Palo Alto Networks CEO Nikesh Arora says businesses will not adopt AI at scale unless the cost of using AI—particularly token or usage pricing—falls substantially. Speaking on CNBC’s “Squawk on the Street,” Arora links enterprise adoption to affordability rather than performance, arguing that demand and efficiency improvements are emerging. He comments that a reported 54% efficiency gain from OpenAI’s latest model is “a good start,” but he adds that costs must continue to drop over the next two years for broader rollout.
Across the outlets, Arora’s central claim is that token costs may need to plunge by roughly 90% before AI becomes widely affordable for businesses. One report also breaks this into a timeline, saying costs should fall about 20% over the next 12 months and then by a further 90% by the following year. The coverage frames the issue as a potential barrier to large-scale enterprise deployments if pricing for AI usage remains too high, despite improvements in model efficiency.
