On March 16, 2026, Alibaba officially announced the establishment of the Alibaba Token Hub (ATH) Business Group, led personally by Group CEO Eddie Wu. With the core strategy of 'creating Tokens, distributing Tokens, and applying Tokens,' it integrates five major sectors: Tongyi Labs, MaaS Business Line, Qianwen Business Unit, Wukong Business Unit, and AI Innovation Business Unit, completing an AI closed loop from foundational model R&D to full-scenario applications for consumers (C-end) and businesses (B-end).

This is not just an organizational upgrade but also Alibaba's bet on the underlying logic of future commerce in the AI Agent era.

'Centralized Collaboration'

In an internal letter, Eddie Wu clearly judged that we are now on the eve of AGI explosion, where AI Agents will become the core interaction carriers in the digital world, and Tokens will be the core fuel supporting agent operations.

From the development trajectory of Alibaba's AI business, this integration marks a shift from 'decentralized competition' to 'centralized collaboration.'

Over the past few years, Alibaba's AI business has been fragmented, with Qianwen, Lingguang, Kuake, and others belonging to different systems, leading to data silos, technology non-sharing, and severe internal duplication of efforts. Even though all were built on the Tongyi large model, they struggled to form synergy.

Now, with ATH's integration, each sector has a clear role: Tongyi Labs handles the technological foundation, MaaS drives ecosystem output, Qianwen focuses on the C-end market, and the newly unveiled Wukong Business Unit directly targets B-end AI-native work platforms. Its goal is to deeply embed large model capabilities into real enterprise workflows, leveraging DingTalk's over 40 million paying enterprise users to create an enterprise-grade gateway for the AI era.

Financially, Alibaba is in a critical phase of AI investment. In FY2025, the group's revenue neared one trillion yuan, with net profit surging by 77%. By mid-FY2026, revenue reached 495.447 billion yuan, with net profit at 62.994 billion yuan. Among core business segments, Alibaba Cloud's AI-related revenue has seen triple-digit growth for nine consecutive quarters.

Meanwhile, Eddie Wu announced plans to invest over 380 billion yuan in cloud and AI infrastructure over the next three years. Under this massive investment, the establishment of ATH essentially aims to translate technological input into actionable, measurable, and scalable commercial returns.

'Utilities' of the AI Era

From an industry perspective, the birth of the ATH Business Group coincides with a critical inflection point where the AI industry is shifting from 'parameter competition' to 'scenario implementation.'

2026 is widely recognized as the first year of AI Agents, with China's AI Agent market expected to surpass 1.001 trillion yuan and grow to 3.3009 trillion yuan by 2028.

Alibaba's core ambition is to define Tokens as the 'utilities' of the AI era, establishing a value measurement system and pricing power in the AI world by controlling Token production, distribution, and consumption.

The strategic framework is already clear: the MaaS platform serves as the 'logistics center' for Tokens, offering model services to the entire industry; Tongyi Labs acts as the 'power plant' for Tokens, continuously iterating multimodal large models; Qianwen and Wukong complete the final application of Tokens in the C-end and B-end, respectively.

This layout elevates Alibaba from a mere model provider and computing power supplier to an AI infrastructure service provider, breaking free from homogeneous model competition and moving toward a higher-level ecological positioning.

Especially in the B-end market, enterprise procurement is shifting from 'buying software' to 'buying digital labor.' The emergence of the Wukong Business Unit precisely aligns with this trend, potentially reshaping efficiency across enterprise operations, supply chains, and e-commerce management.

Looming Concerns

Despite its grand strategic blueprint, ATH still faces triple challenges in technology, market, and organization.

Technologically, large model hallucinations remain unresolved, and B-end scenarios demand extremely high accuracy and stability. A single error can lead to business risks, while standardized products struggle to adapt to highly fragmented enterprise needs, easily falling into the 'customization quagmire.'

Market-wise, Baidu, Huawei, Tencent, and ByteDance are all pressing in. On the C-end, user competition for Doubao and Yuanbao is fierce, while on the B-end, Wenxin and Pangu have deep industry penetration. Although Alibaba has ecological advantages, it lacks absolute barriers.

Commercially, domestic enterprises have weak software payment habits, and the Token-based billing model is still a novel concept. Enterprises care more about 'solving problems' than 'how Tokens flow.' Without killer applications, even the most splendid architecture will struggle to generate real revenue.

Organizationally, Alibaba's frequent structural adjustments bring collaboration costs, risks of core talent loss, and inherent contradictions between open-source technology and commercial profitability, all looming as concerns for ATH.

This Token-centric strategic gamble ultimately determines whether Alibaba can truly transform from an e-commerce giant into an AI tech giant. Eddie Wu's personal leadership reflects the group's utmost determination, signaling that there is no turning back in Alibaba's AI transformation.

For Alibaba, ATH is not just a simple business integration but a redefinition of commercial rules for the next decade: Tokens become the carrier of digital asset circulation, AI Agents become the primary workforce, cloud and models become infrastructure, and B-end and C-end drive growth together.

Success would grant Alibaba infrastructure discourse power in the AI era, unlocking a trillion-yuan growth curve. Failure would mean difficulty in realizing massive investments, technical advantages being caught up by rivals, and missing the industry window.

For China's entire AI industry, Alibaba's exploration also holds benchmark significance. After all, the endgame of AI is not model competition but commercial implementation; not conceptual hype but value creation.

Whether Token Hub can truly succeed and whether Wukong can break the B-end implementation curse will be decisive factors in the AI industry's developments in the coming year.