Rising AI Costs Drive Software Developers to Open-Weight Models

A growing number of software companies are turning to open-weight artificial intelligence models, including Chinese ones, due to the cost of proprietary models from U.S. AI developers, Bloomberg reported Monday (Sept. 21).

Open-weight models are typically cheaper and enable firms to create their own models with their own data, thereby cutting the risk of outsourcing their technology to another company, according to the report.

However, other companies have found that the strategy of using open-weight models to build their own model doesn’t work because of the higher upfront costs and the need for specialized talent, computer infrastructure and data, if they don’t have enough data of their own. In the case of Chinese open-weight models, some companies’ clients are concerned about data privacy, the report said.

Some firms that have built their own models continue using models from big AI labs when there is a need for the best capabilities that are available, per the report.

PYMNTS reported in July that for chief financial officers of middle market companies, the question is whether the savings, flexibility and control offered by open models are sufficient to justify assuming more responsibility for the infrastructure beneath them.

Self-hosting requires computing capacity, storage, cybersecurity controls, monitoring tools and skilled employees, while a proprietary, closed platform typically bundles many of those responsibilities into its price.

It was reported in August that AT&T cut the costs of coding and some other advanced AI tasks by as much as 56% by using tools that route employees’ queries to cheaper models when appropriate. AT&T found that when doing so, the quality of the performance of the AI declined by only 2%.

AT&T aims to increase the share of employees’ queries that are powered by open-source models from the current 40% to between 60% and 70% in the coming years, per the report.

It was reported in June that the rising costs of models from large AI labs have been driven by the shift from chatbots to agents, which consumer more computing power, as well as the AI labs’ shift from flat subscriptions to token-based billing.

Chinese labs are able to charge less than U.S. companies due to their more efficient models and China’s lower energy costs, the report said.