Author: Hao Xin

Editor: Wu Xianzhi

Doubao, a platform boasting 345 million monthly active users and long known for its 'mass-market' approach, has recently taken an unexpected turn.

In early May, Doubao officially announced the introduction of a paid subscription model, sparking widespread discussion among users and industry observers alike.

The paid subscription service is tiered into three options: Standard at 68 RMB/month, Enhanced at 200 RMB/month, and Professional at 500 RMB/month. It is primarily designed for complex task scenarios such as PPT generation and data analysis, while basic functionalities remain accessible for free.

According to sources familiar with Doubao's operations, the paid model will initially undergo a period of grayscale testing, with select users gaining early access to the new membership rules. 'If all proceeds as planned, it will be fully rolled out by the end of this month,' the source added.

However, many users have reported that after updating to the latest version of Doubao, they have yet to encounter the paid model, indicating that large-scale grayscale testing may not have commenced as of yet.

The source further revealed that the Doubao team had indeed received user feedback regarding the demand for AI-powered PPTs and spreadsheets, which served as one of the key motivations for considering a paid model.

Doubao is not the pioneer in adopting a paid subscription model. Overseas, platforms like OpenAI and Claude have implemented subscription models from the outset, with prices increasing as intelligence levels improve. Domestically, startups such as MiniMax and Moonshot AI have partially shifted their focus away from casual users towards productivity scenarios in pursuit of listing and commercialization goals.

With this move, Doubao has not entirely abandoned its 'mass-market' strategy but has opted for a more moderate 'middle ground' approach.

Another insider disclosed that computing resources were the primary driver behind this decision. Professional scenarios such as expert modes, PPT generation, and spreadsheet creation are highly resource-intensive.

'Currently, there are no additional resources allocated, leading to many users encountering errors or receiving 'try again later' prompts. With a paid model, we can ensure a seamless experience without such issues,' the insider explained.

Doubao's Gradual Approach

Various indicators suggest that Doubao's previous free model has been facing resource constraints. While casual users may not have noticed any significant issues, heavy professional users have experienced a decline in service quality.

Internally, the Doubao team believes that the time is ripe for introducing a paid model, based on two key factors: First, genuine user feedback indicates that some users are willing to pay for a stable experience—this is not a fabricated demand. Second, AI assistant usage scenarios have clearly stratified.

The free version can adequately cover most daily conversations and simple Q&A sessions. In contrast, features like AI-powered PPTs and smart spreadsheet editing are resource-intensive, less frequently used, but highly valuable. These low-frequency, high-cost features are naturally suited for a subscription model.

Currently, Doubao's professional features are comparable to those offered by competitors, suggesting that the main selling point of the paid model is not exclusive functionality but stable service quality. This indicates a need to 'catch up' on product maturity—ensuring usability before differentiating features.

Thus, the direct purpose of introducing paid subscriptions is not to exclude free users but to create a 'fast lane' for heavy users, using pricing to filter demand and ensure service stability. In essence: free for the masses, performance for professionals.

The most notable aspect of Doubao's strategy is that it has not alienated free users. The official response emphasized that 'basic features remain free,' and the Doubao team has previously stated that 'most needs of casual users are covered by the free version.'

In contrast, Kimi's pricing strategy is more aggressive. To some extent, Kimi has abandoned traffic and casual users—even for simple questions, users must wait or are informed that they are in a 'peak computing period' and advised to subscribe.

Doubao appears more restrained, aiming to maintain traffic while exploring commercialization opportunities. Its 'gradualist' approach is essentially a commercialization strategy that leverages user scale as its foundation, computing costs as its boundary, and scenario stratification as its leverage. This aligns with ByteDance's product philosophy of scaling first, monetizing later, and cautiously preserving the core experience.

ByteDance's massive user base does not allow for 'shock therapy.' Doubao has 345 million monthly active users, a strategic asset for ByteDance. A sudden shift to full paid access would result in significant user loss, with competitors ready to capitalize on the situation. The 'gradualist' approach maintains the free baseline, protecting traffic while testing the ceiling.

If prices are set too high or pushed aggressively, users will feel that it is not worth it. A more moderate compromise allows charging based on service quality first, with room to gradually raise prices or add tiers as features improve.

At this stage, this approach is rational. It avoids massive user churn, gives the team time to develop differentiated features, and generates revenue to fund computing expansion.

According to sources close to ByteDance, following its usual operational logic, Volcano Engine uses token consumption as a core business metric, and the Doubao team previously included traffic growth in performance evaluations. Once Doubao's paid model is fully implemented, paid conversion rates are likely to be added to the evaluation system.

Computing Power, Still the Key Bottleneck

Looking at the broader timeline, several ByteDance-backed AI products began charging or adjusting prices in the first half of this year. Doubao's move sparked controversy simply because C-end users are most sensitive to such changes.

ByteDance's AI coding tool, Trae, adjusted its pricing from 'per-use' to 'per-token' and introduced five subscription plans. Its AI video generation tool, Jimeng, raised prices three times in a month, effectively devaluing points, while Volcano Ark's Coding Plan suspended its first-purchase discount in March and now offers two paid plans without additional temporary discounts.

Doubao's move ignited public debate, with some saying, 'Even cash-rich ByteDance is running out of money.'

Undeniably, computing resources are currently the biggest bottleneck for AI development. GPU lease prices remain high, and domestic AI applications lack the USD pricing power enjoyed by OpenAI. As a result, all major model providers subsidize AI with profits from other businesses.

Alibaba has e-commerce and cloud, ByteDance has advertising and Douyin, and Tencent has gaming. However, unlike Alibaba Cloud, which has a vast enterprise client base and mature cloud revenue, ByteDance's Volcano Engine started later and has fewer client resources, putting it at a disadvantage in computing leasing monetization.

Thus, ByteDance urgently needs two things: direct C-end charging and point-to-point GPU leasing for major clients. This explains Doubao's pricing—not because the product is perfect, but because the computing bills can no longer wait. Only professional scenarios justify user payments to cover the high costs of individual requests.

There are reports that ByteDance is investing in self-developed GPUs, which could be a game-changer. If successful and deployed at scale, ByteDance's computing costs would drop significantly, leaving Doubao with two options:

1. Lower prices to attract more paying users and squeeze competitors' space.
2. Maintain prices, boost profits, and reallocate saved computing resources to improve the free version's experience, widening the gap with rivals.

Beyond computing power, a clear signal from insiders is that ByteDance and Doubao believe 'China has reached a turning point in AI commercialization, with mass-market and professional segments now diverging.'

ByteDance is building a closed-loop AI business model. P-end and enterprise users pay for capabilities in products like Doubao, Trae, and Jimeng/Dreamina. Commercialization is a step toward optimizing the product ecosystem, stratifying users, and establishing a healthy business model for long-term AI infrastructure investment.

This is not just ByteDance's choice but also marks the establishment of a commercialization paradigm in China's AI industry. The era of free services sustained solely by burning cash is ending, as providers strategically seek paid scenarios.

AI price hikes have surged this year. OpenAI's ChatGPT Plus rose from $20 to $30 per month, Claude Code was removed from the $20/month Pro plan, and Cursor Pro doubled from $20 to $40. Divided into professional, premium, and elite tiers, Doubao's prices are significantly lower than ChatGPT, Claude, and Gemini overseas and slightly below Kimi's domestic pricing.

This is an inevitable result of technological maturity and agent proliferation, indicating that free models can no longer support unlimited, high-load professional demands, necessitating stratified services.

The Third Path and Industry Benchmark

The market currently divides into two camps:

• The free-leaning camp, including Qianwen, Yuanbao, and DeepSeek, backed by Alibaba, Tencent, and DeepSeek, respectively, covers computing costs through other group businesses or investors. Their primary goal is to acquire users, accumulate data, and drive cloud service adoption. Sustainability depends on the backers' willingness to continue funding; if the group cuts budgets, free access may be reduced.
• The fully commercialized camp, such as Kimi and MiniMax, lacks unlimited resources and must quickly become self-sustaining. Their free versions are limited, forcing users to either pay or tolerate lower service quality or usage caps. This model survives but struggles to scale user numbers and risks ceding free users to competitors.

Doubao is now testing a third path: free basic features to maintain traffic, while charging for professional scenarios to cover high computing costs.

Whatever Doubao's commercialization progress, it will serve as an industry benchmark.

AI assistant paid conversion rates are typically low. If Doubao achieves a 1-2% conversion rate among its 345 million monthly active users—yielding 3.5-7 million paying users—it would demonstrate a replicable commercialization model for domestic AI applications. Products with hundreds of millions of monthly users could rely on 1% of core users for operations, significantly boosting capital confidence in the AI application layer.

More importantly, Doubao's model does not require users to change habits or providers to abandon scale. The resources and ecosystems built by major players can become their unique strength.

If this template proves successful with 345 million users, it will become the easiest path for most AI assistant products to replicate, supporting the future valuation of AI application companies.

Doubao's 'third path' resembles a test of the industry's payment threshold.

Doubao's value lies not in how much it earns but in showing the industry whether, under current computing costs, an AI assistant with hundreds of millions of users can achieve sustainable operations through stratified services. It reveals what AI capabilities users are willing to pay for, how much, and where the boundary between free and paid lies.

These answers are needed not just by ByteDance but also by Alibaba, Tencent, Baidu, and all AI startups.