07/30 2026
366
In February 2024, OpenAI's AI text-to-video model, Sora, took the tech world by storm, sparking renewed interest in the previously under-the-radar AI video generation sector.
AIShige, too, experienced rapid growth amid this surge.
Over the past few years, under the leadership of Wang Changhu, the former head of AI Lab Visual at ByteDance, the company has refined its large models and launched PixVerse, an AI video generation platform that has amassed over 100 million users worldwide, achieving commercial success last year.

Image source: PixVerse
However, the benefits of this wave were short-lived.
Today, with more players entering the AI video generation arena, AIShige finds itself embroiled in fierce competition on all fronts.
As Seedance and Kling AI continue to expand their user bases in both B2B and B2C markets, and competitors like Alibaba's Wanxiang and MiniMax close in on image quality and commercialization, the question remains: How can AIShige maintain steady progress?
1
Betting Big on AI Video Generation
Before venturing into AI video generation, AIShige had already laid a solid foundation.
Founder Wang Changhu spent over a decade at Microsoft Research Asia (MSRA), researching "Magic Pen Ma Liang" sketch search technology. He later joined ByteDance as the AI Lab Director, focusing on computer vision, video understanding, and machine learning.

Wang Changhu (file photo)
Wang's experience at ByteDance was instrumental in establishing AIShige.
According to official introductions, Wang joined ByteDance in early 2017 and played a pivotal role in developing video products like Douyin and TikTok from scratch.
During this time, he led the construction of visual technology and business platforms, applying relevant technologies across ByteDance's product lineup. He provided AI technical support and business solutions for various group products, covering the entire lifecycle of visual content production and consumption.
After ChatGPT sparked the AIGC entrepreneurial boom, Wang leveraged his experience with large-scale data processing and deep learning algorithms at Douyin and TikTok to plan his venture in AI video generation.
In April 2023, Wang assembled a startup team.
While language large models and text-to-image were the industry's mainstream at the time, Wang insisted that video content was closest to users. He focused his venture on AIGC's visual multimodal algorithm platform, covering various vertical application scenarios related to vision.
According to Wang's initial vision, AIShige would initially focus on B2B applications.
The algorithms on its visual multimodal algorithm platform could not only generate content but also improve the efficiency of screening and intercepting non-compliant content. Like ByteDance before, with Wang's technical support, it replaced costly manual reviews while enhancing efficiency.
Additionally, the algorithm could reduce the distribution of similar or identical videos on the user side in cases of plagiarism.
However, in practice, this plan faced obstacles. The company's current development strategy prioritizes B2C before B2B and overseas markets before domestic ones.
In its early stages, AIShige also faced financing challenges.
Wang mentioned in a speech that during the startup phase, 99% of investors were skeptical about the video generation sector, believing it couldn't materialize within five years.
This meant that after Sora became a hit, AIShige, wanting to invest more in large model training, could only attempt it once due to limited cash flow.
However, with continuous technological breakthroughs, AIShige has now become a leading AI video generation company both nationally and globally. But at the same time, the company began facing new challenges.
2
Impressive B2C Traffic, But Commercialization Remains Elusive
Pursuing a dual strategy of B2C and B2B development, AIShige launched three categories of PixVerse models. The V series is for general user video generation, the C series for cinematic-quality models, and the R series for real-time world models.
The V series was the earliest to gain traction.
Six months after AIShige's founding, it released PixVerse V1, the industry's first AI video generation model capable of producing 4K-quality videos. In 2024, it launched the web version of the PixVerse AI video generation platform, ahead of Sora.
However, its performance fell short of Sora's. The product's launch ignited the AI video generation industry while subjecting domestic AI video companies to external skepticism.
Facing competitive pressure from Sora, AIShige focused on large model training and performance optimization, setting short-term goals like improving generation speed and quality while establishing feedback mechanisms to verify results.
Subsequently, AIShige maintained high-frequency model iterations, upgrading every few months.
In October 2024, after upgrading its large model to PixVerse V3, the PixVerse platform launched a "Transformation Special Effects" template, allowing users to upload photos to generate cinematic-quality character transformations.

Image source: PixVerse
This feature gained significant user approval, with generated videos going viral on domestic and international social platforms. According to the
Afterward, the company continued to seek technological breakthroughs, iteratively developing large models capable of "generating high-quality 360p videos in 5 seconds," "generating 1080p videos in 30 seconds," and "achieving synchronized audio-visuals," attracting user growth.
AIShige's business scope also expanded from overseas to domestic markets, launching "Paiwu AI" (domestic version of PixVerse) in June 2025.

Image source: Paiwu AI official website
Official data shows that PixVerse now has over 150 million global users, but its commercialization progress lags far behind user growth.
AIShige began commercialization in November 2024. When disclosing its B+ round financing in October 2025, the company revealed it had surpassed 100 million global users, with product MAU exceeding 16 million and ARR (Annual Recurring Revenue) surpassing $40 million, representing over a 10-fold increase in revenue in less than a year.
In response, the
AIShige hasn't placed all its bets on the B2C business.
Earlier this year, the company launched the general-purpose real-time world model PixVerse R1, along with phased exploration and the release of its first real-time video game engine, PixVerse Game. In April, it officially released its first large model for the film and television industry, PixVerse C1.
With these two model series, AIShige successfully entered the B2B market.
Currently, the company's B2B business has penetrated various industry scenarios, including internet, e-commerce, film and television, and gaming, collaborating with companies like Alipay, Baidu, Tencent, Ruyi, Zhangyue, and 37 Interactive Entertainment.
From a financing perspective, capital holds high expectations for AIShige's future prospects.
Since PixVerse's user base grew, the company's financing frequency increased. Its Series C funding round in March this year set a record for the largest single financing in Asia's AI video generation sector.
Public information shows that this round was led by CDH Hong Kong Fund, CDH VGC, and CDH Baifu, with participation from China Ruyi, 37 Interactive Entertainment, Yizhuang State Investment, Zhongwei Capital, and Fosun Rui Zheng, raising $300 million (approximately RMB 2.06 billion).

Image source: AIShige official WeChat account
A few months later, the company secured Series C+ funding led by Alibaba, with over ten institutions, including Lollapalooza Capital (Wang Huiwen's family office) and Ivy Capital, participating. Combined with the March funding, the total reached RMB 2.98 billion.
However, amid intense competition in the AI video generation industry and ongoing exploration of commercialization paths, whether AIShige can develop as capital expects remains uncertain.
3
Facing Competitive Pressure from Seedance 2.0 and Kling AI
Currently, the AI video generation sector has entered the second half of a dual race in technology and commercialization, with domestic large models rising collectively.
MiniMax's Conch AI, Alibaba Cloud's Wanxiang 2.6, Volcano Engine's Seedance 2.0, Kuaishou's Kling AI, and others have sequentially launched new-generation video large models, rapidly approaching the global first tier in terms of image quality, consistency, and long-form generation. Seedance 2.0 and Kling AI, in particular, are pressuring AIShige on both B2B and B2C fronts.
In February this year, ByteDance released Seedance 2.0, gaining widespread market recognition.

Image source: Seedance
According to the
The
Currently, Seedance 2.0 is being implemented in various fields, including film, television, entertainment, short dramas, and advertising. It has previously collaborated with this year's Spring Festival Gala, Jiuzhou Culture, Malt Media, and Kuaizi Technology.
On the B2C front, Seedance models are deeply integrated with ByteDance's AI product entrances like Jimeng and Doubao, leveraging super apps like Douyin and Hongguo Short Dramas to form a closed-loop ecosystem of "technology-application-user." In contrast, AIShige only launched the Paiwu AI platform last year and has yet to build a similar ecological barrier.
Another major competitor, Kling AI, also poses a threat to AIShige.
According to the latest rankings in February 2026 from Artificial Analysis.ai, a globally renowned independent AI benchmarking and analysis platform, in the global image-to-video (no audio) sector rankings, AIShige's PixVerse V5.6 ranked fourth with an ELO composite quality score of 1291 (derived from blind testing duels), trailing behind Elon Musk's xAI's grok-imagine-video, Kuaishou's Kling (Kling AI)'s Kling 2.5 Turbo 1080p, and Kling 3.0 Omni 1080p(Pro).

Image source: Artificial Analysis official website
Commercialization further confirms Kling AI's competitiveness.
As of 2025, Kling AI had surpassed 60 million global users, with annual revenue exceeding RMB 1.04 billion. In December 2025, its monthly revenue surpassed $20 million, corresponding to an annualized revenue run rate (ARR) of $240 million.
In contrast, as of October 2025, AIShige's ARR exceeded $40 million, with 100 million users at the time, 40 million more than Kling AI.
From these comparisons, despite its larger user base, AIShige's per-user monetization capability lags significantly behind Kling AI's.
Besides the B2C front, Kling AI's B2B business is also gaining momentum. Over 20,000 enterprise clients, including Xiaomi, Amazon Web Services, and BlueFocus, have adopted Kling's API services, integrating them into industrial production workflows.
As more internet giants enter this sector and AIShige's "uniqueness" advantage diminishes, it can only break through by continuously iterating technologies to build barriers. Otherwise, with intensifying competition, the company may face growth pressures if it cannot sustainably erect barriers.
AIShige's current situation confirms one issue: in the AI video sector, where "burning money buys time," the gap between independent startups and big firms lies not just in funding and computing power but also in scenario entrances and ecological closed loops. As the traffic dividend from special effects templates fades, how AIShige retains B2C users and excels in the B2B market battle will determine the company's ultimate trajectory.