Chinese AI startup Shengshu’s image-to-video launch challenges OpenAI’s Sora

Even AI is getting on the US vs China competition.

Beijing-based Shengshu Technology is getting a step ahead of OpenAI’s Sora as it unveiled a new feature for its AI-powered tool, Vidu, which can now generate video by combining multiple images into a seamless sequence. 

This new capability lets Vidu synthesize scenes from three separate pictures — for example, transforming a photo of a shirt, a person, and a moped into a short video of the person wearing the shirt and riding the moped through a lively backdrop.

While OpenAI introduced its own video-generating model, Sora, capable of creating up to one-minute videos from text, that model has yet to be released to the public. Meanwhile, Shengshu is setting itself apart with Vidu, an image-to-video tool that is already commercially available and in use by businesses worldwide.

Shengshu’s Chief Technology Officer, Fan Bao, explained that Vidu’s key achievement lies in its “visual consistency,” allowing it to blend separate images into cohesive animations. “Very early on we pinpointed [visual consistency] as the problem, and wanted to solve it well,” Bao said.

Since its launch in April, Vidu has captured significant attention, including a viral moment on TikTok with lifelike videos that transform profile photos into animations of people hugging. The tool is proving popular with advertisers, animators, and other creative industries, and monthly usage per customer can range widely, from 100,000 to 1 million yuan (US$13,871 to US$138,711), according to Shengshu co-founder and CEO Jiayu Tang.

Shengshu is also tackling industry concerns around copyright and ethical use. Businesses can sign licensing agreements with artists to allow the AI to adopt specific styles, Tang said. To protect privacy, Vidu prohibits generating content with celebrity or “sensitive” images, and complies with global data protection regulations by deleting personal photo data post-use.

Founded just last year, Shengshu has strong backing, including from Baidu Ventures, Ant Group, and the Beijing municipal government. The startup’s infrastructure operates on rented cloud servers across China and internationally, marking it as a major emerging player in the AI video generation field.

While OpenAI has been touted as a pioneer in generative AI, the entry of strong competitors into the foray ups the ante for AI services that introduce breakthroughs in creativity and production – and it can only get better from here.

Share this Post:

Accessibility Toolbar