Peter Berezin, Chief Global Strategist at BCA Research, contends that the much-hyped AI boom is primarily a stock market phenomenon, with its broader economic impact yet to be genuinely felt. Speaking on CNBC's 'Closing Bell Overtime', Berezin engaged with the interviewer about his latest research, which suggests that the transformative power of generative AI has not yet translated into significant aggregate productivity gains across the economy, challenging the prevailing narrative of immediate, widespread economic uplift.
Berezin points to a striking disconnect between market enthusiasm and tangible economic data. "You don't see that in the aggregate productivity data yet," he observed, referring to the lack of a noticeable surge in output per hour worked. He noted that current productivity growth merely aligns with pre-AI era projections, lacking the accelerated trajectory one might expect from a revolutionary technology. This echoes the famous "Solow Paradox" of the late 1980s, where economist Robert Solow remarked, "You can see the computer age everywhere but in the productivity statistics." It took over a decade for the widespread adoption of personal computers and the internet to visibly impact productivity, with profits from the internet only truly materializing around 2005.
The implication for investors is critical: patience may be a virtue they lack. While specific companies like Nvidia are undoubtedly reaping substantial rewards from the AI infrastructure build-out, the broader economic benefits and widespread corporate profitability remain elusive. Berezin cautions that "for investors, it's not enough for productivity to rise, you also need the profits to increase." The risk is that investors, driven by short-term gains and the fear of missing out, may grow impatient before AI's full economic potential is realized, potentially leading to market corrections.
Crucially, Berezin differentiates AI from prior technological revolutions like software, which benefited from strong network effects and minimal replication costs. With traditional software, an initial investment yields a product that can be distributed at near-zero marginal cost, fostering economies of scale. AI, however, presents a different economic model. "In some sense, AI looks a lot like airlines," Berezin provocatively stated. This analogy highlights that while AI is an undeniably important and indispensable industry for future economic growth, its underlying economics are characterized by high capital expenditure requirements, such as massive data centers and specialized chips. These significant, ongoing investments, coupled with the commoditized nature of foundational AI models (similar neural net transformer architectures), could constrain broad-based profitability.
The absence of strong network effects, where the value of a product increases with more users, further complicates the picture for AI. Unlike social media platforms or office suites, individual AI users often interact directly with the AI, not with other users of the same AI, diminishing this powerful growth driver. Therefore, while AI will undoubtedly drive innovation and efficiency, the path to widespread, sustainable profitability across diverse industries may be more arduous and less immediate than current market valuations suggest. Investors should meticulously monitor the free cash flow of hyperscalers for early indicators of the AI boom's long-term viability and broad economic translation.
