How GE Vernova builds the turbines powering the AI data center boom

A special view inside GE VernovaThe largest gas turbine plant in Greenville, South Carolina, provides new evidence that the AI ​​boom is continuing strong.

Inside, engineers are working with factory workers to speed up production of this complex machine. The company hired 200 workers last year, and another 300 people are expected to start working in this factory by the end of the year.

It is artificial intelligence that fuels growth.

Hyperscalers — companies like these: Amazon, Google, Microsoft And Seer – they are lining up to buy the company’s gas turbines. With AI data centers requiring significant amounts of energy and causing bottlenecks on the grid, these companies are increasingly relying on independent energy sources such as gas turbines.

“Currently, when you need robust power on a large scale, the industrial gas turbine is one of the leading solutions for this,” says Pablo Koziner., GE Vernova’s chief commercial and operations officer told CNBC.

The AI ​​opportunity is encouraging leaders at OpenAI and other companies to gain a deeper understanding of industrial design and energy generation.

Executives from nearly every major hyperscaler walked the factory floor, according to a person familiar with the visits who asked not to be identified because the details are not public.

The turbines are huge, 31 feet high and weighing 280 tons. One turbine can power nearly half a million homes.

“When we think about what the world needs for electrification and what we need to power this AI surge that we’re experiencing, a lot of it comes straight out of this factory,” Koziner said.

Microsoft purchased seven of them to power its data center in Texas. At 2.7 gigawatts, that’s enough electricity to power nearly 3 million homes.

GE Vernova turbines are already online at Elon Musk’s xAI Colossus 1 campus in Tennessee, and nearly another gigawatt is being deployed at OpenAI’s Stargate project in Texas, according to Cleanview, an organization that tracks data center development.

Demand for these machines far exceeds supply, with the order book due to be filled by 2029. Koziner added that the company is making more reservations for 2030 and even 2031.

“Today, about 20% of our gas power order book goes to the data center, which is an AI-type application,” he said.

One turbine could cost more than $250 million, according to industry estimates. According to analysts at Melius, the price has increased by 300% in the last 3 years. The sharp rise in prices underscores why AI capital expenditure budgets continue to rise; This is one of the main concerns among technology investors.

This spending increase has been a boon for GE Vernova; Its shares are up almost 60% in the last six months.

Public pressure on data center development and growing environmental concerns may challenge the development of artificial intelligence.

GE Vernova said it is working to make its turbines more environmentally friendly.

“We also spend a lot of time and effort on the sustainability of these machines,” Koziner said. “And the turbine you see here is twice as efficient as the turbine we would have built 20 years ago.”