CTO-founded AI Digital Human Startup "Vector Equation" Secures Nearly 10 Million Yuan in Angel Round Financing | 36Kr Exclusive

Written by Yuan Yingliang

Edited by Deng Yongyi

"Intelligent Emergence" has learned that the intelligent digital human platform developer "Vector Equation" has previously completed a nearly ten million yuan angel round of financing, led by Zhencheng Investment, with Beijing Jixin Management Consulting and Shanghai Angel Club as co-investors. The funds will be used to enrich the research and development of digital human product technology.

"Vector Equation" was established on March 14, 2024. The founder & CEO, Shen Renkui, is the former CTO of Dedao / Luo Ji Siwei. He has previously worked at Tencent and Baidu, and the co-founding team has Internet R & D experience in Baidu, Meituan, and other companies. The product "Pomegranate Digital Human" is a one-stop AI digital human video creation platform focused on Asians, and it began commercial operations in June this year.

Short videos have long become the king of traffic acquisition, and AI digital humans add more fuel to the fire. The overseas AI video generation company Heygen's annualized income rapidly grew from one million US dollars to 35 million US dollars within 14 months. It is estimated that in China, the core market size of virtual digital humans will reach 48.06 billion yuan by 2025, and Tencent, Alibaba, ByteDance, etc. have also entered the game.

Shen Renkui disclosed to "Intelligent Emergence" that he had the idea of making digital humans as early as four years ago, but he was just waiting for the arrival of the technological inflection point.

"When I saw the new architecture of digital humans proposed in a paper, I realized that this is the commercializable technology I have been waiting for," he mentioned. "In the past, it usually took at least one day to collect and model data, but now a digital human can be quickly generated in 3 - 5 minutes."

The digital human track is moving closer to the large model from the previous generation of technology stacks such as 3D engines, and the production efficiency of digital humans has been greatly improved - Even the effect of a digital human made for a few hundred yuan at a low price is stronger than that of one worth more than one million yuan before.

In Shen Renkui's view, Because the company has less technical debt, being "new" is an advantage. And in the competition of giants, the opportunities for a new generation of entrepreneurs still exist. The main track of giants is information distribution, not information production, and even if they enter the game, it is difficult for them to monopolize standardized products and services.

"Pomegranate Digital Human" is a typical information production product that can convert text information into digital human videos to improve the efficiency of content creation. In terms of the picture, its simulation degree is high, and it can 1:1 reproduce the characters, scenes, clothing, and actions of real human videos. The high-quality underlying model trained with high-quality data can promote tasks such as lip-syncing in different languages and significantly reduce the required amount of data.

Compared with the current leading video generation manufacturers, The time required for "Pomegranate Digital Human" to record a video is shorter, significantly reduced from the previous 30 minutes to 30 seconds. At the same time, "Pomegranate Digital Human" is more adapted to the Chinese environment, shows advantages in dynamic scenes such as walking outdoors, and can also realize the interaction of multiple digital humans in the same picture.

Pomegranate Digital Human who can play basketball, ride a bike, and speak multiple languages

In terms of sound, the mechanical sound problem of TTS (Text To Speech) has been solved, and the current sound is more real, natural, and with cadence. The self-developed high-end version of the sound with a price of thousands of yuan is comparable to the industry's hundred-thousand-level, which can be personalized to customize the accent and pronunciation habits, with higher quality and better adaptation to the scene.

Shen Renkui introduced to "Intelligent Emergence" that "Pomegranate Digital Human" has achieved a fully automated customized digital human in the entire process. On the one hand, when users record videos, they do not need to pay attention to the details of lip-syncing, and the face twisting angle does not exceed 30 degrees. On the other hand, the system can handle the mixed arrangement of Chinese and English and complex digital scenes, and achieve a natural and smooth output through intelligent sentence breaking and context analysis. Because the cost of manual intervention is removed, the platform only charges according to the video generation time.

For high-paying customers, "Pomegranate Digital Human" will also provide an AI assistant to achieve interactive functions such as capturing subtitles, rewriting, and generating videos in the WeChat conversation window.

Currently, many domestic products focus on the creator ecosystem, and "Pomegranate Digital Human" has also launched a one-click video creation service, which is the most widely used scene at this stage. However, Shen Renkui believes that the opportunity for the enterprise-level market is greater, and the market is not yet saturated, which is the direction that the company pays more attention to.

In this track, "Pomegranate Digital Human" provides innovative video solutions for enterprises, helping enterprises quickly generate a large amount of video content, improve operational efficiency, and accelerate effect optimization with the help of digital humans and automated technology.

In addition, "Pomegranate Digital Human" plans to expand into the field of interactive videos, allowing digital humans to interact with the audience in real-time, not limited to static displays.

Interactive videos are not the same as live broadcasts. Live broadcasts are only one of the application scenarios. Although digital human live broadcasts are the future development direction, Shen Renkui said that he is still waiting for the technology to further mature.

"The core challenge of digital human live broadcasts is not only in technology, but also in the deep understanding of industry needs. The key is to quickly extract operational industry experience and transform it into a user-friendly product experience," he added.

Currently, "Pomegranate Digital Human" has successfully achieved commercial operation. In the future, the company will continue to optimize product functions, expand the market layout, and attract more outstanding talents to promote further development.