As the global AI race accelerates, Huawei's 2025 annual report leaves little ambiguity: AI now sits at the core of its strategy. The 147-page filing references "AI" 421 times, an unusually explicit signal of strategic depth. The company is pursuing a "foundation first, expansion later" model, pairing heavy R&D with infrastructure buildout to scale its AI position.
"Artificial intelligence may be the last technological revolution in human history, destined to be turbulent yet transformative," said Huawei deputy and rotating chair Meng Wanzhou.
Meng described the current environment as "navigating through fog," arguing that direction outweighs speed. She framed AI as the defining opportunity of the next decade, while stressing strategic discipline: "Pulling back now is to strike harder later." The message signals a long-cycle positioning rather than near-term monetization.
AI strategy built on R&D scale
Huawei's AI commitment is anchored in sustained R&D investment. Spending reached CNY192.3 billion (approx. US$27.9 billion) in 2025, accounting for 21.8% of revenue, with cumulative investment exceeding CNY1.38 trillion over the past decade. More than half of employees are in R&D, reinforcing its view of technology as a structural moat rather than a cost center.
Investment is spread across the stack rather than focused on a single model, spanning chips, computing architecture, operating systems, and development platforms. Huawei is commercializing Ascend AI processors for inference, supported by a "cluster + supernode" design to scale data center compute. Modular architecture is also pushing AI into enterprise edge deployments.
The strategy positions Huawei as an AI infrastructure provider rather than an application platform. Its stated goal is to build a "silicon black soil" layer capable of supporting multiple AI applications, effectively anchoring value at the infrastructure level.
AI integration across compute and devices
Huawei is applying AI horizontally across its businesses, from ICT infrastructure to consumer devices and smart EVs. AI is positioned not as a standalone engine, but as a shared layer enhancing product competitiveness across segments.
The shift is most visible in devices. Huawei's HarmonyOS ecosystem has surpassed tens of millions of devices and more than ten million developers, crossing key scale thresholds.
Huawei has introduced AI features such as the Xiaoyi intelligent agent, positioning devices as proactive service assistants.
Meng said "AI on devices" is driving a new battle for user entry points, with device makers and model developers competing across hardware, ecosystems, and experience. Huawei's strength lies in hardware-software integration, but its position in applications and content still depends on ecosystem expansion.
AI ecosystem gaps versus platform peers
Huawei's ecosystem challenge is more complex than its hardware position. Despite building developer ecosystems around Ascend, Kunpeng, and Huawei Cloud, with millions of developers, structural gaps remain versus platform-based peers.
For example, Tencent's WeChat (Weixin) leverages massive messaging traffic to embed AI into daily use. Alibaba and ByteDance rely on e-commerce and cloud platforms to anchor enterprise services and data. All three control high-frequency entry points, allowing AI to convert directly into user engagement and revenue.
Huawei's ecosystem starts from hardware and system layers. While it benefits from integration and technological autonomy, it remains weaker in high-frequency applications and content distribution. This pushes the company toward a more open strategy to attract partners and fill application-layer gaps.
Huawei operates a foundational AI platform, while Tencent and Alibaba run application-driven ecosystems. The models are not mutually exclusive, but at the key battleground of entry points and data traffic, Huawei remains at a disadvantage.
The report frames Huawei's AI strategy as a capital-intensive, long-cycle industrial buildout: laying foundations in R&D and compute, then extending into devices and ecosystems. While Tencent, Alibaba, and ByteDance dominate traffic and applications, Huawei is approaching from the infrastructure layer, aiming to control the "production tools" of the AI era. This places its strategy firmly at the base of the value chain, where long-term leverage may ultimately accrue.
Article translated by Levi Li and edited by Jack Wu
