Amazon is facing significant challenges in its aggressive data center expansion, which is critical to supporting the company’s booming artificial intelligence workloads and AWS cloud services. According to internal documents and sources familiar with the matter, the tech giant is on pace to build 240 new data centers by 2040, representing approximately 75 million gross square feet—equivalent to 27 Empire State Buildings.
Power shortages are emerging as the most critical constraint. An internal Amazon document reveals that “across the AMER region, we are experiencing headwinds in power, zoning and permitting, water, and workforce/labor that are providing challenges to our long-term capacity growth.” US data center electricity consumption is forecast to triple by 2030, largely driven by generative AI demands. Bernstein Research estimates that electricity demand for AI data centers could exceed supply in just two years without intervention.
The power crisis has created what Amazon employees internally call “zombie” data centers—facilities that sit idle or “operationally blocked” because they lack sufficient power to operate. Amazon is experiencing electricity supply issues in Oregon, Ohio, and Northern Virginia. In Portland, utility partner PacifiCorp faces electricity shortfalls and generation constraints forecast until 2030, leaving Amazon with unpredictable energy supplies.
AWS CEO Matt Garman publicly acknowledged these challenges during an August podcast, stating customer demand is “mind-boggling” and requires “a lot of investment,” adding “we’re probably going to be tight for the next little bit of time.” The company is pursuing multiple solutions: conducting power studies with utility partners, exploring partnerships with third-party colocation providers, expanding into newer markets like Indiana, and investing heavily in alternative energy sources, including a $500 million funding round for small modular nuclear reactors.
Water infrastructure and skilled labor shortages compound the problems. Many modern data centers use water cooling systems for AI workloads, straining local water infrastructure. Amazon also faces a critical shortage of skilled electricians, particularly in rural areas, forcing the company to rely on more expensive traveling electricians and prioritize resource allocation. The company has engaged with local colleges, contractors, and unions to develop training programs.
In Santa Clara, California, Amazon is paying up to 50% higher electricity rates under a proposed 15-year, 20-megawatt agreement, demonstrating the premium costs associated with securing power in competitive markets. Internal emails show AWS executives are “highly concerned” about delays at multiple California facilities.
Key Quotes
Across the AMER region, we are experiencing headwinds in power, zoning and permitting, water, and workforce/labor that are providing challenges to our long-term capacity growth
From an internal Amazon document obtained by Business Insider, this quote reveals the multiple simultaneous constraints facing the company’s data center expansion plans critical to supporting AI workloads across its largest market.
The need for more power and a better functioning grid was clear before AI. Now, it is becoming urgent.
David Cahn, a Sequoia Capital general partner, articulated how generative AI has transformed power infrastructure from a long-term concern into an immediate crisis affecting the entire tech sector and venture capital investments.
We’re probably going to be tight for the next little bit of time
AWS CEO Matt Garman acknowledged during an August podcast that Amazon faces capacity constraints due to ‘mind-boggling’ customer demand for AI services, signaling potential service limitations for enterprises seeking cloud AI resources.
It’s almost like mosquitoes to light — we are attracted to power
An Amazon insider’s metaphor captures the company’s desperate search for electricity to fuel its AI ambitions, illustrating how power availability has become the primary driver of data center location decisions.
Our Take
Amazon’s infrastructure crisis exposes a fundamental miscalculation in the AI industry: the assumption that digital transformation could outpace physical reality. While companies rushed to develop increasingly powerful AI models, they underestimated the massive energy, water, and skilled labor requirements to deploy them at scale. The “zombie” data centers represent a new category of stranded assets—built but unusable infrastructure that symbolizes the gap between AI ambition and execution capability.
The willingness to pay 50% premiums for power and invest $500 million in experimental nuclear technology demonstrates how desperate the situation has become. This isn’t just an Amazon problem—it’s an industry-wide reckoning that will reshape competitive dynamics. Companies with existing power agreements and efficient data centers gain significant strategic advantages, while newcomers face insurmountable barriers to entry. The AI revolution may ultimately be constrained not by algorithmic innovation but by 19th-century infrastructure: electrical grids, water systems, and skilled trades.
Why This Matters
This story reveals critical infrastructure bottlenecks threatening the AI revolution’s momentum. As the largest cloud provider, Amazon’s struggles signal systemic challenges facing the entire tech industry as generative AI transforms computing demands. The potential supply-demand mismatch in just two years could slow AI innovation and deployment across sectors.
The emergence of “zombie” data centers represents billions in stranded capital and highlights how physical infrastructure constraints are becoming the limiting factor for AI advancement, not just chip availability or algorithmic breakthroughs. This shift has profound implications for AI companies, cloud providers, and enterprises planning AI initiatives.
The situation is driving massive investments in alternative energy, including nuclear power, potentially accelerating the clean energy transition. It’s also reshaping data center geography as companies move beyond traditional hubs to access power and resources. For businesses relying on cloud AI services, these constraints could mean higher costs, capacity limitations, and delays in accessing cutting-edge AI capabilities, fundamentally impacting digital transformation timelines across industries.
Recommended Reading
For those interested in learning more about artificial intelligence, machine learning, and effective AI communication, here are some excellent resources:
Recommended Reading
Related Stories
- How AI Can Help Make Buildings More Energy Efficient
- Amazon to Invest Additional $4 Billion in AI Startup Anthropic
- EnCharge AI Secures $100M Series B to Revolutionize Energy-Efficient AI Chips
- Wall Street Asks Big Tech: Will AI Ever Make Money?
- The AI Hype Cycle: Reality Check and Future Expectations