NeuroAI NEUROAINEUROAI.SITE
China AI, reported clearly
1,300+ verified stories on models, robotics, BCI and the business of AI.
Browse stories
ESC

AI Credits Inside Your Video Editor: How Wondershare Turned an Installed Tool Into a Habit

A Chinese software company's results show a quieter AI business model: sell small amounts of intelligence inside tools people already use.

2026-11-22 · 1122 words · NeuroAI
AI Credits Inside Your Video Editor: How Wondershare Turned an Installed Tool Into a Habit

A freelance video editor in Manila opens the same editing program she has used for years. There is a new button that removes a noisy background, and next to it a small balance of credits. She does not think of it as buying artificial intelligence. She thinks of it as finishing the job before dinner.

That little balance is becoming one of the most practical business models in generative media. Wondershare (万兴科技), a Chinese software company whose tools are distributed globally, reported in its 2025 annual results that AI-native applications (AI原生应用) brought in ¥130 million (≈ US$18M / HK$142M), up more than 90 percent, while paying users for those apps more than doubled.

Key takeaways

  • Wondershare reported 2025 revenue of ¥1.533 billion (≈ US$215M / HK$1.67B), up 6.46 percent, with its net loss narrowing substantially.
  • Revenue from AI-native applications exceeded ¥130 million (≈ US$18M / HK$142M), up over 90 percent, with paying users up over 100 percent.
  • AI server calls exceeded 1.3 billion during 2025, up more than 140 percent.
  • Instead of selling a separate AI subscription, the company puts AI credits (AI积分) inside existing products such as Filmora, PDFelement and UniConverter.
  • The significance is distribution: an AI feature inside a familiar tool reaches people who would never sign up for a standalone AI site.

The credits model, explained plainly

Most AI products ask you to visit a new website, learn a new interface and pay a monthly fee. Wondershare's approach is different. Its products already had millions of users who opened them to cut a clip or sign a PDF. The company adds AI functions to those products and meters them with credits: a small allowance, then a charge for heavier use.

The company's 2025 report, summarized in Chinese financial media, says AI credit revenue inside Filmora (万兴喵影) accounted for about 6 percent of that product's revenue, and about 10 percent inside UniConverter. In PDFelement, desktop AI credit revenue grew more than 80 percent. Those are not enormous proportions, but they are early numbers in products with large existing user bases.

Why does this matter? Because the hardest thing in consumer AI is not building a feature; it is getting people to change where they work. A credit balance inside a tool you already trust removes that step. It also changes the purchase decision from "is AI worth a subscription?" to "is this one export worth a few credits?" That is a much smaller question, and people answer small questions more easily.

What the company actually sells

Wondershare's portfolio is unglamorous by AI standards: video editing, PDF handling, mind maps, diagrams, file conversion. Its own materials say the products are used in more than 200 countries. That global footprint is the real asset. A standalone AI image generator has to compete with every other new AI site. A "remove background" button inside a video editor has a captive audience.

The 2025 results describe a dual-model approach: the company builds its own vertical audio-visual model, called Tianmu (天幕), while also plugging in outside models from OpenAI, Google, Alibaba and others. That is a sensible hedge. A user who wants a particular visual style can get it from whichever model performs best, while the company keeps the interface and the customer relationship.

The same report says research and development spending reached ¥485 million (≈ US$68M / HK$529M), about 31.65 percent of revenue. For a company of this size, that is a heavy bet on keeping the tool stack current.

Comic drama as a test case

One of the more revealing products is Wondershare Drama Factory (万兴剧厂), a tool aimed at producing comic drama (漫剧) — a serialized, often illustrated short-form format adjacent to the broader microdrama (短剧) boom. The company says the product supported thousands of episodes after launch and that weekly AI credit consumption rose 63 percent between the first and last week of its first month after the March release.

That number is a company claim and a small sample, but the logic is sound. Comic drama needs many consistent characters, many scenes and constant iteration. It is exactly the kind of workflow where a creator would happily spend credits on a specific task — a new background, a changed expression, a cleaner voice-over — rather than pay for unlimited access to a general chatbot.

Wondershare also made strategic investments upstream and downstream: a significant investment in video-model developer Shengshu Technology (生数科技) and a smaller one in a comic-drama production team. That is an attempt to control both the model that generates the material and the workflow that packages it.

The honest reading of the numbers

A 6 percent or 10 percent AI credit share is not a transformation. It is an early signal that users will pay for embedded AI in small amounts. The company still made a net loss in 2025, though it narrowed sharply. The global tools business remains the base; AI is currently an add-on that improves the economics of that base.

There is also a question worth holding onto: credits are easy to sell when they feel like a convenience and hard to sell when users perceive them as a tax on a feature they used to get free. A product that meters every minor function can annoy the same users it hopes to convert. The better versions of this model give a generous everyday allowance and charge for genuinely expensive work, such as long video generation.

For the wider industry, Wondershare's results suggest that the consumer AI race will not be decided only by model benchmarks. Distribution, billing and habit are equally important. A mediocre feature in a tool people open every day may earn more than a brilliant feature buried in an app nobody remembers to launch.

Honest limitations

This is a business-model deep read built from the company's 2025 annual report as summarized by Chinese financial media. Financial figures, product percentages, user-footprint claims and the comic-drama growth rate are company-reported and have not been independently audited here. No claim is made about the quality of the company's models relative to competitors, and the article does not reproduce any ranking or superlative from vendor marketing. It also does not assess the legal position on training data or generated content in any specific market.

What readers can do now

  • If you already pay for a creative tool, check whether your subscription includes AI credits before buying a separate AI service.
  • Try one metered feature on a real task and compare the credit cost with the time it saves; that is the only conversion rate that matters to you.
  • For creators, consider formats with many small repeated tasks, such as comic drama (漫剧); they tend to get more value from credit-based tools than one-off projects do.

Related coverage

Cite this article

NeuroAI (脑机网), "AI Credits Inside Your Video Editor: How Wondershare Turned an Installed Tool Into a Habit," 22 November 2026, https://neuroai.site/a/na-aigc-wondershare-ai-credits

Machine-readable: BibTeX · RIS · JSON · Markdown

More in “Generative Media” → · Back to home · Markdown version

China AI, reported clearly
1,300+ verified stories on models, robotics, BCI and the business of AI.
Browse stories