According to the Financial APP, OpenAI, the leader in generative artificial intelligence, has released a new feature that allows enterprise customers to customize or fine-tune their AI startup company's most powerful AI large model, GPT-4o, using their own company's data. As the competition for enterprise customers by AI large model companies intensifies, OpenAI has launched this new feature specifically for enterprises.

Analysts widely believe that OpenAI's move to allow enterprise customers to customize the most powerful AI large model, GPT-4o, which is considered the most comprehensive AI large model to date, is mainly due to the increasing competition in the commercialization of AI products and the need for OpenAI itself to generate larger revenue to support research and development expenditures. Enterprise customers are also under increasing pressure to prove that investments in artificial intelligence can bring profit growth, thus pushing OpenAI to allow customization of AI large models for enterprise customers.

OpenAI launched this customization feature, known as 'fine-tuning', on Tuesday local time in the artificial intelligence industry. The fine-tuning feature allows enterprise customers to train their existing AI large models on additional relevant information specific to particular tasks or subject areas.

For example, a company focused on skateboard production may fine-tune the AI model to use it as a chatbot for customer service related to skateboard functionalities, solving detailed aspects of professional wheels and skateboard care.

For OpenAI's flagship large models, this capability is brand new - its predecessor GPT-4 or other flagship models did not offer fine-tuning. Previously, the company only allowed enterprise customers to fine-tune other smaller versions of AI large models, including GPT-4o mini, which is a smaller, more simplified version of GPT-4o with reduced capacity. By opening up the fine-tuning feature for GPT-4o and its predecessor GPT-4, enterprises can now optimize OpenAI's most powerful AI large models more directly, without relying on external IT services.

The opening of the fine-tuning feature by OpenAI is considered a milestone.

Many AI technology companies are committed to providing customized AI large-scale models of different sizes. Olivier Godement, the product owner of OpenAI's API, said that OpenAI is striving to make it easier for customers to adjust their most powerful AI models by working directly with enterprises rather than using external services or weaker products.

"We have always been very focused on reducing barriers, reducing functional friction, and starting the required workload for fine-tuning," emphasized Godement in a statement.

To fine-tune the AI large-scale models, enterprise-level customers need to upload their data to OpenAI's servers. John Allard, the software engineer responsible for the customization work, stated that the average training time for large-scale model fine-tuning is one to two hours. Allard said that in the early stages of this feature, users can only fine-tune AI large-scale models using text-based data, not images or other content for training.

From the potential market and market competition trends of AI large-scale models, OpenAI's move is of far-reaching significance. Enterprise-level customers often need AI models that can adapt to their specific industries or business needs. Through fine-tuning, companies can optimize the GPT-4o model to perform better in specific tasks, such as improving the responsiveness of customer service systems or providing more accurate advice in complex technical fields such as software engineering or financial analysis.

With the intensification of AI market competition, each company is seeking tangible business benefits through AI investments. By allowing customers to customize GPT-4o, OpenAI not only enhances the applicability of its model but also improves its competitiveness in the enterprise market. Companies can integrate AI into their operations more directly and efficiently, thereby achieving value faster.

OpenAI simplifies the customization process, making it easier to fine-tune and personalize AI, and lowering the technical barrier for enterprises to use AI. This will attract more enterprise customers, especially those small and medium-sized enterprises without strong technical team support, and expand the application scope of GPT-4o in the market.

The "explosion moment" of enterprise-level AI applications is getting closer and closer!

The Chief Operating Officer of OpenAI predicted earlier this year that 2024 will be the "year of application" for artificial intelligence (AI) technology, which means that various generative AI applications such as ChatGPT and Sora developed by companies like OpenAI may be on the verge of "explosive growth" worldwide.

As enterprises and individual users gradually realize the significant improvement in work efficiency brought by AI, and AI leader OpenAI introduces the most powerful AI large-scale model customization feature to enterprise-level customers, it is expected that AI technology will soon be widely used in economies and markets worldwide. Therefore, the threshold for the outbreak of various types of AI software seems to be getting closer and closer. Cloud giants, such as Amazon and Microsoft, who are committed to building the AI software development ecosystem, as well as AI large-scale model leader OpenAI, are expected to usher in an "iPhone moment" - a moment when both performance and market cap soar.

Enterprise-level AI is undoubtedly at the forefront of AI software development, with the total market size of enterprise artificial intelligence software expected to grow rapidly at an astonishing rate of 52% annually from 2024 to 2029, according to a report by Mordor Intelligence.

Mordor Intelligence pointed out that the Fourth Industrial Revolution (Industry 4.0), characterized by physical and advanced digital technologies such as AI and ubiquitous mobile supercomputing power, has had a significant impact on industries worldwide. Especially, global enterprises are gradually realizing the tremendous value of integrating AI into their business processes, improving operational efficiency and reducing costs through AI automation, and predicting business outcomes and enhancing profitability with AI assistance.

In a recent research report, UBS Group, an international major bank, pointed out that the global technology industry has just begun a large-scale performance growth cycle. UBS expects that by 2027, AI technology will achieve extremely wide applications in various industries in major economies around the world, thereby driving the market value of large AI models and AI software applications to a subdivided market worth up to $225 billion, an epic leap compared to the $2.2 billion scale in 2022, and the compound annual growth rate is expected to reach 152% during this period. UBS also predicts that the total revenue of the AI industry will increase 15 times, from approximately $28 billion in 2022 to $420 billion in 2027.