"The AI spending looks impressive but it's probably the peak," stated Jonathan Millar, Barclays' senior economist, during an interview on CNBC's "Power Lunch." Millar spoke with the program's hosts about Barclays' perspective on artificial intelligence capital expenditures and the potential for a break-even point for AI investments. His analysis suggests that while current AI spending is substantial, the rate of growth may be unsustainable, drawing parallels to past technological booms.
The core of Millar's argument centers on the idea that the current surge in AI investment, particularly in hardware and infrastructure, might be reaching its zenith. He points out that the sheer scale of capital expenditure required to build out the necessary AI capabilities is immense. "We calculate that you'd probably have to raise the level of investment by about 20% per year in perpetuity for that to continue," Millar explained, highlighting the difficulty in sustaining such rapid growth indefinitely. This implies that while AI is transformative, the current pace of investment might outstrip the economy's capacity to absorb it without a proportional increase in productivity gains.
A key insight from Millar's commentary is the comparison to the dot-com bubble of the late 1990s. He drew a parallel, noting that during that period, there was significant investment in internet infrastructure and technology, which ultimately led to a boom. However, the subsequent "productivity miracle" that justified those investments took time to materialize. "The 1990s were great in terms of productivity, even though we had a boom and bust cycle. But in this case, the issue here is that it's going to be hard to sustain that kind of growth, and I think that's important as well," Millar elaborated. This suggests that while AI has the potential to drive significant productivity gains, the current investment levels might be front-loading those gains, making it challenging to maintain the same trajectory without a clear and immediate return on investment.
Another critical point Millar raised concerns the foundational infrastructure required for AI. He emphasized that the demand for computing power, particularly from hyperscalers like Google, Meta, Microsoft, and Amazon, is a primary driver of AI capital expenditure. This infrastructure build-out, while essential, requires massive investment in data centers and specialized hardware. "You really need it to continue to grow very fast to continue to provide the kind of thrust to annualized GDP growth," Millar stated. This underscores the dependency of AI's economic impact on the continuous expansion of its underlying technological backbone.
The analysis suggests that the current high levels of AI investment may not translate into immediate, broad-based economic growth in the same way that previous technological revolutions did. While the potential for AI is undeniable, the economic models and projections need to account for the significant lead time and capital intensity involved in building out the necessary infrastructure. The comparison to the 1990s highlights the risk of overestimating the immediate impact of new technologies. The "boom and bust" cycle mentioned by Millar serves as a cautionary tale for investors and policymakers alike, stressing the importance of realistic expectations and sustainable growth models.
Related Reading
- AI Spending: Peak Hype or Broadening Foundation?
- AI Spending Frenzy Sustainable, But Job Market Nuance Needed, Says Meeks
- AI Cycle's Long Runway: Why It's Not Your Grandfather's Tech Bubble
The data presented during the discussion, showing stock performances of major tech companies and indices like the S&P 500 and Nasdaq, provided a backdrop to Millar's economic outlook. While these indices showed some positive movement, Millar's focus remained on the underlying drivers of that growth and their long-term sustainability. His assessment suggests that the impressive AI spending might be a temporary surge rather than a sustained engine of economic expansion, at least in its current form.
Millar's perspective offers a pragmatic view on the current AI investment landscape. He cautions against an overly optimistic projection of immediate economic returns, emphasizing the significant infrastructure requirements and the historical patterns of technological adoption. The long-term impact of AI will likely depend on how effectively the foundational infrastructure can be scaled and how quickly productivity gains can be realized to justify the substantial capital outlay.
