Largest financing in the industry: A former ByteDance veteran ventures into AI video, securing a 430 million yuan investment from Alibaba with over 100 million users.

The AI-generated video track has recently broken a record: Aishi Technology completed a financing round (Series B) of over $60 million, led by Alibaba. This financing (approximately 426 million yuan) has broken the record for the largest single-round financing in the AIGC video field (domestically).

So, the question arises. What kind of super opportunities are hidden behind this financing that made Alibaba invest so heavily?

- 01 -

The founder of Aishi Technology is Wang Changhu, an 80s generation (born in 1982) from Benxi, Liaoning. In 2009, he graduated from the University of Science and Technology of China (with a doctorate) and joined the Microsoft Research Asia, where he worked for eight years.

Until 2017, he switched to ByteDance, serving as the director of the AI Lab and leading the construction of the video AI capabilities of Douyin and TikTok from scratch.

From this experience, Wang Changhu's background is focused on AI video. At the University of Science and Technology of China, his research directions included computer vision, video understanding, multimedia retrieval, and machine learning; at ByteDance, he also specialized in AI video.

However, he then switched from ByteDance to Longfor Group. This experience had a little episode: he served as the general manager of the AIoT engine team, responsible for digital and intelligent transformation.

It wasn't until April 2023 that he founded Aishi Technology, returning to the AIGC video track.

In terms of product strategy, Aishi Technology first targeted the overseas market and then the domestic market. In January 2024, Aishi Technology launched its overseas product, PixVerse. It is a tool that uses AI to create videos: users input photos or text, and PixVerse generates high-quality videos.

In June 2025, Aishi Technology launched "Paiwo AI" - it can be regarded as the domestic version of PixVerse.

This is a highly competitive track with many giants around. Not to mention international giants like Runway, Google, and OpenAI abroad, just looking at the domestic leading players, the competition is tough enough for entrepreneurs: ByteDance has CapCut and Jimeng, and Kuaishou has Keling.

Of course, ByteDance is Wang Changhu's former employer. Historically, there are many stories of entrepreneurs leaving their former employers to start a business and then surpassing them. The probability is low but not uncommon.

From the perspective of entrepreneurial experience, for a startup in such a track, it's best to find a strong backer. Sure enough, in the recent financing round, Aishi Technology got the support of Alibaba.

Although Alibaba itself has also made arrangements in AI video. For example, Tongyi has launched the video large model "Tongyi Wanxiang", and Alibaba Marketing has released tomoVideo, entering the field of AI video for e-commerce marketing. But for innovation, it's best to carry out in-house incubation and external innovation simultaneously, and equity investment is essential.

For Aishi Technology, the ecological resources behind Alibaba are also a good medicine to compete against other giants, including but not limited to computing power resources, AI technology, cooperation in application scenarios, and financial support.

- 02 -

There are several ways to make money in the video AIGC field. Taking different paths means different outcomes. One is the ToC model, and the other is the ToB model.

How to make money in the ToC model? For example, through subscription memberships, which cost 98 yuan per month (see the table below for details).

How to make money in the ToB model? For example, by selling SAAS, video production services, industry solutions, etc. (see the table below for details).

How does Aishi Technology make money? It takes a two-pronged approach.

One is the ToC subscription membership service: it is said that its global user base has exceeded 100 million, and the subscription revenue has covered the costs.

The other is the ToB service. At the beginning of 2025, Aishi Technology launched its B2B business, focusing on fields such as advertising, short dramas, and games. The specific service forms are API and customized video generation services.

For example, in the game track, it can provide solutions for game companies, such as character action design and batch generation of scene assets; in the advertising field, it can produce SKU customized advertisements and real human voice-over materials for enterprises.

These two directions require completely different corporate genes. Aishi Technology's two-pronged approach is a bit of a trial-and-error strategy.

The ToC subscription model tests the product ability and the ability to acquire customers and invest in traffic. The competitors are all giants. More specifically, Aishi Technology's product experience must surpass that of giants like ByteDance; at the same time, in terms of customer acquisition through traffic investment, it cannot be inferior to giants like ByteDance.

This tests the team's R & D ability, product ability, and financing ability, and each link is not easy.

The ToB model also has insurmountable challenges.

Previously, Pencil News interviewed Luo Jiangchun, the founder of Yilan Technology. The main challenges of the ToB model can be summarized as follows.

1. Deviation between customer perception and requirements: "Want results, not tools."

The core goal of ToB customers is to "achieve business indicators through content" (such as customer acquisition, activation, and revenue increase), rather than "using AIGC tools."

More specifically, customers only care whether the finally delivered content can meet the KPIs (such as e-commerce customers focusing on video conversion rates and bank customers focusing on user usage time), and they don't care whether the content is generated by AI or manually produced. Therefore, the technological advantages of AIGC are difficult to directly become the motivation for customers to pay.

2. There are bottlenecks in technological evolution.

It is difficult to "efficiently generate videos" in the short term. At least in the next 3 - 5 years, it is impossible to generate content that satisfies customers with one click.

From the initial draft generated by AI (such as hundreds of pictures and multiple versions of scripts), to the screening and optimization by professional teams (refining pictures and rewriting scripts), and then to multiple rounds of review and delivery, each link requires human judgment, and it is difficult to achieve the ultimate goal of "cost reduction and efficiency improvement."

3. "Non-standardized" customer needs: high matching costs and great difficulty in scaling up.

The needs of ToB customers are characterized by "personalization and scenarization." AIGC has difficulty covering them through standardized tools or services, making "scalable replication" a core problem.

4. Chaotic competitive landscape: squeezed by large companies + industry involution, limited survival space for startups.

There are many players in the track (including startups and large companies), but there is no "absolute leading player." Some enterprises compete for customers by "reducing prices and providing free additional services," squeezing the profit margins of the entire industry; at the same time, many players "enter and exit quickly," resulting in unstable market demand and service quality.

- 03 -

From global practical cases, some potential profitable tracks have emerged.

1. The virtual human direction.

Synthesia is a UK-based company specializing in AI virtual human video production. Through face and voice synthesis technology, it creates new video expression methods for marketing, training, and media communication. According to a report in April 2025, when Synthesia received strategic investment from Adobe, its ARR had exceeded $100 million.

2. AI video generation and creative tools.

Runway is a comprehensive US platform focusing on AI video generation and creative tools, providing a full set of solutions from image-to-video conversion, multi-scene special effect generation to real-time mobile editing. According to The Information, in December 2024, Runway's estimated monthly annualized revenue reached $84 million.

3. Video advertising.

Tencent Advertising performed strongly in 2024, with its annual advertising revenue reaching 121.4 billion yuan, a year-on-year increase of 20%. Its AIGC creative platform, Miaosi, has improved the production efficiency and click-through rate of advertising materials through functions such as text-to-image, image-to-image, and digital humans.

In these cases, the B2B profit model stands out. Enterprises such as Synthesia and Runway have achieved significant revenue by providing enterprise-level AI video solutions, such as AI digital human video SaaS. Synthesia's ARR reached $150 million, and Runway's ARR was also $80 million.

This article does not constitute any investment advice.

This article is from the WeChat official account “Pencil News” (ID: pencilnews), written by Aiyu, and published by 36Kr with authorization.