Every transformative technology cycle comes with a familiar debate: Are we witnessing the next great economic opportunity—or the formation of a bubble?
We have had this conversation before. From the dot.com era to emerging technology booms that followed, periods of rapid investment often spark concerns about whether expectations are outpacing reality. Today, AI (artificial intelligence) finds itself at the center of that discussion.
The most important question may not be whether AI itself is a bubble. Instead, it is whether the market can support the pace of growth that businesses, investors, and customers expect. New research from AI infrastructure provider CUDO Compute shows 98% of U.S. firms believe there is an AI bubble, but the concern is being driven less by hype and more by mounting pressure around infrastructure, energy, and geopolitics.
As AI moves from experimentation to enterprise-scale implementation, attention is shifting toward the practical challenges that accompany widespread adoption. The conversation is becoming less about possibility and more about sustainability.
In that new research, energy is emerging as one of the defining constraints shaping the next phase of AI growth. More than a quarter of respondents, 28%, said energy is becoming a growing concern when planning AI workloads, while 25% said rising costs are forcing them to slow or pause AI training activity altogether.
This is usually the tip of the spear when we see things come crashing down. Many of you reading this now might not recall the pain of the dot.com bubble. However, it wiped out companies, bankrupted investors, and left a lot of people asking who they turn to then when everything around them crumbled like a house of cards.
That shift signals an important moment for the market. Mature technology ecosystems require the infrastructure, resources, and operational foundations necessary to turn innovation into long-term economic value.
The findings suggest businesses are becoming increasingly concerned about whether the physical infrastructure underpinning AI can keep pace with demand. Among U.S. respondents, the leading drivers behind concerns around an ‘AI bubble’ were high energy and infrastructure costs (35%), rapid investment in GPUs without matching infrastructure readiness (34%), long planning and permitting timelines (29%) limiting building, geopolitical risk and export controls (28%), and shortages of affordable power or grid capacity (28%).
Geopolitics and policy are also becoming increasingly influential, although they are not yet overriding economic realities. While 41% said trade restrictions, tariffs, or export controls are influencing deployment decisions, and 28% cited geopolitical instability as a reason to keep workloads closer to home, 37% said cost and performance still take priority over sovereignty concerns.
The debate surrounding an AI bubble is likely to continue, but history suggests market concerns are often most useful when they force industries to confront underlying realities.
What makes the current AI conversation particularly interesting is that many of the concerns now being raised are not about the technology’s potential. Instead, they center on the processes and strategy needed to support that potential in the long term.
The future of business should not be made solely on AI outputs. AI can be tricky and therefore fool people into offering up answers that imply they are rooted in sound business reasoning. But true strategic decisions need human interpretation to determine a strong strategic outcome from department to department and company to company.
Whether the market ultimately views this moment as a bubble or a breakthrough may depend less on AI itself and more on how effectively the industry addresses the challenges that come with rapid growth and human understanding.
Want to tweet about this article? Use hashtags #construction #IoT #sustainability #AI #5G #cloud #edge #futureofwork #infrastructure