The gap is widening between corporate AI adopters and laggards

A new divide is forming among global companies based on their adoption of artificial intelligence (AI), according to recent data from OpenAI. The report highlights a significant disparity between enterprises that have rapidly embraced AI and those that are adopting it at a slower pace. Firms in the top 10% of AI usage in June consumed output tokens eight and a half times more per user than companies in the middle of the pack, the study found. In January, the top firms were using two and a half times the median amount of tokens.

OpenAI's report, shared exclusively with Semafor, indicates a shift in corporate AI usage from the traditional ChatGPT interface to their agentic coding platform, Codex. This suggests businesses are increasingly turning to AI for more specialized tasks rather than general use. Ronnie Chatterji, OpenAI's chief economist, noted this trend to Semafor.

OpenAI is ramping up its efforts to attract corporate customers, partly to compete with Anthropic, a rival company that has focused early on businesses and coding tools alongside consumer applications. Both OpenAI and Anthropic are preparing for major initial public offerings (IPOs). The widening gap between AI leaders and laggards echoes past technological revolutions, such as the internet, where early adopters gained a significant advantage, and later laggards eventually caught up once the productivity benefits became evident.

Chatterji pointed out that the primary challenge for CEOs and Chief Technology Officers (CTOs) today is demonstrating to their boards and shareholders that the increased spending on AI is justified. However, he emphasized that high token usage does not automatically translate to improved productivity, and accurately measuring return on investment (ROI) remains a significant challenge.