AI "Rewrites" ZTE

On the evening of March 6th, ZTE Corporation released its annual financial report for 2025. The company achieved an annual operating revenue of 133.896 billion yuan, a year-on-year increase of 10.38%, and returned to the fast - growth track.

Behind the performance, the real changes are hidden in the structure. The company's computing power business revenue increased by about 150% year - on - year, and its proportion in the overall revenue rose to 24.6%. Among them, the revenue of servers and storage increased by more than 200% year - on - year, and the revenue of data center products increased by 50% year - on - year. The enterprise business achieved double - digit growth, reaching 37.222 billion yuan.

In the past year, the ICT industry has been talking about AI, and ZTE has presented a report on "fulfillment". From the financial report, it can be seen that the company has completed the original capital accumulation and technical foundation construction required for AI transformation. At this year's MWC Barcelona, the full - scenario AI product matrix outlined a commercial closed - loop from the "computing power base" to "terminal applications".

The other side of the coin is also worthy of attention. In 2025, ZTE's net profit attributable to the parent company was 5.62 billion yuan, a year - on - year decline of 33.3%, and the gross profit margin was also under pressure. The decline in profit was mainly affected by the dual impacts of the industry cycle switch and the change in business structure.

On the one hand, the capital expenditure of operators has entered a contraction period. The capital expenditure of the three major operators last year decreased by about 20% compared with the peak in 2023, and the investment in 5G decreased by more than 40%, which directly affected the company's revenue from traditional network equipment.

On the other hand, the prices of global storage chips have continued to rise, and the supply - demand imbalance of DRAM and NAND Flash has directly pushed up the hardware costs of computing power equipment. Mid - stream equipment manufacturers are facing the dual pressures of rising prices from upstream and strong bargaining power from downstream, and the industry is generally facing challenges in gross profit margin.

In the face of such a cycle, ZTE's choice is to increase R & D investment. The annual R & D expenses reached 22.76 billion yuan, accounting for about 17.0% of the revenue, and were invested in key technologies such as computing power nativization and seamless migration.

AI Nativization: Terminal Completes "Species Evolution"

At the MWC in Barcelona, a large number of domestic and foreign media and audiences crowded in front of a mobile phone to experience it. An experiencer said to the phone, "Help me book a car, a hotel, and then help me choose the best local restaurant and make a reservation." A few seconds later, the phone automatically opened the car - hailing app to set the destination, switched to the hotel platform to complete price comparison and reservation, and finally launched the review app to select a restaurant - all without manual operation, and the software ran automatically in the background.

This mobile phone is the Nubia M153, a collaboration between ZTE and Doubao under ByteDance. Borrowing the description of Ni Fei, the president of ZTE's Terminal Business Unit, behind this "self - driving" - like intelligent experience, multiple technologies such as end - side large models, operating system - level API opening, and privacy computing are integrated, which tests a technology company's control over software - hardware integration.

For a long time in the past, most of the AI mentioned by mobile phone manufacturers was in - app functions, such as AI beauty, AI translation, and AI photo retouching. However, the breakthrough of the Nubia M153 lies in that it turns AI from a "tool" into an "Agent". Users no longer need to learn how to operate apps. They just need to tell AI what they want, and the rest will be completed by the system - level intelligent agent.

The reason why the mobile phone has become the preferred carrier for AI intelligent agents is that it has the most sensors, the most frequent interaction scenarios, and the richest application ecosystem. It is the first entrance for everyone to access AI and the bridgehead for AI capabilities to sink from the cloud to users' fingertips. ZTE's decision to launch a system - level intelligent agent mobile phone at this point is based on the trend that AI will shift from "passive response" to "active service".

ZTE's layout of AI terminals is not limited to mobile phones. At the MWC booth, an AI pet named iMoochi also attracted attention. It has five personalities, yawns, gets hungry, is happy when stroked, and even has its own "life rhythm".

iMoochi targets the emotional companionship market, providing users with a "furry friend" that doesn't need feeding, won't get sick, and can provide emotional value 24/7. Currently, this product has five members with different personalities: Hopami, Mimiu, Cynomi, Mogogo, and Morin. The supporting APP can manage the device, view the companionship diary, and understand its "mood".

As the group of urban single - dwellers continues to expand and emotional connection becomes a scarce commodity, the emergence of iMoochi points to a judgment - the end - game of AI may not necessarily be a smarter tool, but also a warmer partner.

The AI mid - screen series products covering the home scenario play another role. By integrating entertainment, communication, security, and whole - house intelligent control capabilities, it becomes the center of home intelligence. In 2025, the sales volume of ZTE's smart mid - screens exceeded one million units, the cumulative shipments of FTTR products exceeded 30 million sets, and ZTE's Wi - Fi 7 home terminals ranked among the top in the industry. These figures mean that ZTE has enough user touchpoints to support the implementation of the AI ecosystem.

The three types of products, AI mobile phones, AI pets, and AI mid - screens, outline ZTE's AI map for its terminal business: with the mobile phone as the core entrance, the AI pet to expand the emotional companionship scenario, and the AI mid - screen to cover the home intelligent space. By connecting personal and home scenarios through different forms of terminals, ZTE attempts to establish an AI interaction system covering multiple terminals and spaces.

Over 150% Revenue Growth of Computing Power Products: The "TCO Logic" Behind

The most eye - catching figure in the financial report comes from computing power products: in 2025, the computing power achieved leap - forward growth, with the annual revenue increasing by about 150% year - on - year, and accounting for nearly 25% of the overall revenue.

The direct driving force for this growth is that ZTE has seized the opportunity of leading Internet companies and industry customers to increase investment in intelligent computing. The explosion of AI large models is reshaping the global demand for digital infrastructure - whether it is operator networks, cloud computing centers, or enterprise - level computing power facilities, they are all entering the "AI native" cycle.

The reason why customers firmly choose ZTE lies in the TCO (Total Cost of Ownership).

At the chip level, ZTE's self - developed chipset enables customized optimization at the device level, reducing dependence on high - cost external devices and controlling hardware costs from the source, thus compressing costs as much as possible during the period of rising storage prices.

At the algorithm level, ZTE's continuous optimization of key algorithms enables the same hardware to produce higher performance. The GigaMIMO technology uses AI integration to improve the experience of edge users several times. It doesn't rely on adding antennas but on the precise scheduling of channels by algorithms. The embedded AI Radio can reduce the power amplifier energy consumption by about 40%, which directly translates into savings on customers' electricity bills.

At the architecture level, ZTE's efficient orthogonal architecture super - node supports 32 to 128 GPUs in a single cabinet, and the "0 - cable" design inside the cabinet not only improves the integration level but also reduces the failure points and maintenance costs. Based on a single cabinet, a cluster scale from hundreds to tens of thousands of cards can be flexibly built, allowing customers not to pay for long - term needs at once but to expand capacity on demand and invest in installments.

At the delivery level, ZTE's service network covering more than 160 countries and regions means that customers don't need to rebuild local support teams when deploying globally. In 2025, ZTE's international market revenue reached 44.16 billion yuan, a year - on - year increase of 12.4%. This delivery network itself is a cost advantage, and the construction of intelligent computing centers for overseas operators and the global deployment of multinational enterprises can all reuse the existing channels and service systems.

At the standard level, ZTE's in - depth participation in the formulation of standards such as high - speed interconnection enables it to maintain compatibility in the new - generation technology iteration and reduce customers' migration costs. In the field of Wi - Fi 7, ZTE has more than 110 global core patents, and its number of IEEE 802.11bn proposals ranks among the top four globally.

This set of capabilities has been verified at the product level. The AIR MAX solution uses AI to empower the wireless network and the core network, realizing intelligent network operation and maintenance and improving traffic value. Currently, AIR Engine has been deployed in more than 150,000 sites globally. In the TMF's self - intelligent network rating, its intelligent capabilities in wireless energy - saving and fault management have reached the leading level in the industry.

As of now, ZTE has built more than 500 green data centers globally, launched ten - thousand - card - level intelligent computing centers, and its intelligent computing servers have entered the core scenarios of leading domestic enterprises in the Internet, telecommunications, finance, and power industries. Together with more than a thousand partners, it has created multiple intelligent benchmarks in fields such as industrial manufacturing and urban governance.

Behind the explosive growth of computing power is ZTE's "TCO logic". Its accumulation in the five aspects of chips, algorithms, architecture, delivery, and standards precisely addresses the cost proposition that customers care about the most.

A Re - evaluation of "Full - Stack Competitiveness"

After "switching tracks" to AI, how should the market re - evaluate ZTE? In the short term, it is a fact that the profitability is under pressure. But in the long term, the real value anchor lies in: when the evolution of AI technology enters a new stage of "system - ability competition", what irreplaceable moats does ZTE have?

The first moat is the end - to - end integration ability. Compared with pure software companies, ZTE's AI advantage lies in its depth of "reaching down to the chip base and sensing scenario applications upward". From base stations to mobile phones, from the cloud to terminals, from networks to chips - this reconstruction of the "underlying logic" means that when AI applications require software - hardware integration optimization, ZTE can achieve full - link optimization from the underlying instruction set to the upper - layer business logic.

The GigaMIMO technology uses AI integration to improve the experience of edge users several times, proving that ZTE is not building an isolated AI but a digital - intelligent base that "can sense, think, and self - operate". The embedded AI Radio reduces energy consumption by nearly 40%. It doesn't rely on more expensive components but on the extreme optimization of power amplifier scheduling by AI.

The second moat is the global implementation ability. ZTE's business covers more than 160 countries and regions globally, and it has accumulated rich experience in going global and local service capabilities. When AI applications enter the stage of global competition, this "implementation ability" is becoming a scarce resource. Whether it is deploying intelligent computing centers for overseas operators or promoting AI terminals to the global market, ZTE's global network is an important lever for its AI strategy.

The communication network is the foundation of AI construction. "Computing power without a network is an information island." What ZTE is doing is to transform the coverage advantage of the communication network into the delivery advantage of AI services.

The third moat is the extreme pursuit of TCO. The current competition in the AI industry has shifted from "can it be done" to "can it be used well". Customers are no longer paying for peak computing power but for the comprehensive benefits over the entire life cycle. The five major capabilities of chips, algorithms, architecture, delivery, and standards that ZTE has accumulated over more than forty years all point to the same goal: to turn AI from a "luxury" into "infrastructure", enabling customers to use it affordably, effectively, and sustainably.

At present, ZTE is at a critical stage of transforming old and new driving forces. The traditional network business provides a stable cash flow, the AI computing power business contributes high - speed growth, while the profitability is under structural pressure during the transformation period. This is a stage that every enterprise trying to cross the cycle must go through.

However, the market may need to view this pressure from a different perspective: when a company is willing to sacrifice short - term profits for long - term system - ability construction, this investment itself is a bet on the future competition pattern.