36氪 was informed that “TIAMAT”, the first generative AI platform in China, has completed a multi-million dollar angel round of financing, and the investor in this round is DCM. Vision Capital acted as exclusive financial advisor to the company’s financing.
“TIAMAT” is a young generative AI platform established in 2021. As the first batch of technical teams in the field of AI generation in China, it is committed to providing individual users and corporate customers with services in the form of creative production, advertising, digital collections and other scenarios through self-developed algorithms, underlying engines, and platforms. It has only been established for one year. Time has accumulated considerable community users and business customers.
Globally, the AI-generated heat continues. The image generation platform featuring Text-to-image generation mode and Diffusion model is subverting the traditional fields of content creation and art design. 36 Krypton has also recently combed the industry mapping of the global AI generation field.
From the original pen and paper palette to the digital tablet, drawing software, iPad, and then to the AI era, anyone can create high-quality works of art with just a few lines of text. Being lowered, the industry’s attitude towards AI generation also shows a great contrast. For example, in the protests of some artists, most users who have experienced the product believe that AI generation has set off a wave of art democratization.
“Does everyone have the right to use simple tools and technologies to unleash their creativity and express what they want to express?”
The answer is obvious. As the first batch of entrepreneurs to step into the field of generative AI in China, Qinggan, the founder of “TIAMAT”, told 36氪 that AI generation is not creating a new demand, but using breakthrough technology to solve the efficiency problem of creation .
AI-generated content: Data volume changes under creative efficiency
The early term in the field of Text-to-image is called the application of multi-modal technology, but since the second half of last year, the conversion and application between multi-modality has become the point that the entire industry wants to break through. Based on this technical environment, “TIAMAT” began to explore how to make better products and landing scenarios.
In the first half of this year, “TIAMAT” began to release Chinese Text-to-image products on social media such as Xiaohongshu and Douyin for internal testing. In terms of time difference, it was in the same period as the overseas AI generation boom.
“TIAMAT” generates images
With the explosion of imagination, the issue of commercialization has become the biggest constraint on AI generation that may fall into rational recovery for a period of time in the future. For example, in the domestic market, different from overseas C-side payment habits, it is determined that few individual users are willing to continue to pay for creative efficiency. Therefore, although this field still exists as a “black technology” on social media, ToB is widely used. Taking into account is also inevitable. In this regard, “TIAMAT” has a very positive attitude.
“The enterprise service market definitely exists. For example, the top 10 game companies, advertising companies, and well-known IPs that we are deeply cooperating with are all looking for us to cooperate continuously.”
Although it is still in the community beta test stage, the commercialization confidence and exploration of “TIAMAT” is tantamount to illuminating a part of the way for the capital that is still waiting and the technology entrepreneurs who want to enter the game. At present, “TIAMAT” has established long-term cooperative relationships with first-line fashion magazines such as well-known sci-fi IP, CCTV, Harper’s Bazaar, and Marie Claire, and has cooperated with many leading game companies, especially those holding Forbidden City Culture-related IP Intangible cultural heritage IP agency Zhizao Bureau and others have carried out strategic cooperation. The content generated by AI through it is widely used in the dissemination and display of urban public environment and public transportation.
As an overlapping part of the entire image market and AIGC (AI Generated Content), Text-to-image is highly scalable. Take the image and content production industry as an example: at present, the global stock image market size is close to 300 billion US dollars, but even for paid materials, it is often “use whatever users the platform has”, rather than “whatever users want”. what” needs. This also leads to the fact that the traditional picture industry is still a “digital handicraft industry”, which means that from large enterprises to self-media bloggers, all the content produced is a little bit of hard work. “Seven posts a week, which is what we call the daily update, is already quite fast.”
But with the aid of AI capabilities, the background data of “TIAMAT” shows that a single user can generate 3,000 to 10,000 images in a month. Quantitative changes of orders of magnitude may lead to qualitative changes, and future content production and content distribution methods will be further changed due to the explosion of data, and this is what “TIAMAT” wants to do from the technical side.
“TIAMAT” generates images
“The creative work that many people understand is the production of ideas itself, but in fact, most practitioners are doing repetitive, mechanical and meaningless work, which suppresses their creativity.” And this, she herself The original intention of stepping into the AIGC field: even if you are tired of real life, you want to use AI to turn your spiritual needs and creative inspiration into reality.
Community strategy: better understand Chinese, develop simultaneously with overseas
Different from the previous product model that only required a landing page and a download/experience button to release, in recent years, digital products in all walks of life have increasingly required users’ “participatory design”. The community is like the badge that every AI company has on its chest. It is increasingly clear that in a fair environment of open source technology, community activity and richness determine the success or failure of competition.
In the field of AI generation, Midjourney built on Discord and Stable Diffusion loaded the test version on the AI community Hugging Face are the best proofs. In the emphasis on community resources, “TIAMAT” is no exception.
“Because what AI does, unlike software like Photoshop, it can be done once and for all as long as it is online.”
This is also the reason why AI companies have a more obvious first-mover advantage, and it is difficult to compete once they miss the window. Many underlying technologies are open source, but an AI platform like “TIAMAT” needs to constantly observe users’ behavior every day, such as what kind of images they want to generate and what kind of sentences they want to use to describe them. Popular art styles, Internet contexts, and frequently-popular buzzwords are constantly changing, and the fluidity of culture is bound to be reflected in the adjustment of data and models. Therefore, launching a product may only be on the starting line, and it is more important for the community’s real-time feedback and understanding of needs.
“20% of the user groups are small B-end designers, 60% are people in the general design industry, and the rest are people who are more interested in new technologies and new data.”
No matter which part of the group is cut out, the main body of the community of “TIAMAT” is the young people after 95-05. They have a strong sense of identity for things such as creativity, aesthetics, IP, and scene interaction under the metaverse. At present, “TIAMAT” is committed to incubating AI Art’s Kol and profitable users through the community, and continues to cultivate the user habits of the Chinese Internet, thereby triggering the breaking circle effect.
Compared with the most popular overseas Text-to-image services Stable Diffusion, DALLE2, Midjourney, etc., the model and community advantages of “TIAMAT” are mainly reflected in two points:
- Continue to accumulate in Chinese natural language processing, and plan to expand to other Asian markets and languages in the next stage;
- It can provide images that are more suitable for the needs of domestic users, and locate core users in professional groups and business behaviors.
At present, on the product side, “TIAMAT”, which is working hard to accumulate users and build a community, has achieved thousands of active users in the internal test and tens of thousands of new monthly users. The community users cover domestic design, art creation, and multi-modal technology. Wait for the top bloggers; at the same time, more than one-third of users have the willingness to continue to pay, and the C-side commercialization capability is considerable.
In terms of team, “TIAMAT” was incubated from Shanghai University of Science and Technology, Chinese Academy of Sciences. The core members of the team are from the School of Information, Shanghai University of Science and Technology, Chinese Academy of Sciences and the University of the Arts, London. Rich experience in project planning and community operation.
“A great technology company will not be separated from art, like Apple, because the end of science is a kind of art of technology, and this characteristic is most obvious in the field of AI generation, which is both future-oriented and It belongs to the integration of localized humanities and social sciences and technology.”
“TIAMAT” generates images
Up to now, the exposure of “TIAMAT” on the entire network in the past 4 months has reached 50 million, and this data is still rising. The new model supports the generation of 100,000-140,000 pictures per day, and the unit generation speed is controlled within a few seconds to two. Within minutes, you are ready for “quantitative changes in data lead to qualitative changes” at any time.
Talking about the future of generative AI, Qinggan said frankly: the future is conceivable and unimaginable. In the future envisioned by “TIAMAT”, AI will redefine the possibility of creation and release human imagination and creativity.
————————————
Lin Xinhe, Founding Partner and Managing Partner of DCM China, said, “Generative AI provides an accurate, efficient, professional and easy production method for people’s active expression, further unlocking the potential of content production. It can be seen that AI It will be involved in more fields in the form of basic technology and become the driving force for great changes.”
media coverage
36Kr , a new budding entrepreneurial state in the investment world
Related events
- TIAMAT completes multi-million dollar angel round of financing2022-10-12
- Luchen Technology Completes $6 Million Seed and Angel Round Financing2022-09-01
- Yuanchu Smart received tens of millions of US dollars in A+ round of financing2021-12-24
- 99 companies raised more than 45 billion yuan in 100 days, the overseas AI boom continues, and the four directions are the most popular 2021-11-11
- 37 Degrees Data received tens of millions of Pre-A round financing2021-01-20
This article is reprinted from: https://readhub.cn/topic/8k2PxfS6Kxa
This site is for inclusion only, and the copyright belongs to the original author.