(Yicai) June 26 -- There are not enough vertical applications that use large language models, so developers find it difficult to become profitable, despite huge investments, the founder and chairman of Chinese software developer Neusoft said recently.

If only a few people use a phone, for example, it is hard for a telecoms company to make money, despite its investment in communication infrastructure, Liu Jiren told Yicai at the Annual Meeting of the New Champions 2024.

The application of all new technologies needs to go through a starting-up phase before they become prevalent, Liu said.

"Artificial intelligence will not be an option but a must," he said. "The demand for AI will only get stronger, just like that for the internet. We have connected many customers with the internet and we will connect even more with AI technology.

"It is a complicated process for tech companies to go from investing to making profits in the AI sector. While some companies may be able to turn the technology into money, others may go under."

The value created by AI will reach nearly USD1.2 trillion by 2027, according to a report released by Swiss banking giant UBS earlier this month. Companies using computational resources to develop algorithms and LLMs are likely to see the strongest growth, while those embedding AI technology into specific applications have the greatest potential to generate profit.

The rapid growth in the AI sector means that many investors are still investing, Zurich-based UBS said. But, on the other hand, competition in the sector may trigger a capital expenditure cycle, leading to investment bubbles.

"When a new technology emerges, there are different views on whether investment will lead to a bubble or remain insufficient," Liu said. "For example, in the new energy vehicle industry, successful investments are seen as the right ones, while unsuccessful ones may be seen as bubbles or lacking.

"We don’t need to worry about investing in the wrong things because successful investments will achieve returns," he said.

Editor: Kim Taylor