According to Gizmodo, Derek Strine, owner of Pennhurst Asylum haunted house in Pennsylvania, plans to convert his 130-acre property into an AI data center facility. The project has already seen $16 million invested in conversion work, with the first phase budgeted at $60 million for engineering and permitting alone. This represents a dramatic shift from Strine’s original $3 million investment in the haunted house business, which currently hosts historical tours, photography sessions, and paranormal investigations. The project faces significant community opposition over noise pollution and water usage concerns, reflecting broader tensions as data centers expand into new territories. This unusual conversion story highlights the speculative nature of the current AI infrastructure boom.
Table of Contents
The Data Center Gold Rush Meets Reality
What Strine’s story reveals is a broader pattern in commercial real estate where traditional property owners are pivoting toward data center development without necessarily having the technical expertise or industry experience. The CBRE survey showing 95% of real estate investors planning increased data center investments indicates we’re witnessing a classic speculative bubble pattern. Historically, when such high percentages of investors chase the same asset class, it often precedes market corrections. The fundamental mismatch here is that while real estate developers understand land acquisition and construction, they frequently underestimate the complex technical requirements of modern AI data centers, particularly around power distribution, cooling infrastructure, and network connectivity.
Community Resistance and Environmental Concerns
The opposition from Pennhurst’s neighbors reflects a growing national trend where communities are pushing back against data center development. As research indicates, data centers bring substantial environmental impacts that extend far beyond the “noise pollution” mentioned in the local concerns. Modern AI data centers consume staggering amounts of water for cooling—often millions of gallons daily—and draw enormous electrical loads that can strain regional power grids. What’s particularly concerning about projects like Pennhurst is that they’re often proposed for locations without existing infrastructure to support such intensive operations, requiring massive public investment in power and water systems that taxpayers ultimately bear.
The Ghost Center Phenomenon
The statistic that 90% of announced data centers never get built deserves closer examination. These “ghost centers” represent a significant market inefficiency where capital gets allocated to projects that lack fundamental viability. The reasons for this high failure rate include inadequate power availability, zoning and permitting challenges, community opposition, and changing market conditions during the lengthy development timeline. Many developers announce projects primarily to secure land options or attract investment, creating a misleading picture of actual data center capacity growth. This phenomenon isn’t new to infrastructure development—we saw similar patterns during the fiber optic boom of the late 1990s and the crypto mining facility rush of 2017-2018.
The AI Infrastructure Reality Check
While the AI boom has created genuine demand for computing infrastructure, the market is becoming increasingly stratified between hyperscale operators with established expertise and speculative developers chasing the trend. Projects like Pennhurst’s conversion face particular challenges because retrofitting historical or unconventional structures for data center use often involves compromises in efficiency, security, and scalability. The cooling requirements alone for AI workloads are substantially different from traditional data center needs, with GPU clusters generating concentrated heat loads that demand sophisticated thermal management solutions. Many speculative developers underestimate these technical requirements while overestimating their ability to secure anchor tenants in an increasingly competitive market.
Broader Market Implications
The Pennhurst story serves as a microcosm of larger trends affecting the data center industry. We’re seeing a geographic dispersion of development as prime locations become saturated, pushing projects into secondary markets and unconventional sites. This creates both opportunities and risks—while it can bring economic development to new areas, it also increases the likelihood of projects failing due to infrastructure limitations or community resistance. The current environment resembles previous technology infrastructure bubbles where the combination of hype, readily available capital, and FOMO (fear of missing out) drove overinvestment in projects with questionable fundamentals. As the market matures, we should expect consolidation and a shakeout where only the most viable projects survive.
