Baidu, SenseTime, and Zhipu are the top three. IDC releases the first report on the market share of large model platforms and applications

On August 21, International Data Corporation (IDC) today released the first edition of ChinaLarge model platformMarket share, 2023: The first year of large models - the initial stage. Data shows thatThe market size of China's large model platform and related applications will reach RMB 1.765 billion in 2023.

The report states that in the past year, the industry has made more early investments in big models, and even waited and watched rather than invested heavily, so the overall market size in 2023 is not significant; and the market structure is still mainly composed of early investors such as Baidu, SenseTime, Zhipu, Baichuan, etc. In 2024, the leading Internet companies increased their investment in big models and launched a price war, which brought certain competitive pressure to early big model startups.It is expected that the market structure will undergo multiple rounds of major changes in the next 2-3 years..

The report shows that the market share of China's large model platform and application in 2023 is as follows:

  • 1. Benefiting from years of heavy investment in the field of AI and early investment in large models, Baidu Smart Cloud's large model market size will reach RMB 350 million in 2023.Ranked first in the market, with a market share of 19.9%;
  • 2. SenseTime also benefits from its focus on big models and its Shanghai AIDC, which can provide customers with integrated big model computing and application solutions. It ranks second in the market with a market share of 16.0%.
  • 3. Zhipu AI is the winner among startups in 2023, ranking third in the market;
  • 4. Next are Baichuan Intelligence, Fourth Paradigm, etc.

Baidu, SenseTime, and Zhipu are the top three. IDC releases the first report on the market share of large model platforms and applications

Looking ahead, the report states that although big models and generative AI are the biggest focus of today's technology market, the actual capabilities of big models are still relatively low, and the big model capabilities of different technology providers also vary to a certain extent. Failure to continuously invest in foundational big models can easily lead to companies falling behind and being eliminated in the next round of model capability competition.

The report also mentioned that the innovation of large language basic models has reached the convergence stage.Multimodal large models and video large models have wider application scenarios, but the technology is not yet mature, and the market needs such products.In addition, not all data types are text or image vision types. In the field of scientific computing, the data types are special and the current large models may not be applicable. A completely new base large model is still needed.

statement:The content is collected from various media platforms such as public websites. If the included content infringes on your rights, please contact us by email and we will deal with it as soon as possible.
Information

Anthropic accused of using pirated books to train AI, authors file class action lawsuit

2024-8-22 9:15:14

Information

Meta deploys new web crawler bot to collect massive amounts of data for its AI models

2024-8-22 9:17:21

Search