Analysis of Commercialization Challenges and Breakthrough Strategies for MiniMax's Consumer-facing AI Applications

According to MiniMax’s Hong Kong IPO prospectus and financial data, the company’s overall gross margin in the first three quarters of 2025 was 23.3%, but

[1][2]. Behind this data, there are obvious structural characteristics:

93% of MiniMax’s cost of sales comes from cloud computing service expenditures[2], meaning that computing power costs form the main burden of its C-end business. In stark contrast, the gross margin of its B-end business reaches as high as 69.4%[1]. Notably, after excluding the impact of Xingye (Domestic Version of Talkie), which has not yet been fully commercialized, the actual gross margin of the C-end business is close to 50%[3], indicating that MiniMax has a healthy profitability foundation in its commercialized products.

In the first three quarters of 2025, revenue from C-end native AI products reached $38.02 million, accounting for 71.1% of total revenue[1]. Among this, subscription revenue from Conch AI was approximately $14.147 million, and advertising revenue from Talkie/Xingye was approximately $11.188 million[2]. Talkie and Conch AI contributed 35.1% and 32.6% of total revenue respectively[1], forming the core growth engines.

A more severe challenge for MiniMax lies in the weakening momentum of user growth. Data shows that

[2]. Most importantly:

: In the first nine months of 2025, the year-on-year growth rate of new users of MiniMax’s main application decreased by 46.36%, while the year-on-year growth rate of new users of Talkie/Xingye decreased by 7.4%[2].

: The average monthly active users (MAU) of MiniMax’s main application decreased by 34.03% year-on-year[2], meaning that the expansion of user scale has not translated into higher usage frequency; the MAU of Talkie/Xingye increased by 61.71% year-on-year, which is significantly lower than the 124.35% growth rate of its user base[2], indicating that user quality rather than quantity is becoming the core issue.

: Marketing expenses peaked at $86.995 million in 2024, and although they dropped to $39.325 million in the first nine months of 2025, they still accounted for more than 10% of the current period’s revenue[2]. The company has accumulated over 100,000 promotional materials[2], and user expansion still relies heavily on paid user acquisition to a considerable extent.

---

Globally, successful AI applications are achieving commercial breakthroughs at an astonishing rate.

[4]. These cases reveal the core laws of AI application commercialization:

: Successful AI applications do not directly sell generative capabilities themselves, but instead sell “integration of credible results” and “final delivery forms”[4]. For example, Perplexity sells integrated search results, and Replit sells application operation and hosting delivery. This model effectively reduces the decision-making cost for users to pay.

: Investors’ evaluation criteria have shifted from traditional SaaS profit margins to ARPU[4]. Even if the gross margin is not high, as long as it can take over large-scale human resource budgets and form high user value, sustainable growth can be achieved.

: Represented by Perplexity, it follows the path of “First To-C, Then To Team, Finally To Enterprise”[4]. It first allows individual users to pay for outcomes, then increases ARPU through collaboration, management, and permission systems. MiniMax’s current C-end user base exactly has this potential.

Product	Core Model	Paid Conversion Strategy	ARPU Characteristics
Character.AI	Freemium + Subscription	Free basic features + premium feature subscription ($9.99/month)[4]	Relies on high-frequency usage scenarios
Gamma	Tiered Subscription	Accumulate 50 million free users → 1% paid conversion rate[4]	25% of subscriptions come from teams, boosting ARPU
HeyGen	Subscription Model	Driven by feature iterations (digital human live streaming doubled paid users)[4]	Focuses on commercial scenarios
Suno	C-end Subscription-focused	Paid access to higher generation quotas and commercial usage rights[4]	Targets creator groups

---

MiniMax should learn from the experience of global successful cases,

[4]. Specifically:

• Talkie/Xingye
  : Upgrade from “AI Companion” to “Emotional Solution Provider”, providing proof of quantifiable emotional support effects
• Conch AI
  : Upgrade from “AI Chat” to “Professional Content Production Assistant”, delivering directly usable video and copywriting outcomes
• Open Platform
  : Upgrade from “API Calling” to “Enterprise AI Transformation Solution”, undertaking complete business workflows

Currently, subscription revenue of MiniMax’s products mainly comes from Conch AI ($14.147 million)[2], but the average expenditure per paid user is approximately $56, which is only 11.2 times that of Talkie[2], indicating that the potential of high-value tool-type paid users has not been fully tapped. Recommendations:

• Basic Version
  : Retain core chat functions, limit the number of generations and precision
• Professional Version
  : Unlock high-definition video generation, unlimited chats, and brand-level content output
• Enterprise Version
  : Provide API access, team collaboration, and private deployment

Currently, MiniMax’s paid conversion rate is less than 1%[2], which is far lower than the 1% conversion rate of industry benchmark Gamma[4] (whose registered users have reached 50 million). This means that with a sufficient user base, there is huge room for conversion rate improvement:

• First-order Conversion
  : Provide limited-time free trials or first-month discounts to lower the threshold for first-time payment
• Value Anchor
  : Design multi-tier subscription plans (e.g., $9.99/month Basic Version vs. $49.99/month Professional Version) to use the price anchoring effect to boost mid-to-high-end conversions
• Scenario Trigger
  : Pop up a prompt to “Upgrade to unlock HD export/commercial authorization” after users generate satisfactory content

• Establish a points system: Earn points through daily check-ins, inviting friends, and content sharing
• Points redemption benefits: Free generation quotas, exclusive characters, priority customer service
• Tier privileges: Premium members enjoy exclusive models and customized services

Learning from Perplexity’s revenue-sharing advertising model[4] and Character.AI’s brand cooperation experience[4], MiniMax can explore:

• Advertising Revenue
  : Embed native advertisements (such as brand virtual characters) in Talkie’s companion scenarios
• Value-added Services
  : Provide high-priced services such as AI character customization and exclusive chat training
• Content Revenue Sharing
  : Share commercial proceeds from generated content with content creators
• E-commerce Commissions
  : Embed product recommendation functions in Conch AI and collect commissions based on conversions

Currently, the gross margin of MiniMax’s B-end business reaches as high as 69.4%[1], and the number of paid clients has surged from 100 in 2023 to 2,500 in the first three quarters of 2025[1]. Recommendations:

• Industry-specific Solutions
  : Launch customized solutions for vertical industries such as film and television, education, and marketing
• Private Deployment
  : Provide localized deployment services for large clients to increase customer unit price
• Ecosystem Cooperation
  : Integrate with office software and SaaS platforms to expand reach channels

93% of cost of sales comes from cloud computing services[2], which is the core factor restricting gross margin. Recommendations:

• Model Distillation
  : Deploy lightweight models on edge devices to reduce cloud calls
• Hybrid Deployment
  : Use self-built computing power for high-frequency scenarios and public cloud for low-frequency scenarios
• Dynamic Scheduling
  : Allocate computing power resource priorities based on user payment tiers

Currently, marketing expenses account for more than 10% of revenue[2], and user growth has reached a bottleneck. Recommendations:

• Word-of-mouth Propagation
  : Optimize product experience and encourage spontaneous user sharing
• Community Operation
  : Establish user communities to reduce marginal customer acquisition costs
• Content Marketing
  : Attract organic traffic through high-quality content and reduce reliance on paid promotion

---

Action Item	Objective	Key Metric
Optimize subscription tiers	Design 3-4 subscription plans	Increase paid conversion rate to 1.5%
Launch first-order discount	Lower first-time payment threshold	30% increase in number of new paid users
Strengthen enterprise sales	Establish B-end sales team	Increase B-end revenue share to 35%

Action Item	Objective	Key Metric
Launch Enterprise Version	Customize for industry clients	Increase customer unit price to $5,000 per year
Launch membership system	Enhance user stickiness	Increase 30-day retention rate to 40%
Launch open platform value-added services	Increase API call revenue	Increase ARPU to $80

• Integrate the user systems of Talkie, Conch AI, and the open platform
• Achieve interconnection of user data and benefits
• Form a complete value chain of “Tools → Content → Services”

• Increase C-end business gross margin to over 50% (already close after excluding Xingye)
• Increase overall paid conversion rate to 3-5%
• Surpass $200 million in Annual Recurring Revenue (ARR)

---

: MiniMax has been involved in multiple copyright disputes[2]. Issues regarding the legality of training data and the similarity of generated content may bring legal costs. It is recommended to establish a content review mechanism and build authorization cooperation with copyright holders.

: The differentiation paths of China’s “Six AI Dragons” have become clear[3]. Zhipu (GLM) is betting on enterprise privatization + API, and MiniMax’s flagship coding/voice model public cloud calls are facing direct competition. It is recommended to continue investing in technological innovation to maintain leading model capabilities.

: The 46.36% decline in new user growth rate[2] indicates that the pure scale expansion model has failed. It is recommended to shift the focus from “user quantity” to “user value”.

• Compliance First
  : Establish a legal team and lay out copyright cooperation in advance
• Differentiated Competition
  : Focus on advantageous scenarios such as multimodal models and emotional companionship
• Efficiency First
  : Shift from pursuing growth rate to pursuing unit economic benefits

---

The gross margin dilemma of MiniMax’s C-end AI applications is essentially a microcosm of the “scale before efficiency” development model in the AI 2.0 era.

[1][2]. The key lies in whether it can complete the strategic transformation from “scale expansion” to “value deep cultivation” before user growth slows down.

Learning from the experience of globally successful AI applications, MiniMax should focus on three directions:

. If it can achieve the phased goals of increasing the paid conversion rate to 3% and restoring the C-end gross margin to 50% within the next 18-24 months, MiniMax is expected to become a benchmark case for global AI application commercialization.

---

[1] Eastmoney.com - MiniMax’s Hong Kong IPO: The Fastest-listed AI Company, Overseas Revenue Accounts for Over 70% (https://caifuhao.eastmoney.com/news/20251231104413440297380)
[2] PopCJ - MiniMax’s Hundred-billion Market Value: A Victory for AI 2.0 or a New Capital Illusion? (https://www.popcj.com/depth/2509532601509570)
[3] Sina Finance - Going Overseas and Going Public: China’s First Batch of Large Models Have Made It (https://finance.sina.com.cn/tech/roll/2026-01-04/doc-inhfcasw7648852.shtml)
[4] OFweek AI Network - Reaching $100 Million in 8 Months! Review of the 9 Most Profitable AI Applications Globally, AI Business Logic Has Changed Completely (https://m.ofweek.com/ai/2026-01/ART-201700-8420-30678839.html)