Training AI large models requires a large number of AI corpora, but due to insufficient data mining and inability to circulate freely in the market, high-quality Chinese data is still scarce, and AI corpus-related enterprises have received widespread attention.
In this paper, a total of 33 AI corpus companies are selected as research samples, and the return on net assets, gross profit margin, and net profit margin are used as evaluation indicators.
The data is based on history, does not represent future trends, is for static analysis only and does not constitute investment advice.
Top 10 AI corpus profitability companies:
10th Tors
Profitability: Return on equity 10.30%, gross profit margin 61.30%, net profit margin 20.80%
Main products: Big data software products and services are the main source of revenue, accounting for 40.11% of revenue and gross profit margin of 69.69%
Company highlights: The "Chinese Internet Corpus" jointly built by Tors and KLCII has been officially released.
9th Beyondsoft
Profitability: Return on equity 10.75%, gross margin 25.36%, net profit margin 6.70%
Main products: R&D engineering is the main source of income, accounting for 57.09% of revenue, and gross profit margin of 21.47%
Company Highlights: Beyondsoft provides corpus, data annotation, and data training services related to AI large model training for some customers.
No. 8 Newtouch Software
Profitability: Return on equity 7.33%, gross profit margin 28.07%, net profit margin 5.26%
Main products: IT solutions are the main source of revenue, accounting for 84.83% of revenue, and gross profit margin is 22.01%
Company Highlights: The R&D of Xinzhi Software products is mainly supported by the Xinzhi Xinzhi corpus and robot platform.
No. 7 Mango Supermedia
Profitability: Return on equity 15.64%, gross margin 34.48%, net profit margin 13.60%
Main products: Mango TV's Internet video business is the main source of revenue, accounting for 76.85% of revenue, and gross profit margin of 42.67%
Company Highlights: Mango Supermedia's rich copyright content data assets provide high-quality training corpus for the research and development of "Mango Model".
The 6th Chinese science and technology
Profitability: Return on equity 10.77%, gross profit margin 29.00%, net profit margin 18.31%
Main products: Publication import and export business is the main source of income, accounting for 52.23% of revenue, gross profit margin of 6.54%
Company Highlights: China Science and Technology has successively launched a number of digital products and knowledge service platforms such as "Science Library", "Science Think Tank" and "China Biological Library".
5th Desheng Technology
Profitability: Return on equity 10.56%, gross profit margin 44.95%, net profit margin 12.98%
Main products: all-in-one card and AIOT application are the main sources of revenue, accounting for 80.03% of revenue, and gross profit margin of 48.57%
Company Highlights: Desheng Technology's intelligent knowledge operation service has been deployed in more than 60 cities across the country, accumulating a large number of corpora for the company's training of large models.
4th Chinese Online
Profitability: Return on equity -5.32%, gross margin 61.14%, net profit margin -4.10%
Main products: cultural products are the main source of income, accounting for 97.73% of revenue, and gross profit margin is 49.58%
Company highlights: Based on the open source AI model technology, Chinese Online uses the company's corpus data in the literary field and deploys it on its own server for AI large model development and training.
Section 3 Health Information
Profitability: Return on equity 18.25%, gross profit margin 53.88%, net profit margin 33.47%
Main products: project income is the main source of income, accounting for 78.76% of revenue, gross profit margin of 61.46%
Company Highlights: Jianlu Information has mature data desensitization technology, which can ensure that the masked data can be used to support the training of large-scale models in China while protecting user privacy.
The 2nd Shanghai Ganglian
Profitability: Return on equity 13.97%, gross margin 1.85%, net profit margin 0.54%
Main products: supply chain service business is the main source of revenue, accounting for 96.35% of revenue, gross profit margin of 0.60%
Company highlights: The corpus of Shanghai Ganglian model is based on the company's own data, and the high-value corpus data generated by a large number of data governance work.
1st straight flush
Profitability: Return on equity 32.50%, gross profit margin 90.90%, net profit margin 54.20%
Main products: Value-added telecommunication services are the main source of revenue, accounting for 42.89% of revenue, and gross profit margin of 86.14%
Company highlights: The HithinkGPT large model developed by Flush is the company's design from the training corpus, training framework and model structure, all of which are built from scratch and innovatively, and the pre-trained financial corpus has reached trillions of tokens.
The top 10 companies in AI corpus profitability, return on net assets, gross profit margin, and net profit margin in the past three years:
