Introduction

The recent pricing announcement from Doubao App Store, revealing subscription rates for its AI services, signifies a pivotal shift in the domestic AI industry. The standard annual package is priced at 688 yuan, with other tiers at 2048 yuan and 5088 yuan for enhanced and professional versions, respectively. Although the details are still being tested, the framework of “free basic + paid premium” is becoming clear.

Transition from Free to Paid

This transition is not surprising. Free services and subsidies have always been competitive strategies rather than the norm. The recent AI red envelope competition during the Spring Festival was about gaining market entry and user habits, representing the last major traffic harvest of the mobile internet era. However, the reality is that there is no such thing as a free lunch. As the saying goes, “Token shortages” have become a consensus, and the cost of computing cannot be indefinitely covered by capital. Thus, charging for services has become an inevitable commercial model, marking a rite of passage for technology commercialization.

Industry Trends

A recent report from the Daily Economic News corroborates this industry trend. With the conclusion of the A-share annual report season, AI application companies have shown a mixed but converging performance for 2025: 35 companies reported a year-on-year net profit growth of over 50%, while 40 companies saw a decline in revenue. On one hand, the revenue scale continues to expand, with AI capabilities deeply embedded in core products, leading to rapid growth in user engagement and usage metrics. On the other hand, high R&D investments are still compressing short-term profits, placing the industry at a critical juncture from technological breakthroughs to commercial validation.

Three Key Realities

The swift transition from free services to paid plans reveals three key realities:

1. Computing Costs Are Real: Subsidies are essentially a countdown where capital is exchanged for time, which cannot be sustained indefinitely.
2. Model Capability Maturity: The iteration of model capabilities has reached a monetization phase, where complex task scenarios justify charging fees, and the disparity between free and paid experiences supports tiered offerings.
3. Industry Consensus: There is a growing disconnect between the grand narrative of “AI changing civilization” and the reality of “subsidy-based marketing”. If charging does not begin, even those telling the story will lose faith.

Doubao’s Strategic Move

As one of the largest domestic AI applications, Doubao’s decision to charge for its services reflects confidence in its technological value and represents a breakthrough in a highly competitive industry. If Doubao does not take this step, its competitors may find it difficult to follow suit. In contrast, international players like ChatGPT and Claude have already adopted similar pricing strategies.

Implications for the Future

This is not merely a price adjustment; it signals a critical shift for domestic large models from “burning money to capture market share” to “self-sustaining growth through iteration”. The competition logic has shifted from “traffic accumulation” to “value realization”. The pressing question is not whether AI should charge fees but rather how to ensure that users continue to pay after the transition.

When users begin to pay monthly, the era of “story-driven user acquisition” ends, and the era of “value verification” begins. In the free era, the focus was on daily active users (DAU), where users tolerated basic functionality. In the paid era, the focus shifts to user retention and depth of engagement, where users demand a product that is not only good but also irreplaceable.

Value Verification

Doubao’s paid features are anchored in complex productivity scenarios such as PPT generation, data analysis, and film production. These tasks require robust foundational model capabilities and test the limits of context understanding, multi-modal collaboration, and professional logic. Any shortcomings will be magnified under the scrutiny of paying users. If the paid experience remains at a “toy” level, users will vote with their feet.

Thus, pricing is merely the entry ticket; true competitiveness lies in the ability to deliver top-tier experiences to paying users. This includes providing fast-track access during peak times rather than causing anxiety from long wait times, which is crucial for retention rates and average revenue per user (ARPU).

Sustainable Growth

Moreover, the revenue generated from charging must be reinvested in computing power and R&D iterations, creating a positive feedback loop of “technology upgrades – experience enhancements – user payments – reinvestment”. This is not only vital for Doubao but also for establishing a sustainable competitive edge for domestic AI on the global stage.

Conclusion

Charging does not equate to abandoning accessibility. The free version will continue to support everyday use, while the paid version will meet professional productivity needs. This tiered approach is reasonable, but companies must recognize that the right to charge is granted by users, not bestowed by the platform. In this hard-hitting phase of value verification, only by consistently demonstrating real capabilities and providing visible upgrades can companies ensure users feel their spending is justified, thus encouraging ongoing payments. This may be the core challenge that all domestic large models must face moving forward.