Recently, the sale of AI courses has once again stirred up controversy.

The catalyst? Film Storm, a prominent technology video creator with tens of millions of followers, launched an AI practical course priced at 49 RMB during a live stream. Within just a day, sales soared past 100,000 units, projecting daily course revenue at over 4.9 million RMB. However, the product page was inundated with negative reviews:

Complaints ranged from rigid, template-based operations, results not aligning with tutorials, to fans feeling let down after coming for Tim...

It's worth noting that this isn't Film Storm's first foray into course sales. A glance at their knowledge service flagship store reveals courses like "Comprehensive Essential Course for CapCut" and "Cinematic iPhone Filmmaking" have sold over 300,000 and 800,000 units, respectively. In the AI era, widespread anxiety has even driven top creators to participate firsthand.

According to the "2026 China Knowledge Monetization Industry Development Report," the market size is projected to hit 57 billion RMB in 2026, marking an 18.8% year-on-year increase, with paying users exceeding 432 million. Among these, AI tool application courses have emerged as the fastest-growing segment, with users' average annual expenditure on knowledge consumption rising to 380 RMB and repurchase rates surpassing 65%.

Amidst this AI learning craze, tech creators, education creators, business creators, and even motivational speakers have become pivotal distributors of AI courses. Course sellers are ubiquitous across the content landscape.

Do AI Courses "Rescue" Declining Monetization Capabilities?

Beyond Film Storm, numerous creators are venturing into AI course sales.

Li Yizhou, a Tsinghua Ph.D. dubbed the "First-Generation AI Guru," sold "Everyone's AI Course" for 199 RMB in 2023, amassing approximately 250,000 sales within a year, generating 50 million RMB in revenue. Similarly, top business influencer Zhang Qi launched the "AI + Business Growth" course priced at 520 RMB. Li Xiaolai, a renowned New Oriental instructor and author of "The Road to Financial Freedom," also trial-taught the "AI Beginner's Course (Transition)" within his community.

Last year, a knowledge monetization creator reportedly earned 200,000 RMB in just four days selling a course on DeepSeek.

In reality, creators relying on robust content output, such as tech, knowledge, and business creators, have struggled with monetization in recent years. During tough times, platforms like Bilibili, which host a high concentration of such creators, witnessed a wave of top creators leaving or halting updates.

KAS Data specifically analyzed knowledge accounts like "Chief of Top Secret Research Institute," which gained 3 million followers in 15 days. However, in terms of monetization, the account's cumulative sales over 30 days were a mere 8,541 RMB, with some products showing zero sales.

In 2025, "Sina Finance" reported that 300 million self-media creators were halting updates en masse, with commercial orders as low as 10 RMB fiercely contested. Against this backdrop, major creators are scrambling to find new survival strategies.

A prime example is Film Storm. Founded by Tim, public data indicates the account boasts over 40 million followers across platforms, including 16 million on Bilibili and 14 million on Douyin. It has been honored as Bilibili's "Top 100 UPs" for six consecutive years, covering equipment reviews, film and television education, original short films, commercial advertisements, and more.

In 2024, the team's revenue surpassed 100 million RMB. E-commerce, self-branded products, and course sales constitute Film Storm's commercial segment. To drive traffic, the account also launched the "Survival on a Desert Island" challenge live stream, themed "Send Tim to the Island for 100 Hours," attracting 24 million viewers and generating 500,000 RMB in live stream sales.

However, not all accounts can swiftly establish a vast commercial foundation. In contrast, selling courses currently represents the lowest-threshold, highest-efficiency, and zero-infrastructure monetization shortcut for creators. The AI knowledge monetization market spans various segments, including AI drawing, AI video, large model applications, and prompt engineering.

Rapid technological iteration-driven anxiety allows each segment to spawn a plethora of courses.

Currently, courses sold by major creators focus on diverse areas.

Li Yizhou's beginner-oriented "Everyone's AI Course: Zero to Hero" generated 50 million RMB in sales. His marketing-focused "Super Traffic Marketing Course - IP + AI," priced at 399 RMB, sold 78,000 units, amassing 30 million RMB in sales. Growth instructor Zhou Zuoluo's monetization-focused "ChatGPT AI Monetization Circle" generated 2.79 million RMB in sales in 17 days.

Leveraging the traffic dividend of large model applications and the allure of side hustle monetization, more creators are achieving efficient monetization, especially as traditional advertising and live stream e-commerce approach their growth ceilings.

According to QuestMobile's "2025 Internet Advertising Market Half-Year Report," the internet advertising market is growing moderately, with advertisers' budgets remaining conservative, top traffic growth peaking, and ad unit price growth slowing down. Live stream e-commerce has also significantly cooled. According to iResearch data, the growth rate of live stream e-commerce fell from 245.9% in 2019 to double digits in 2023, with an expected growth rate of around 18% from 2024 to 2026.

Amidst these signs, AI course sales have emerged as a pivotal monetization method in the new content landscape.

But can AI courses truly "rescue" the declining monetization capabilities of creators like Film Storm?

Firstly, traffic acquisition costs for knowledge monetization are skyrocketing. Data indicates that in 2025, the repurchase rate for knowledge monetization users was below 8%, with course completion rates under 15%, while traffic acquisition costs rose by 120%. Secondly, the proliferation of low-quality courses is backfiring on creators, with many long-time fans of Film Storm expressing disappointment.

Previously, several knowledge creators, including "Li Yizhou," "Haishen Ge," and "dada," faced account suspensions or bans due to controversies over course sales.

When all top creators become AI distributors, can ordinary users still trust their recommendations? This is a question worth pondering.

Do KOLs Dominating Traffic Entries Render AI Competition Meaningless?

As creators flock to the course-selling arena, top content creators who command user trust have become AI "distributors" controlling user mindshare entries. A thought-provoking question arises: Is the focus of competition in the AI application layer rapidly shifting from product capabilities to channel control?

Firstly, the AI tools industry is facing an increasingly severe homogenization dilemma.

In early 2026, "National Business Daily" reported that despite the AI application boom, homogenization is rampant, with many AI products on the market sharing highly similar underlying logic. Rapid model updates and converging tool functionalities make it nearly impossible for AI tools to establish long-term moats solely through technological differences. A significant portion of industry attention must now shift towards traffic and channels.

Data supports this notion.

As of June 2026, domestic mobile AI native apps and application plugins reached approximately 765 million users before deduplication, forming a nationwide wave of AI tool adoption. In the first half of the year, the top tier of AI applications—Doubao, Qianwen, and DeepSeek—had monthly active user bases of 382 million, 167 million, and 130 million, respectively.

Under the Matthew effect, year-on-year growth rates are highly concentrated among top players, with Doubao growing by 172.1% and Qianwen by 5792.9%. Qianwen's explosive growth is directly tied to its integration with Alibaba's ecosystem businesses, including Taobao, Alipay, Taobao Deals, Fliggy, and Gaode.

In other words, entry competition has begun to intensify in this highly concentrated market.

The situation is similar overseas. In May 2026, ChatGPT continued to lead the AI assistant market, becoming the fastest app in global mobile application history to reach 1 billion monthly active users (MAU). In the first quarter of 2026, ChatGPT, Google Gemini, and DeepSeek accounted for nearly 90% of total usage time in the AI assistant app category.

As competition for user acquisition intensifies, advertising spending on generative AI-related ads in the U.S. market more than tripled in the first quarter of 2026 compared to the previous year. Data shows that OpenAI and Anthropic's ad spending surged by over 800% year-on-year.

Currently, partnering with top KOLs to launch customized courses has become one of the most efficient customer acquisition methods for AI tool vendors. The "2026 AI Product Global Expansion Core Guide" by InMega MCN specifically mentions methods for building an efficient KOL matrix for AI global expansion, categorizing potential KOLs into three types:

Hardcore analytical KOLs, including tech review creators, efficiency tool experts, and industry programmer UPs; scenario-based implant KOLs, including lifestyle vloggers, digital nomads, and workplace experts; and vertical niche KOLs, including designers, editors, e-commerce experts, and programmers.

Unlike shallow exposure through regular advertising, KOL course distribution achieves deep user mindshare occupation: after users follow tutorials step-by-step, form usage habits, and become deeply engaged, the cost of switching to other tools increases significantly. Some AI products have already validated this logic, with the typical example being Jasper AI, an AI marketing tool and writing assistant founded in 2021.

Jasper initially accumulated users through communities, reviews, and teaching content on Facebook.

Today, "course-driven traffic" for AI products has evolved to be more direct, especially for smaller applications.

It is reported that Film Storm's AI course was not independently developed but launched in collaboration with TapNow, a third-party AI generation service platform. Developed by Shenzhen Tianke Intelligent Technology Co., Ltd. and launched in 2025, TapNow positions itself as an AI visual creation platform for "professional film and television process tooling."

However, users are gradually seeing through the nature of "courses as lead generation ads," accelerating the depreciation of creators' trust assets.

In recent reports, many users who purchased Film Storm's course complained that a significant portion of the content introduced the functions and usage methods of the partner platform TapNow. Some users joked, "I spent 49 RMB watching hours of ads" and "I bought a TapNow manual for 49 RMB..."

Additionally, the model of tying course access to platform subscriptions has raised suspicions of indirect lead generation.

More intriguingly, this "course-driven traffic" model may not represent a stable state and could evolve in two directions.

Firstly, top creators may extend upstream, incubating self-developed AI tools based on their IPs to internalize channel profits. After all, the barriers to AI entrepreneurship are lowering. In July 2026, "Youth Daily" reported a surge in small-scale, low-capital AI startups in 2026.

According to Tianyancha Professional data, approximately 472,000 new AI-related enterprises were registered in China as of 2026. Top teams like Film Storm, possessing both technical capabilities and user bases, may launch their own AI film and television creation tools in the future to reduce reliance on third-party platforms.

Secondly, tool vendors may infiltrate downstream, building standardized teaching content and user communities to reduce dependence on top KOL channels.

The "influencer distribution model" stuck in the middle is ultimately a transitional product of the technological iteration window. From this perspective, AI courses are not a new story for knowledge monetization; instead, they represent the first battle in the AI channel war. After all, everyone wants to control user mindshare entries and industry pricing power.

Affordable AI Courses, Unaffordable AI Bills

49 RMB, 89 RMB, 99 RMB... The flood of AI practical courses creates the illusion for ordinary people that they can "enter the AI era at a low cost." Creators and course sellers continually instill the idea that a few dozen RMB can unlock AI creativity, but few users calculate the true costs in advance.

In reality, courses are merely one-time AI entry tickets. The real rigid expenses are the ongoing tool subscriptions, computing power credits, and model upgrade fees. After low-priced courses fulfill their lead generation mission, users discover that the daily usage costs of AI production far exceed the initial course fees, representing the true long-term expenditure.

Currently, AI applications are rapidly moving away from the free era.

For example, mainstream platforms in China have formed multiple "price tiers" for memberships. Doubao offers monthly subscriptions at 68 RMB (standard), 200 RMB (enhanced), and 500 RMB (premium); annual subscriptions cost 688 RMB, 2048 RMB, and 5088 RMB, respectively. Without an annual subscription, the premium package costs 6088 RMB per year.

Costs are even higher for video tutorials and short drama tutorials frequently featured in courses. According to data disclosed by ChineseAll, AI short dramas and video generation account for 55% of all Token consumption in China, far surpassing e-commerce and marketing (24%) and software development (15%).

According to CSDN, in March 2026, Seedance 2.0 announced its pricing: generating a 15-second standard video (720P, 24fps) consumes approximately 308,880 Tokens. At the standard rate of 46 RMB per million Tokens, the cost per video is about 15 RMB, or 1 RMB per second.

Although a 50-episode AI short drama, with a total runtime of 50 to 75 minutes, may seem to have a modest scope, the video generation process alone can burn through tens of millions of Tokens. For individual users operating solo, this expense is far from negligible. Even for teams, attaining the desired film quality frequently necessitates multiple rounds of adjustments and regenerations.

It's worth noting that a wave of AI price hikes swept across the market in 2026, with most mainstream applications in China opting to raise their prices.

As reported by Sina News, Tencent Cloud's Hunyuan HY 2.0 saw a staggering 463% increase in input prices in one fell swoop, while AI computing and storage products experienced maximum price hikes of up to 34%. Zhipu AI, not to be outdone, raised its prices three times in the first quarter alone, with core API prices soaring by 83%. The output price for GLM-5.2 reached 28 RMB per million Tokens, quadruple that of lightweight models.

Similarly, Alibaba Cloud discontinued its most basic Coding Plan package and increased the prices of AI computing services by 5% to 34%. MiniMax boosted the prices of its core models by 30% to 50%. Claude's security-enhanced Fable 5 output pricing hit 50 USD per million Tokens, double that of the Opus series. Seedance 2.0 even raised its prices three times in a single month.

Meanwhile, The Paper reported that Jimeng AI underwent a round of 'three consecutive price rises' in April 2026.

At a time when AI is once again fueling the growth of the knowledge payment sector, an increasing number of people are embracing this model and are even willing to up their spending.

In February 2026, the '2025 Annual Trend Insight Report on China's Knowledge Payment Industry' was released. The report's data revealed that from 2023 to 2025, the average amount spent per course purchase by knowledge payment consumers rose from RMB 84.37 to RMB 138.81, and the annual frequency of course purchases increased from 3.8 times to 4.7 times.

Consequently, many individuals perceive spending a few dozen yuan or less on AI courses as a great bargain.

However, once they actually dive into the practical application phase, they often find that the money spent on courses doesn't even cover a month's worth of tool usage fees.

And this is precisely the inconvenient truth that bloggers in this wave of AI course sales are least likely to volunteer.