HSBC is pushing back against mounting claims that artificial intelligence has entered “bubble territory,” pointing to fresh evidence that companies are already generating meaningful returns. The bank says new data offers “further evidence of ROI from AI,” countering one widely cited—but flawed—statistic that has fueled skepticism.
In a recent note, analyst Yuning Bai argues that critics of the AI boom have leaned too heavily on a “startling” and “rather weak” data point. The concern, Bai explains, stems from a July report by MIT NANDA claiming that “95% of organizations are getting zero return” on their GenAI investments.
HSBC contends that this figure has had “a disproportional impact on the AI bubble debate,” even though the research behind it is “rather weak.”
More recent findings tell a different story. HSBC highlights a new Wharton-GBK study showing that “many enterprises are already seeing tangible benefits in productivity and performance,” with 74% of surveyed firms reporting positive returns from their generative AI deployments.
The bank notes that it is “implausible” for AI success rates to jump from 5% to nearly 75% in just a few months. Instead, Bai emphasizes that “measuring the success of AI implementation is difficult and the results are very sensitive to precise methodological choices.”
As a result, the original 95% figure “can probably be taken with a pinch of salt,” Bai writes.
HSBC argues that what matters most is methodological consistency—an area where the Wharton-GBK study excels, now in its third iteration. According to the bank, this research “provides a useful counterpoint” to bubble fears and offers clearer proof that AI investments are already yielding real, measurable benefits for most companies.