ZTE has launched its M8000 intelligent computing micro-module data center and a new-generation integrated AIDC IT POD solution (ZXED) for global markets. The offering targets rapid, modular deployment of AI computing capacity, integrating power, cooling and IT infrastructure into pre-engineered pods suited to high-density GPU workloads.
Overview
ZTE Corporation has launched the ZXEDC M8000 Intelligent Computing Micro-Module, a new-generation integrated AIDC IT POD solution, at Data Center World Asia. The company positions the M8000 as a response to the structural pressures now shaping AI data centers: soaring rack power density, uncertain business iteration, and the shift from air cooling to liquid cooling. Rather than treating power, cooling, racks, cabling and management as separate procurement packages, ZTE has bundled them into a pre-engineered micro-module intended for rapid, modular deployment of AI computing capacity.
The launch matters because AI infrastructure is increasingly constrained less by server availability than by the physical environment surrounding those servers. High-density GPU workloads require power distribution, heat rejection and operational visibility to be designed together. ZTE’s proposition is that a standardized POD can shorten delivery timelines while preserving enough flexibility to support multiple generations of IT hardware.
Architecture and Technical Composition
The M8000 integrates an enclosed cold aisle, liquid-cooled racks, network racks, a coldplate liquid cooling CDU, secondary-side piping, high-voltage DC, intelligent busbars, a thin-sheet air wall, monitoring, lighting, cabling and fire linkage into one unified solution. This is a system-level design rather than a collection of components. The inclusion of network racks and fire linkage is particularly relevant for AI clusters, where compute, switching and safety systems must be coordinated inside the same high-density footprint.
ZTE says the solution supports rack power from 20–60kW per rack, with a peak of 250kW per rack. Racks are decoupled from the frame and can accommodate full racks and special-form-factor racks. The architecture also allows on-site air-to-liquid cooling retrofits at later stages. Standardized, productized components are intended to enable fast on-site assembly and shorter delivery timelines.
Power and Cooling Economics
The M8000 adopts an air-and-liquid shared-source architecture. Its CDU pairs with thin-sheet air walls using dual cooling sources, allowing a dynamically adjustable air/liquid cooling ratio. This is significant because AI data centers rarely move to liquid cooling in a single step; many operators need a hybrid environment that can support both conventional and high-density liquid-cooled equipment.
ZTE claims the solution maximizes free cooling and delivers a PUE as low as 1.15 across all regions. It also combines 800V HVDC with intelligent busbars for flexible in-row power supply, achieving up to 98% power supply efficiency. Those figures should be read as vendor specifications, but they point to the central design goal: reducing conversion losses and cooling overhead while keeping power distribution modular enough to change as rack densities evolve.
Safety, Certification and Operations
Liquid cooling introduces new operational risks, particularly leakage around CDUs, racks, servers and switches. ZTE says the M8000 is the industry’s first solution to obtain UL system-level leak-proof certification. Full-chain leak protection covers the CDU, liquid-cooled racks, liquid-cooled servers and switches, and the system includes leak detection and automatic liquid replenishment to mitigate leakage risks.
On the operations side, the M8000 provides unified control of air and liquid cooling, integrated cooling and power management, and automatic optimization via AI algorithms. ZTE says these capabilities reduce cooling CLF energy consumption by more than 10%. The iDCIM management platform provides end-to-end visibility of power and cooling systems, simplifying operations and maintenance for intelligent computing infrastructure.
Deployment Model and Reference Case
ZTE’s customer benefits center on investment protection, reliability and operating expense. The company says one architecture supports three generations of IT hardware upgrades, avoiding large-scale retrofits during computing capacity expansion. Full-chain liquid cooling safety protection is intended to deliver stable operating environments for high-power AI clusters, while free cooling and AI optimization are positioned to lower energy consumption and PUE.
The reference case illustrates the deployment model. In 2025, ZTE won the bid for 666 IT modules for a computing center project in the Yangtze River Delta for a leading Chinese internet enterprise. According to ZTE, 345 sets of M8000 IT PODs have been successfully delivered. The project uses single-row racks plus single-row thin-sheet air walls, factory prefabrication and skid-mounted transportation for on-site hoisting. ZTE says it took only one month from PO issuance to site arrival, supporting rapid rollout of large-scale intelligent computing clusters.
Market Context and Industry Impact
The M8000 arrives as data center operators across Asia Pacific and globally confront a difficult transition. Air cooling remains prevalent, but high-density AI racks are pushing conventional designs beyond their practical limits. Liquid cooling is becoming more viable, yet it requires new supply chains, safety practices and operations skills. ZTE’s emphasis on a hybrid air/liquid architecture, factory prefabrication and system-level certification reflects that transition rather than assuming an immediate all-liquid future.
The competitive significance is twofold. First, ZTE is packaging AI data center infrastructure as a productized POD, which could make high-density capacity faster to order and deploy. Second, by integrating power, cooling, cabling and management, the company is competing in the broader AIDC infrastructure layer, not just in individual cooling or power components.
Implications for AI Data Centers and Future Outlook
For AI data center planners, the M8000 highlights three priorities: evolvability, thermal safety and energy efficiency. The ability to support 20–60kW racks today while accommodating a 250kW peak suggests a design intended to absorb uncertainty in AI hardware roadmaps. The UL system-level leak-proof certification and full-chain leak protection address a key barrier to liquid cooling adoption. The reported PUE of 1.15 and 98% power supply efficiency indicate where vendors are focusing differentiation.
ZTE says the solution is built for high-density air-liquid hybrid AIDC scenarios and is aimed at customers across Asia Pacific and globally. The next test will be execution at scale: whether prefabricated PODs can be delivered, hoisted and commissioned repeatedly across regions while maintaining safety and performance. The Yangtze River Delta reference case provides an early indication, but broader adoption will depend on how quickly operators move from evaluating hybrid designs to standardizing them in new AI capacity plans.