Amazon, Google, Microsoft and Meta are accelerating investments in AI data centers and electricity infrastructure, but their expansion is facing growing opposition over power costs, water consumption, environmental impacts and pressure on local electricity grids.
Amazon leads the four technology companies in projected 2026 capital expenditure, followed by Alphabet, Microsoft and Meta. Their strategies include nuclear power agreements, renewable energy procurement, grid infrastructure and multigigawatt AI computing campuses.
However, communities across the United States and other markets are questioning whether the economic benefits of AI data centers justify their electricity requirements, water consumption and potential effects on household utility bills.
Figures represent company-wide capital expenditure forecasts or estimates, not spending exclusively on electricity. Microsoft’s figure is a calendar-year estimate rather than directly comparable fiscal-year guidance.
1. Amazon: $220 Billion Investment and Opposition to Data Center Expansion
Amazon’s approximately $220 billion capital expenditure outlook makes it the largest spender among the four companies.
AWS is expanding AI infrastructure, including Project Rainier in Indiana, while pursuing long-term electricity agreements.
In September 2026, Amazon and Constellation announced a 20-year agreement covering 690 MW of nuclear power associated with Maryland’s Calvert Cliffs plant.
Approximately 190 MW represents additional nuclear capacity expected between 2030 and 2032.
The Amazon–Constellation nuclear power agreement supports more than $3 billion in infrastructure investment.
Protests Against Amazon Data Centers
Amazon’s data center expansion in Northern Virginia has faced opposition from residents and environmental organizations concerned about electricity infrastructure, water demand and land development.
In Virginia’s Prince William County, the proposed Digital Gateway data center development attracted substantial community opposition. Residents challenged the project’s effects on residential neighborhoods, rural land and nearby Manassas National Battlefield Park.
Amazon is a major regional data center operator, but the Digital Gateway controversy should not be described as a protest against an Amazon-owned project.
In Oregon, AWS-related data center expansion has also attracted scrutiny over water consumption and the use of local infrastructure.
Amazon’s challenge: Securing electricity and community acceptance while expanding AI computing capacity.
2. Google: $200 Billion Spending and Nuclear Power Expansion
Alphabet expects capital expenditure of approximately $195 billion–$205 billion in 2026.
Google is expanding AI infrastructure for Gemini, Google Cloud and other services while investing in renewable electricity and nuclear generation.
In October 2026, Google and Constellation announced a 20-year agreement supporting 890 MW of additional nuclear generation through upgrades at 11 existing units.
A separate agreement covers 2,700 MW of existing generation capacity.
Together, the agreements cover 3,590 MW, although only 890 MW represents incremental capacity.
The Google–Constellation electricity agreement highlights Google’s strategy of securing long-term power supplies.
Community Opposition to Google Data Centers
Google’s data center developments have faced scrutiny over water consumption, energy requirements and environmental impacts.
In The Dalles, Oregon, local residents and environmental advocates have questioned water use associated with Google’s facilities, particularly in a region where water availability is an important concern.
Google’s proposed data center expansion in Chile also encountered environmental opposition over groundwater consumption. A Chilean environmental court ordered a reassessment of the project in 2024.
These cases demonstrate how water availability can become a significant constraint on AI infrastructure development.
Google’s challenge: Expanding electricity procurement while reducing the environmental impact of increasingly power-intensive AI operations.
3. Microsoft: $190 Billion AI Spending and Data Center Protests in Wisconsin
Microsoft’s estimated calendar-year 2026 capital expenditure is approximately $190 billion, reflecting investments in Azure, AI computing infrastructure and data center expansion.
Its energy strategy includes a 20-year agreement with Constellation supporting the planned restart of the 835 MW Crane Clean Energy Center in Pennsylvania.
The facility, formerly Three Mile Island Unit 1, is targeted to resume operations in 2028, subject to regulatory and project requirements.
Microsoft is also expanding its Fairwater AI infrastructure in Wisconsin.
Microsoft Faces Protests, Noise Complaints and Legal Challenges
Microsoft has encountered significant opposition to data center development in Wisconsin.
In Caledonia, Microsoft abandoned plans for a 244-acre data center project following community objections.
In Kenosha, residents organized opposition to another proposed Microsoft development covering approximately 240 acres. Protesters raised concerns about electricity consumption, water demand, noise and transparency.
Microsoft plans additional data center campuses in Mount Pleasant. The company indicated that its facilities could eventually require approximately 2 GW of electricity when fully developed.
Water requirements across the planned campuses could reach 8.4 million gallons annually.
Residents also complained about noise from Microsoft’s operational Fairwater facility, resulting in legal action in July 2026. Microsoft acknowledged that cooling fans were producing an unexpected humming sound and announced mitigation measures.
The opposition demonstrates that even financially strong technology companies can encounter community resistance affecting project approvals and construction schedules.
4. Meta: $145 Billion Investment and Opposition to Louisiana’s Hyperion Project
Meta expects 2026 capital expenditure of approximately $130 billion–$145 billion.
Its electricity strategy includes agreements involving Vistra, TerraPower, Oklo and Constellation, supporting a portfolio of existing and planned nuclear generation.
Meta has indicated that its nuclear initiatives could support up to 6.6 GW of new and existing clean energy by 2035.
Its Hyperion AI data center development in Louisiana represents another major electricity commitment.
Meta Faces Protests Over Gas Power Plants and Electricity Bills
Meta’s Hyperion project in Richland Parish has attracted opposition from environmental organizations and electricity consumer advocates.
In 2026, Entergy Louisiana proposed seven additional natural gas generating units to support the project, supplementing three previously approved units.
Together, the proposed and approved facilities could provide approximately 7.5 GW of generating capacity.
The Alliance for Affordable Energy and Union of Concerned Scientists challenged aspects of the expansion, raising concerns about electricity costs, emissions and limited public disclosure.
In September 2026, the organizations appealed a regulatory decision concerning access to information about Meta’s electricity requirements.
The advocates’ September 2026 legal challenge highlights the dispute over transparency and electricity infrastructure costs.
Residents have also raised concerns about construction traffic, dust, land development and the potential effects of the project on surrounding communities.
Meta’s challenge is demonstrating that its AI infrastructure expansion will generate local economic benefits without shifting unreasonable infrastructure costs onto other electricity customers.
Why Are Communities Protesting AI Data Centers?
The disputes involving Amazon, Google, Microsoft and Meta reflect several recurring concerns.
| Major concern | Potential impact on communities |
| Electricity consumption | Pressure on local generation and transmission infrastructure |
| Electricity prices | Risk of higher household tariffs if costs are not appropriately allocated |
| Water consumption | Competition for water resources in water-stressed regions |
| Natural gas generation | Additional emissions and local air-quality concerns |
| Noise pollution | Disturbance from cooling equipment and backup generators |
| Land development | Loss of agricultural land and changes to local landscapes |
| Infrastructure costs | Additional spending on substations, transmission and roads |
| Transparency | Limited access to power-demand forecasts and customer agreements |
Opposition does not necessarily mean a project will be rejected. However, permitting disputes, lawsuits and environmental reviews can delay construction and increase development costs.
What Does 1 GW of AI Data Center Electricity Cost?
A data center drawing 1 GW continuously consumes approximately 8.76 TWh annually.
| Electricity price | Annual electricity cost |
| $50/MWh | $438 million |
| $75/MWh | $657 million |
| $100/MWh | $876 million |
| $125/MWh | $1.095 billion |
These calculations exclude transmission charges, backup generation, taxes and other infrastructure expenses.
At $100/MWh, a continuously operating 2 GW facility would incur approximately $1.75 billion in annual electricity costs.
The figures illustrate why long-term electricity agreements and grid infrastructure are central to AI investment strategies.
Which Energy Companies Benefit From AI Infrastructure Spending?
AI data center expansion creates opportunities for electricity producers, renewable developers and equipment manufacturers.
Constellation Energy, NextEra Energy and Vistra are positioned to benefit from growing demand for long-term electricity procurement.
GE Vernova, Siemens Energy, ABB, Hitachi Energy, Eaton and Schneider Electric supply equipment used in power generation, transmission and electrical distribution.
Vertiv, Caterpillar and Cummins provide cooling, electrical infrastructure and backup-generation technologies.
However, community opposition and regulatory delays can affect the timing of supplier orders and project revenue.
Final Verdict: Amazon Leads Spending, but Energy and Community Risks Could Decide the Winner
Amazon leads projected 2026 capital expenditure among the four companies at approximately $220 billion, followed by Alphabet at $195 billion–$205 billion, Microsoft at approximately $190 billion and Meta at $130 billion–$145 billion.
These figures rank overall infrastructure spending, not direct investment in electricity generation.
Google is supporting 890 MW of additional nuclear capacity, Amazon’s Calvert Cliffs agreement includes approximately 190 MW of incremental generation, Microsoft is supporting an 835 MW nuclear restart, and Meta is pursuing a nuclear portfolio potentially reaching 6.6 GW of new and existing generation by 2035.
Community opposition is becoming another important factor in AI data center investment decisions. Protests over electricity bills, water resources, noise, emissions and land use can increase project costs and delay construction.
The companies best positioned to benefit from AI infrastructure expansion will be those that secure reliable electricity, control costs, obtain regulatory approvals and address local community concerns while delivering profitable computing capacity.
SHAFANA FAZAL
