Building and training the artificial intelligence systems we use today takes an immense amount of electricity. Data centers, which house the thousands of interconnected computer chips that power AI, are drawing so much power that they are straining local electrical grids. To keep up, major tech companies are bypassing the public grid and building their own private natural gas power plants to run their systems.
SpaceX is building a large factory in Texas to manufacture critical components for natural gas turbines. These turbines are the engines that burn gas to generate electricity. Currently, there is a global shortage of the specific metal blades and vanes required to build these machines. By producing these parts in-house, SpaceX aims to slash the wait time for new power plants by up to 18 months, helping get more data centers online much faster than current supply chains allow.
The challenge of extreme heat
A gas turbine works by burning fuel to spin a shaft, which creates electricity. Inside the hottest part of the turbine, temperatures reach over 3,000 degrees Fahrenheit. This is significantly hotter than the melting point of the metal used to make the turbine blades. Engineers solve this using two primary methods. First, they create the blades using a specialized casting process where the entire part is grown as a single, solid crystal, which prevents structural weaknesses that would cause the metal to snap or crack under extreme pressure. Second, they use intricate internal cooling channels and special heat-resistant coatings to keep the metal from vaporizing. Because this process is so difficult to perfect, only a handful of factories worldwide can produce these blades at a large scale.
The push for faster AI infrastructure creates a direct trade-off with public health. Natural gas turbines do not just produce electricity; they also emit pollutants like formaldehyde and smog-forming chemicals. Residents living near data centers that rely on these turbines have already begun protesting and filing lawsuits over air quality concerns. As tech companies rush to secure power for their AI projects, they are creating a new geography of industrial pollution. The shift highlights a difficult reality for the industry: the rapid advancement of digital intelligence is becoming tethered to old-fashioned, high-emission physical infrastructure.
Liked this one? The next lands at breakfast.
Every story in tomorrow's AI news, rebuilt in plain English — five minutes, sources linked, free forever.
By joining you agree to receive Article's daily newsletter — unsubscribe in one click. Privacy