字节不认输 豆包手机重出江湖 已大规模量产

This article covers ByteDance’s official entry into the AI agent hardware space via a mass-produced, compliant second-generation Doubao AI phone co-developed with Nubia, marking the start of the third wave of internet traffic entry transformation.

1. Billed as the world’s first mass-produced AI agent smartphone, the new Doubao phone is powered by Snapdragon 8 Gen 3 Extreme Edition, equipped with a 7000mAh large battery, and comes in four color options. A distinct bright orange AI key boosts brand recognition, with Doubao AI deeply integrated into the system底层 to support automatic cross-application task execution. The device has obtained regulatory approval for on-device large model deployment, has an initial stock of nearly 200,000 units, and is currently available for public testing at the World Artificial Intelligence Conference (WAIC).

2. The industry has widely recognized AI agents as the next major traffic entry point. ByteDance has built a complete AI ecosystem spanning large models, developer ecosystems, terminal hardware, and e-commerce monetization. Doubao, its consumer-facing AI chatbot, has surpassed 100 million daily active users in China and already launched paid subscriptions. Going forward, AI is expected to gradually replace the current interaction logic of traditional apps and reshape user habits.

AI has emerged as a new high-growth track for consumer electronics and e-commerce, and brands can draw the following key insights from this industry development:

1. On the consumer trend side: User acceptance of AI-enabled products is already high. The 30,000-unit first batch of the first-generation Doubao test phone sold out quickly, with resale prices on secondhand platforms soaring as high as 36,000 RMB. Combined with the fact that domestic consumer-facing AI applications have surpassed 100 million daily active users, user habits are rapidly shifting toward "letting AI get things done." AI will quickly evolve from a differentiating selling point to a standard industry requirement.

2. On the product R&D side: Compliance and security must be top priorities. The second-generation Doubao phone addressed regulatory gaps ahead of launch, secured one of the first batch of official compliance approvals, and adjusted its functional logic to require manual user confirmation for all sensitive operations. This avoided the platform风控 restriction issues that plagued the first-generation model, a lesson all brands developing AI products should learn.

3. On the business strategy side: Brands can learn from ByteDance’s full-closed-loop approach, integrating AI capabilities with transaction scenarios to deliver end-to-end services and improve user retention and conversion rates.

AI e-commerce and AI hardware are opening up new growth opportunities, but also carry clear risks that need to be avoided. Key takeaways for sellers are as follows:

1. Policy and risk: The Interim Measures for the Administration of Artificial Intelligence Anthropomorphic Interaction Services took effect in July. Any AI-related business must obtain official compliance approval first. In addition, cross-application AI operations that violate platform security rules can trigger风控 restrictions — the first-generation Doubao phone faced full-platform App restrictions for this reason. Sellers should adapt to rules in advance and require user confirmation for all sensitive operations.

2. Growth opportunities: AI is reshaping e-commerce interaction logic, and all major platforms are pushing AI-enabled upgrades. AI can help sellers complete core operational work including graphic design, traffic placement, and product selection, effectively cutting operating costs. Sellers should integrate capabilities from leading AI platforms early to optimize operational efficiency and capture the AI traffic dividend.

3. Business model: AI is no longer free, as leading AI applications have already launched paid subscription services. Sellers can select AI services that fit their own needs to match business growth.

The mass market adoption of AI-native hardware is bringing new business opportunities and transformation insights for consumer electronics factories, outlined as follows:

1. Business opportunities: AI-native hardware has moved beyond testing and prototyping to large-scale mass production. The total stock of the second-generation Doubao phone reaches 200,000 units, nearly 7 times that of the first generation, indicating rapidly growing market demand for AI hardware. Factories can adjust capacity allocation ahead of time, partner with brands on AI hardware orders, and capitalize on this market opportunity.

2. Product design and manufacturing requirements: AI hardware needs distinct visual cues for AI interaction — the Doubao phone uses a bright orange AI key as a differentiating identifier. Meanwhile, AI computing places higher demands on performance and power consumption, requiring high-end chips and large-capacity batteries. Factories need to adapt their design and manufacturing processes to these new requirements.

3. Transformation insights: AI has moved from cloud-based software to consumer hardware, and the industry’s interaction logic is shifting. Factories can advance digital transformation and integrate AI capabilities to optimize production and design workflows to keep pace with industry changes.

The AI agent industry is moving from concept to commercial deployment, opening up new development directions for service providers. Key insights are as follows:

1. Industry development trends: The internet traffic entry point is shifting from app stores and super apps to AI agents, and competition has extended from software to hardware. AI-native hardware and AI e-commerce are both rolling out rapidly, with market demand entering the pre-boom stage, creating substantial room for overall industry growth.

2. Core client pain points: Security and compliance are the common top pain points for the current AI industry. The first-generation Doubao phone triggered风控 restrictions across major platforms because its cross-App simulated click permissions shared similar characteristics to external cheating tools, leading to severe functional limitations. Operating without compliance makes normal business impossible, and users also have concerns about the security of AI operations.

3. Solution directions: Service providers can follow the second-generation Doubao approach: optimize functional logic to require active user confirmation for sensitive operations, and launch products to market only after obtaining compliance approval. Service providers can also develop integration solutions that connect AI capabilities with major platforms and transaction scenarios to meet clients’ commercial deployment needs.

The AI interaction revolution brings new competitive challenges as well as clear development directions for platforms, outlined as follows:

1. Market demand for platforms has shifted: Users now want AI to complete all types of tasks in one stop without switching between multiple apps, while merchants want AI to cut operating costs and improve efficiency. Platforms need to adjust their strategic layouts to adapt to these new demands.

2. Three distinct AI platform business models have emerged in the industry, all worthy of reference: Alibaba’s full-scenario enablement model, which generates ecosystem revenue from AI cloud services; the JD-Tencent model of traffic and supply chain complementarity, where the two partners leverage each other’s strengths to deliver cross-scenario services; and ByteDance’s content-commerce closed-loop model, where AI connects the full user journey from product discovery to checkout.

3. Risk mitigation: Cross-App AI operations carry inherent security risks. Platforms need to establish reasonable风控 rules, align their own rules with evolving regulatory requirements, balance security and innovation, and lay out AI entry points early to avoid being sidelined in the new interaction transformation.

The internet industry is undergoing a new entry point transformation, with many new developments and emerging issues in AI agent deployment worth studying. Key insights are as follows:

1. New industry developments: The internet has already gone through two waves of traffic entry transformation — app stores and super apps — and is now entering the third wave centered on AI agents. AI has moved from cloud-based demos to mass-produced hardware, and competition has extended from software to hardware. ByteDance has built a complete AI closed loop covering large models, developer ecosystems, terminal hardware, and e-commerce monetization, while other leading players including Alibaba, JD.com, and Tencent have all completed their own strategic布局s.

2. New industry problems: The security and compliance of cross-App AI operations remains unresolved, as seen in the first-generation AI phone that faced full-platform风控 restrictions. Balancing AI functional deployment, user security, and platform rules remains the core open problem the industry needs to solve.

3. Business model developments: Leading domestic AI applications have launched paid subscriptions, officially ending the era of free AI. Three distinct mature business models have also emerged in AI e-commerce, providing new research samples for industrial research.