem partners including semiconductor R&D teams, AI application developers, industry solution providers, and overseas service supporting vendors to jointly build full-stack AI service capabilities and better serve real-economy industrial clients.

3. Operational pitfalls to avoid: When developing AI-related businesses, platforms should avoid vague concept-driven marketing. All service design must be closely tied to real customer needs: business process integration, tangible efficiency improvements, and practical technology R&D implementation, while reducing investment in conceptualized services that deliver no actual value.

The full-stack AI cooperation between New Tsinghua Unigroup and Alibaba Cloud is a landmark case of deep integration between China’s AI technology and hard technology industries, carrying high industrial research value. Core research-worthy dimensions are as follows:

1. New trends in industrial development: Domestic AI industrial implementation has entered a stage of full-stack, deeply bound cooperation between cloud vendors and large technology industrial groups. Cooperation scope has expanded beyond early-stage basic computing resource deployment to cover the full value chain: joint definition of chip requirements, collaborative exploration of cutting-edge storage and computing technologies, deep integration of LLMs with business processes, joint expansion of overseas markets, and layout of computing-power/power coordination and on-device intelligence, marking a significant increase in cooperation depth.

2. Traceable new business models: This deeply integrated model—where cloud vendors provide cloud infrastructure and LLM capabilities, while industrial groups open up application scenarios and industrial resources—serves as a typical research sample for large-scale AI deployment in key industries and the validation of closed-loop commercialization.

3. Implications for industrial development: Providers of computing power and LLM services need to strengthen global node layouts and compliance service capabilities to meet the common overseas expansion needs of the technology industry. Future competition in the AI implementation space will center on joint technology R&D capabilities, depth of scenario integration, and level of ecosystem collaboration.