AI developer OpenAI set to utilize gas turbines at its initial Stargate facility for enhanced power generation, mirroring Elon Musk's approach.
Tech Giants Turn to Gas Turbines and Small Modular Reactors for AI Power Needs
Tech giants like OpenAI, Microsoft, Google, and Meta are investing in gas turbines and small modular reactors (SMRs) to power their AI data centers, addressing growing energy demands driven by the enormous and rapidly increasing workloads of AI computing.
The need for reliable, scalable, and fast-deployable power sources that can bypass constraints of traditional electric grids is at the heart of this trend. These investments aim to ensure continuous, high-quality energy availability critical for AI computing.
Grid Constraints and Interconnection Delays
Existing power grids often struggle with the massive, variable power demands of AI data centers, resulting in bottlenecks, higher costs, and power quality issues. Deploying on-site or dedicated power generation like SMRs or gas turbines helps bypass these limitations.
Small Modular Reactors (SMRs)
SMRs, factory-built, modular nuclear reactors, offer a small footprint, long fuel cycles, and scalable power output (1 MW to 300+ MW). They are ideal for diverse data center sites requiring stable baseload power. Big tech companies have already started investing heavily in SMRs as a clean, reliable energy source to meet surging AI workloads.
Gas Turbines
Gas turbines, such as GE Vernova’s LM2500XPRESS, provide quick response times to sudden changes in demand, supporting smooth and continuous power without interruptions crucial for AI systems. They also have emission reduction technologies to reduce environmental impact.
Alternative Onsite Solutions
Some companies are also exploring fuel cell technology powered by hydrogen as a flexible, modular power source capable of quick onsite deployment with potential climate benefits.
Strategic Energy Independence
Owning dedicated power sources reduces reliance on utility grids, avoiding price volatility and supply uncertainty, which are critical for the uninterrupted operation of AI data centers.
The first deployment of gas turbines at a data center site is happening at OpenAI's Stargate data center in Abilene, Texas. If all the turbines are installed, the total output would be 986MW, nearly enough energy to power 500,000 GB200 NVL72 chips.
In the interim, companies are resorting to importing power plants and running gas turbines as a temporary solution to power supply issues. Elon Musk has bought a power plant overseas and plans to ship it to the U.S. for his data center's electrical requirements.
The first deployment of SMR technology is expected in the 2030s, but its market entry is uncertain. If it enters the market, it would provide a solution for quick deployment of nuclear power near data centers.
Crusoe, the AI infrastructure company building the first OpenAI Stargate data center, has acquired the turbine generators to provide flexible and efficient power. Elon Musk has deployed portable power generators at the Memphis Supercluster to ensure sufficient supply, but this has raised pollution concerns among residents.
The GE LM2500EXPRESS units come with Selective Catalytic Reduction (SCR) technology, reducing nitrogen oxide emissions. Stargate 1 currently has permits for only 10 turbines at the site, but it's unclear whether permits for the other 19 turbines are still in process or if they are planned for other sites.
As power supply remains a significant issue for data center deployments, with the grid sometimes struggling to provide the required power, leading to increased prices and reduced power quality, investments in gas turbines and SMRs offer a promising solution for the future of AI infrastructure.
Investments in data-and-cloud-computing technology, like gas turbines and small modular reactors (SMRs), are critical for addressing energy demand challenges caused by AI computing's enormous and rapidly increasing workloads, as these technologies provide reliable, scalable, and quick-deployable power sources to bypass limitations posed by traditional electric grids.
These investments in technology, such as SMRs and gas turbines, aim to ensure continuous, high-quality energy availability essential for AI computing, thus addressing power quality and demand issues faced by AI data centers when relying on existing power grids.
