"Hello, new app!"
Standing in the first year of the landing of the large model, it is time to shout this sentence.
From software apps, smart terminals to embodied intelligence, AIGC has begun to sweep everything.
Large model players, Internet giants, terminal manufacturers, and vertical scene players have entered the market one after another, and the office, creation, marketing, education, and medical fields have been infiltrated one after another...... A trillion-dollar market is brewing.
Standing at the starting point of the wave, what is the current development, where are the opportunities, and what are the challenges?
At the China AIGC Industry Summit held by qubits, an industry reference was officially released - "China AIGC Application Panorama Report".
It draws a comprehensive and three-dimensional application panorama for the current AIGC application market. Key takeaways include:
- The scale of China's AIGC application market will reach 20 billion this year and trillion yuan in 2030;
- 80% of B-end products achieve revenue, and nearly 50% of C-end products are mainly free;
- The proportion of AI native applications is higher than that of X+AI, accounting for nearly 57%;
- Multimodality is a major trend, with application products accounting for nearly 50%;
- Scenario and technology matching is the primary factor for product implementation, and technology maturity determines the implementation speed.
-
The investment resources of the AIGC industry are concentrated at the head, and the potential capital focuses on the application layer.
……
Also unveiled is the 2024 AIGC Enterprise & Product List worth paying attention to. They provide an objective perspective for understanding the current AIGC application market in China.
Let's take a look at the specific details.
What is the current situation?
The current situation of venture capital: resources are gathered to the head
First, let's look at investment and financing.
Last year, there were at least 5 representative financing events in the domestic AIGC circle, among which Zhipu AI received more than 2.5 billion yuan in financing in one year with Zhipu Qingyan, and Kimi intelligent assistant on the dark side of the moon attracted more than 1 billion US dollars in Series A.
It can be said that the cycle of becoming a unicorn is getting shorter and shorter.
On the whole, the investment in the AIGC industry is showing a directional shift: the snowball effect of investment and financing at the model layer is obvious, resources are gathered at the head, and potential capital is focused on the application layer.
Qubit Think Tank predicts that China's AI capital market will further converge to the top in 2024, and at the same time, we believe that more AICG projects that can be applied will have opportunities in the future.
Business model: The B-end realization is clear, and 80% of the revenue is realized
In terms of implementation, 50% of the existing business models are AIGC applications for the C-end.
Among them, the B-end products are evenly distributed from the general scenario to the vertical track, and the revenue model is mainly based on membership subscription and on-demand payment, and the business model is relatively clear, although the proportion (pure B-end) is only 31%, but more than 80% of the products have achieved revenue.
Although the number of users is large (more than 50% of the pure C-end products), the profitability is generally not optimistic, and nearly 50% of the products still have no clear revenue model, mainly free.
According to the analysis of the qubit think tank, this is mainly because the needs of C-end users are not clear, and it is often the supply that stimulates the demand, and the experience of the product itself is required to be higher, emphasizing "ease of use".
Here's the full business model map:
Application type: AI native > X+AI
AIGC application products can be divided into AI native and X+AI according to the type of AI application, and the former accounts for nearly 57%, which is larger than the latter.
The so-called AI native easy to understand refers to the creation of generative AI technology, and X+AI is to add generative AI-related technologies to the original products that do not contain AI.
According to the type of large model application, AIGC application products include three categories: self-developed basic large model, self-built vertical large model, and API access.
As shown in the figure below, it can be seen that products based on self-built vertical large models occupy the mainstream in the entire application layer.
These enterprises use their own data accumulation and technical capabilities to take the lead in finding the direction of AIGC implementation.
From a cross-cutting point of view, AI native products are mainly self-built and self-developed large models, while X+AI products are based on self-built vertical large models and feature API access.
Technology popularity: multimodality is the trend, currently accounting for 48%
The popularity of technology determines the maturity of the application to a certain extent.
At present, more than 44% of the modalities generated by AIGC applications focus on text generation, which is the most popular, followed by image generation at about 29%.
The proportion of audio generation is 6% higher than Sora's popular video generation, and the recent birth of two music generators, Suno and Udio, has indeed shown us the huge potential of AI audio (music) generation.
Finally, 3D generation is still a niche modality, accounting for less than 3% of the product market.
As an inevitable trend, multimodality (products that can identify and understand two or more modal data) has reached 48%, and there is still a lot of room for growth in the future.
If we look at the product presentation ability, analogous to the eight basic capabilities of human beings, the popularity of AIGC products is like this:
"Reading and writing" is the most popular, "seeing, drawing, listening, speaking" is developing rapidly, and "thinking and moving" are the future directions.
What is the player/product distribution?
Back to the present, there are seven major tracks worth paying attention to in AIGC applications.
According to the maturity of the technology, it can be divided into two categories:
- AIGC applications are relatively mature, with a clear business model and good revenue: marketing, retail, education, film and television, and office collaboration.
- At present, the technology is not mature, but it has great potential: gaming, medical care.
According to the product form, 90%+ AI applications are currently in the form of software, and AI hardware has begun to emerge in an endless stream, but it has not yet ushered in the "iPhone moment".
Hey 硬件代表品类中,"I+万物"包括I手机、I PC等;I新物种包括、Rabbit R1等.
According to the function attributes and the form of software and hardware, manufacturers and representative products are divided as follows. Products with productivity attributes account for more than 95%.
In terms of player distribution, according to the type of company, it can be divided into four categories:
- Internet companies
- AI2.0 company
- AI1.0 company
- ICT companies
The representative player graph, application type, and revenue are as follows:
Among them, the commercial revenue represents the company's product realization speed.
From a geographical point of view, Beijing, Shanghai, Hangzhou, and Shenzhen have the most AIGC application products.
The geographical distribution map of AIGC application products in China is as follows, and Beijing is dominated by Internet companies and AI2.0 companies, representing the most players.
What are the opportunities?
For the application development of AIGC, Qubit Think Tank divides the commercialization process into three stages.
- Phase 1 (2024-2027): Product landing
- Phase 2 (2028-2029): The business model matures
- The third stage (after 2030): large-scale profitability
Within three years, the player pattern of the model layer was basically settled, and the "AI native" products at the application layer began to grow explosively. The business model of the B-end market is clear, and the C-end market has shifted from customer acquisition first to profit first.
Starting in 2028, with the maturity of video generation and multimodal technology, it will further help the realization of AGI. The C-end business model is mature, and the head products will appear in companies with the ability to develop their own large models, and the competition for waist products is fierce.
After 2030, it will officially enter the AGI stage. AIGC products will have eight abilities of "listening, speaking, reading, writing, reading, reading, painting, thinking, and moving" at the same time, which can better simulate human beings. The head products of the application layer touch multiple industry scenarios, and the waist products focus on the vertical industry.
In terms of market size, the mainland AIGC application market will reach 20 billion in 2024.
It is estimated that by 2026, the scale of China's AI market will reach 100 billion yuan.
By 2030, the application of AIGC in mainland China will become a trillion-dollar market, with an average compound growth rate of more than 30% in five years (2024-2028).
Finally, the Panorama of China's AIGC Application 1.0 is attached.
In 2024, who is worth paying attention to?
Who will create such a trend and the future?
In the first year of the implementation of AI applications, a reference answer has been newly released - the results of "AIGC enterprises & products worth paying attention to in 2024" have been announced.
There are 50 AIGC companies and 40 AIGC products selected, and the full list is as follows.
We hope that this list can provide an objective perspective for understanding the application trend of AIGC in China, and that it can become a new footnote in the wave of AIGC revolution.