AT&T, one of the largest telecommunications companies in the United States, had been using artificial intelligence models from Anthropic and OpenAI to assist with customer service, call transcription, and coding. These models were charged per usage and considered "closed," meaning the technology companies did not share the source code behind them.
This year, the company changed its strategy. Andy Markus, chief data and artificial intelligence officer at AT&T, decided to look for cheaper alternatives given the soaring costs of AI. The solution found was "open" models, which can be downloaded and modified without the need for payment or prior authorization.
The result was a significant change in the company's AI usage. In May, open models represented 20% of AT&T's AI usage. Since then, that share has risen to 40%, and Markus stated in an interview that it could reach 60% in the coming months.
This migration represents a problem for companies like OpenAI and Anthropic, which depend on the commercialization of their closed proprietary models. The growing adoption of open-source alternatives by large corporations suggests a shift in the AI market, where companies seek to reduce costs by choosing solutions that do not require ongoing payments.
