OpenAI has just announced a new plan to control energy costs for artificial intelligence data centers, in the context that electricity is becoming one of the biggest barriers to the expansion of global AI.
The Stargate Community initiative is designed to ensure that OpenAI's operation does not put pressure on the power system and electricity prices in the data center areas.
According to OpenAI, Stargate is a long-term investment project worth about 500 billion USD, focusing on building large-scale AI data centers to serve model training and inference.
The project has received the support of many large investors, including Oracle, and was publicly supported by US President Donald Trump when it was launched in January 2025.
A noteworthy point in the new plan is the commitment to self-financing energy costs. OpenAI said that each Stargate location will have its own community plan, built based on local conditions as well as contributions from people and local authorities.
The goal is to ensure that AI data centers do not increase electricity prices or cause energy shortages for the surrounding community.
Depending on the region, solutions may include building a new power source and energy storage system fully funded by the project, or investing in expanding existing power generation and transmission infrastructure.
OpenAI emphasizes that all additional costs related to energy will be paid by the project, instead of shifting the burden to local consumers.
OpenAI's move took place not long after Microsoft announced a similar initiative. Last week, Microsoft said it would pay a sufficient electricity price to offset actual costs, and coordinate with power companies to expand supply when necessary for the company's data centers in the US.
This company also aims to reduce water use and limit negative impacts on the surrounding community.
As AI consumes more and more resources, the fact that technology corporations actively invest in energy infrastructure shows that the sustainable AI development problem is being prioritized, instead of just chasing scale and speed.
