According to Nordea analysts Kirsti Sunde Midttun and Ole Håkon Eek-Nielsen, the profitability of artificial intelligence (AI) companies is under structural pressure due to high inference costs, rapid model depreciation, and intensifying competition from free and open-source alternatives [1]. The analysts question the durability of current AI business models, noting that the central economic challenge for AI developers remains the high cost of inference, which is a key reason why leading model companies are not currently profitable [1].
Nordea describes frontier AI models as infrastructure with unusually short useful lives, requiring value extraction before technological obsolescence sets in [1]. The proliferation of capable models being published free of charge is suppressing market willingness to pay and undermining the business models of developers who charge for access [1]. As a result, the industry faces expensive build costs, rapid depreciation within months, and growing competition from both commercial and free alternatives [1].
Over the summer, Nordea observes a rise in investor skepticism toward AI-related equities, which has contributed to a notable rotation out of technology stocks and into cyclical, defensive, and value-oriented sectors [1]. The analysts present a more skeptical view of the industry's prospects, emphasizing the challenges to profitability despite AI's rapid growth [1].
CONCLUSION
Nordea highlights significant structural challenges to AI profitability, including high costs and competitive pressures, which have fueled investor skepticism and a rotation away from tech stocks. The outlook for AI equities remains uncertain as questions persist about the sustainability of current business models and the industry's ability to achieve durable margins.