US gas-fired power development has surged to 378 GW, with 189 GW dedicated to AI data centers. This expansion signals a major shift in energy planning, though it faces supply chain bottlenecks and risks of rising infrastructure costs.
The United States is currently experiencing a rapid buildup of gas-fired power capacity, driven largely by the massive energy demands of artificial intelligence infrastructure. As of August 2026, the total pipeline of gas-fired power projects under development in the country has expanded to 378 gigawatts (GW). Of this total, 189 GW is explicitly designated to power data centers, highlighting an aggressive strategy by technology firms to secure reliable, captive power supplies.
This trend underscores a shift in how major tech firms are approaching energy needs. While renewable energy remains a priority for many, the continuous, reliable power requirement of AI data centers—often referred to as baseload power—has led developers to favor gas-fired plants. Texas has emerged as the central hub of this expansion, with 122 GW of capacity in the pipeline, and nearly two-thirds of that specifically targeting data processing facilities. This represents a significant acceleration in local power infrastructure development over the last six months.
Infrastructure and Supply Chain Challenges
The sheer scale of this buildout, estimated to cost over US$647 billion, has created unexpected pressure on the supply chain for power generation equipment. Traditional heavy-duty gas turbines, which are the standard for large-scale power generation, are facing multi-year lead times. This bottleneck has forced project developers to adapt. Many are now opting for smaller, aeroderivative units and reciprocating engines. While these alternatives allow for faster installation and circumvent production backlogs, they are generally less efficient than traditional combined-cycle plants and may generate higher emissions per unit of output.
Financial and Strategic Risks for Investors
For investors monitoring the energy and technology sectors, the rapid expansion brings both opportunities and significant risks. The massive capital spending involved raises concerns about the potential for stranded assets. If the projected growth in data center electricity demand slows or if technological improvements reduce energy intensity, these newly built plants could become underutilized liabilities. This creates a risk where utility companies and ratepayers might end up bearing the burden of these substantial, locked-in costs.
Furthermore, the competition for hardware is keeping upward pressure on electricity prices. As developers bid aggressively for limited equipment, the cost of building power projects has risen globally. While the US government has encouraged large tech firms like Microsoft, Google, and Meta to pursue independent power solutions to mitigate the impact on retail electricity bills, the long-term viability of many projects remains dependent on whether actual demand matches the current optimistic projections.
Investors should track the commissioning timelines of these 378 GW worth of projects. The next important monitorable will be how many of these plans actually move to the construction and operation phase, versus how many are cancelled due to regulatory hurdles, emission concerns, or shifts in the energy market. The balance between meeting AI's immediate power hunger and managing the long-term financial and environmental costs will be the defining factor for this sector in the coming years.