Time：2026-09-10
Kuala Lumpur, Malaysia, September 9, 2026 – ZTE Corporation (0763.HK / 000063.SZ), a global leading provider of integrated information and communication technology solutions, successfully hosted the 2026 ZTE Global Intelligent Computing Ecosystem Congress, under the theme "Infinite Intelligence, Thriving Together". Centered on the co-evolution of intelligent computing infrastructure and AI models, the congress brought together representatives from global chip and computing vendors, cloud and internet companies, as well as academic experts. Participants engaged in in-depth discussions on cutting-edge topics including computing evolution in the era of large models, rising demands for Agents and Tokens, inference efficiency enhancement, next-generation cloud infrastructure, SuperPOD architecture, and open intelligent computing ecosystems, jointly exploring the technological trends and industrial practices of future intelligent computing infrastructure.
AI is currently accelerating its transition from model capability evolution to large-scale deployment. Concurrently, computing power demands are shifting from training to inference, and expanding from standalone models to new-form applications such as Agents. In this process, Tokens are increasingly becoming a crucial metric for measuring the efficiency and costs of AI applications. Therefore, how to continuously improve computing power efficiency and reduce token costs has become key to advance AI from technological innovation to democratize computing power at scale. Centered on this core issue, ZTE and its global industry partners are exploring multiple dimensions-including computing infrastructure, chips and software, cloud platforms, intelligent agents, and industry applications to jointly forge new pathways for AI industry development.
AI in the Token Era: Accelerating Infrastructure Restructuring
Zhang Wanchun, Senior Vice President of ZTE, stated in his opening speech that AI is entering a phase driven by the Tokenomics. Token has become the core unit of measurement for this new economic model. AI technology architectures, computing models, and business paradigms are being reshaped at an accelerating pace. ZTE follows the paths of open decoupling, diverse computing, and extreme synergy to build end-to-end intelligent computing solutions. With continuous deployments across industries such as telecommunications, finance, and healthcare, ZTE is committed to working with global partners to build an open and win-win intelligent computing ecosystem. With the continuous development of AI applications, computing infrastructure must further adapt to surging inference demands. A more open, diverse, and efficient infrastructure is essential to provide solid supports for the large-scale deployment of AI.
If Tokens are becoming as the new "unit of measurement" in the AI era, then continuously reducing Token costs has become a core challenge that must be addressed for AI to achieve large-scale applications.
To address that, Yang Rui, Technical Director of Cloud & Core Network Product of ZTE, stated that the prosperity of AI applications and the reduction of Token costs are mutually reinforcing drivers of a virtuous "flywheel" cycle. The flourishing application ecosystem generates massive inference demands, which in turn compel continuous cost optimization. In turn the decline in Token costs further lowers the barriers to innovation, unlocks more use cases.
Centered on this trend, ZTE is committed to unleashing full-stack innovation through a flexible and open architecture, and achieving ultimate performance optimization via end-to-end system-level synergy. This further enhances inference efficiency and optimizes computing costs, delivering industry-leading intelligent computing infrastructure solutions, enabling every developer to embrace the infinite possibilities of the AI era at a more optimal Token cost.
From Standalone Computing to Systematic Engineering: Industrial Synergy as the Key to Infrastructure Evolution
As AI computing demands continue to surge, reducing Token costs is no longer just about optimizing a single chip or device. Instead, it has evolved into a systematic engineering effort that encompasses chips, software, clusters, networks, and service applications.
Guoliang Sun, Senior Vice President of MetaX, said, "Computing is far more than hardware. It's a systematic endeavor involving chips, software stacks, AI clusters, and industry applications. MetaX is committed to developing proprietary GPU architectures and the MXMACA software stack, connecting computing infrastructure with real-world business needs and advancing AI computing from availability to usability".
He also highlighted that MetaX has established a "1+6+X" ecosystem strategy to accelerate the adoption of AI computing across industries including finance, healthcare, energy, education & research, transportation, and digital entertainment, while exploring emerging frontiers such as embodied intelligence and the low-altitude economy. MetaX empowers industries with stable, reliable, and efficient computing infrastructure, unlocking new opportunities for the digital economy.
From GPU chips and software stacks to the overall architecture of AI infrastructure, the upstream and downstream players of the industry are forging closer collaborative ties. Tan, Lai Teng, Country Sales Manager of Intel Malaysia, mentioned that Intel powers the reshaping of AI infrastructure, serving as the all-round powerhouse for tomorrow-ready intelligent computing centers. Companies should choose optimized infrastructure for sustained AI evolution. Intel Xeon delivers significant benefits for Traditional AI, Generative AI, Agentic AI, and Frontier AI. Furthermore, Intel will partner with its ecosystem to co-create tomorrow-ready intelligent computing centers.
Convergence of Cloud, Data, and Local Computing: AI Core Capabilities Accelerate Industrial Adoption
As computing infrastructure continues to evolve, the industrialization of AI is further entering a phase of synergistic development across cloud, data, computing power, and models.
Jayson Li, Director of Apsara Stack Product Solutions of Alibaba Cloud Intelligence, attended the event and pointed out that demands for compute and tokens continue to grow in the AI era, while data and AI sovereignty are becoming long-term priorities for countries and industries worldwide. Through close collaboration, Alibaba Cloud Apsara Stack and ZTE are supporting two development tracks on a unified cloud foundation. On one track, they help customers turn GPU assets into sustainable businesses through compute, token and agent ecosystem operations. On the other, they enable continuously evolving Sovereign AI capabilities built on local data, local compute and local models—allowing AI to create business value while remaining rooted in and serving local markets.
If computing power, cloud, and data constitute the foundational infrastructure for AI development, then intelligent agents are emerging as a crucial gateway for AI to further integrate into enterprise production workflows.
Garry Sien, Chief Innovation and Solutions Officer, International Business of Ant Digital Technologies, shared his views on the next phase of enterprise AI. He stated that agent-driven business would take off in the next 6 to 12 months. AI is getting better fast - from chatting to searching, coding, and now actually getting things done. Enterprises don't want another AI tool, they want a system.
Ant DT's answer is the Agent Factory — its Agentar 2.0 platform ships with nearly 200 role-based digital expert templates and hundreds of subscribable skill modules, enabling businesses to assemble AI teams like a production line and put them to work on day one.
Meanwhile, the continuous evolution of AI capabilities is also accelerating the formation of regional innovation ecosystems. Dr. Sara Huang, COO of Hong Kong Generative AI Research and Development Centre (HKGAI), stated that as we enter the era of sovereign AI, HKGAI is committed to leveraging Hong Kong's unique strengths as both a "Super Connector" and a "Super Value‑adder". Drawing on its core product portfolio, including large models and AI agents, together with accumulated practical experience in smart government services, the Centre empowers enterprises to build digitally‑intelligent organisations. She added, "We look forward to deep collaboration with ZTE and our Malaysian partners to jointly foster an open, inclusive, and universally beneficial global AI ecosystem".
An Open Ecosystem Unites Industry Forces to Unleash AI's True Potential
From Token costs to computing architectures, from GPU chips and software stacks to cloud, data and local computing, and from intelligent agents to various industries, this conference has highlighted a clear trend: the AI industry is transitioning from isolated technological innovations to collaborative innovation across infrastructure, hardware and software, models, applications and ecosystems.
For the AI industry, computing efficiency determines how far AI applications can go, while an open ecosystem dictates the ultimate reach of AI's value. Only through open collaboration across the upstream and downstream value chain, and by continuously driving synergy among computing power, networks, cloud, models and applications, can we further lower the barriers to AI innovation and application, enabling AI capabilities to truly permeate more industries and service scenarios.
ZTE Global Intelligent Computing Ecosystem Congress, driven by technological innovation and connected through ecosystem collaboration, fully embraces the development vision of "Infinite Intelligence, Thriving Together". ZTE showcases its key technological achievements in intelligent computing and engaged in in-depth exchanges with global industry partners on intelligent computing infrastructure, AI models, and application ecosystems. This effort is accelerating the evolution of computing infrastructure from isolated capabilities to systemic synergy, and from technological supply to tangible industry value.
Looking ahead, ZTE will continue to build upon open and decoupled intelligent computing infrastructure, using ecosystem collaboration as a vital link. This will accelerate the integration of chips, computing power, networks, cloud, models and applications. Through continuous open cooperation and industrial collaboration, ZTE aims to further enhance AI infrastructure efficiency and reduce Token costs. Ultimately, ZTE will drive AI capabilities from the infrastructure level to various industries, enabling more enterprises to access AI at optimal costs, and jointly build an open, inclusive, accessible and sustainable intelligent ecosystem.
