Summary of Key Points
The widespread adoption of AI technology has transformed the business models of C-side SaaS software (such as Canva and Figma), which were originally considered "low-cost" operations, into more capital-intensive ones. The more active the users are, the greater the consumption of AI computing power, putting increasing pressure on company profits. The traditional revenue models, which relied on attracting free users and then charging for premium subscriptions, have become ineffective. As a result, companies have been forced to adjust their pricing strategies to adopt a hybrid model that combines basic subscriptions with additional fees, similar to how gyms balance the need to retain customers with the need to control costs.
1. AI Turns SaaS from Profitable to Cost-Intensive: Free Users Become a Burden
Before the advent of AI, C-side SaaS was a relatively low-risk business model. For example, Canva could attract 260 million monthly active users through its free editor and then convert them into paid customers by offering advanced templates and licensed materials. The more users there were, the lower the cost per user (since using the templates hardly incurred any expenses). However, with AI, the situation has changed:
- Free users now use AI to generate images and documents, and each usage consumes GPU computing power, which represents a real financial expense.
- Canva offers 200 free AI usage credits per month, but this has turned free traffic from a asset into a financial burden. The more users utilize these services, the more expensive their bills become.
- Figma experienced similar challenges: although its revenue increased by 48% in the second quarter, its AI-related infrastructure costs rose by 117%, leading to a shift from profitability to loss.
2. Pricing Hurdles
With rising costs, raising prices seems like an obvious solution, but each pricing approach comes with its own problems:
- Pure Subscription Model: Revenue is fixed, but costs can fluctuate significantly. Adobe increased its prices after adding AI features, and while subscription revenue grew by 13.7%, costs rose by 16%. The more users used AI, the greater the company's losses.
- Pay-As-You-Go Model: Users are wary of being overcharged. Replit charges based on the amount of AI tasks completed, leading to complaints such as $32 being deducted for a failed task or prices increasing from $2 to $30 for a single project. Additionally, calculation errors sometimes result in overcharging.
- Pay-Per-Result Model: Defining what constitutes a "solved" issue is subjective, leading to disputes between companies and users. For example, Zendesk charges for AI-assisted customer service issues, but users question whether 72 hours of no interaction counts as a resolved issue or whether 40% of cases require manual review.
3. Hybrid Pricing: The Most Practical Solution
After experimenting with various pricing models, companies have settled on a hybrid approach that combines basic subscriptions with additional fees:
- A fixed base fee is charged to ensure a minimum level of revenue, and then extra usage of AI services or successful outcomes are billed separately. This helps spread the costs among more users.
- For instance, Figma and Canva have split their subscriptions into basic packages and optional add-ons. Intercom charges a monthly fee for using its services and then charges $0.99 per resolved issue.
- The essence of this model is to balance risks: companies avoid bearing all the costs, and users don't have to worry about unexpected expenses.
4. The Gym Metaphor: The Dilemma of AI-Based Software Business Models
The analogy with gyms highlights the contradictory relationship companies face with their users:
- Gyms want customers to subscribe to annual memberships (for higher loyalty) but fear high daily usage that would wear out their facilities and staff.
- AI-based software companies desire active, loyal users but are concerned about excessive use of AI resources, which can lead to significant costs.
- This contradiction reflects a fundamental shift in the nature of software businesses in the AI era: the focus is no longer on attracting as many users as possible but on ensuring that users use the services in a balanced manner that maximizes profits.
In Conclusion
AI has transformed C-side SaaS from a low-cost, easy-to-manage business model into one that requires careful cost management. Companies must find a new balance between providing valuable services to users and controlling expenses. The current reality for AI-based software companies is akin to operating a gym: they need to keep their customers engaged without incurring excessive costs.