How The AI Data Center Buildout Is Creating Boom For The Gas Turbine Industry
AI data centers operated by hyperscalers like Amazon, Google, and Meta are driving explosive demand for natural gas turbines, with GE Vernova leading production at its South Carolina facility. Turbine prices have surged 300% in three years, and the company now books orders extending to 2030-2031, though environmental concerns about this technology are mounting.
Summary
The transcript discusses how major technology companies are increasingly relying on natural gas turbines to power their AI data centers due to the massive energy demands of training and running AI models. Hyperscalers need power generation faster than local utilities can provide, making industrial gas turbines an attractive solution. GE Vernova operates the world's largest natural gas turbine manufacturing plant in South Carolina and has experienced rapid expansion to meet demand from Silicon Valley tech giants. The facility has been visited by executives from nearly every major hyperscaler seeking to understand power generation and industrial design. Notable deployments include Microsoft's purchase of seven turbines for a Texas data center (2.7 gigawatts capacity, enough to power over 3 million homes), turbines already operational at XAI's Colossus 1 campus in Tennessee, and nearly a gigawatt being deployed at OpenAI's Stargate project in Texas. The demand far exceeds supply despite turbines costing over $250 million each. GE Vernova's stock has risen nearly 60% in six months, and competitors like Caterpillar, Siemens, and Mitsubishi have also seen explosive growth. Turbine prices have increased 300% over three years. The company claims modern turbines are twice as efficient as those from 20 years ago, generating more power with less fuel consumption and reduced pollution. However, opposition is growing due to environmental concerns. Experts note that while gas turbines are currently a leading solution for large-scale firm power, they alone cannot satisfy AI's total energy demands and must be part of a broader energy ecosystem.
Key Insights
- Executives from nearly every major hyperscaler have visited GE Vernova's South Carolina factory floor to understand industrial design and power generation, indicating a strategic priority shift in how tech companies approach infrastructure
- About 20% of GE Vernova's gas power order book is now dedicated to data center and artificial intelligence applications, representing a significant portion of their business
- Turbine prices have increased 300% over the last three years despite persistent supply shortages, with individual turbines costing over $250 million
- GE Vernova's current turbines are two times more efficient than turbines produced 20 years ago, delivering more power with less fuel consumption
- Experts contend that industrial gas turbines are only one solution in a broader energy ecosystem needed to satisfy AI energy demands, and cannot alone meet total power requirements
Topics
Transcript
[0:01] Your AI chatbot may be powered by these massive natural gas turbines. Hyperscalers like Amazon, Google, Meta, Oracle, and SpaceX are building huge data centers that run AI models made by companies like Anthropic and OpenAI. Those data centers require a lot of energy, and in order to meet customer demand and stay competitive, they need it fast. Faster than local utilities can provide. So, this is what they're buying instead. Behind me is the inside of one gas [0:33] turbine, which in length is 31 ft long. Where the gas enters, it meets the water to create a mini explosion. On the other side is what comes, electric energy to power those data centers. GE Vernova is the…
Full transcript available for MurmurCast members
Sign Up to AccessMore from CNBC
GTA 6's Billion-Dollar Hype Machine
GTA VI is projected to be the most anticipated game launch in history, with an estimated $1 billion+ development budget expected to be recouped on day one through 25 million pre-sales at $100 each. The game's release is anticipated to have massive economic ripple effects across entertainment, streaming, and traditional business sectors, with nearly 1 in 3 gamers planning to purchase it.
The rise, reinvention and uncertain future of Nvidia without Jensen Huang
Nvidia founder Jensen Huang has been removed as CEO amid allegations of multiple affairs, including with direct reports, and connections to Alice Wong, who may be a Chinese spy involved in distributing AI chips across Southeast Asia. The scandal represents the company's biggest crisis in its 33-year history, occurring during a period of strict U.S. export controls on advanced chips to China.
Could Democratic Gains Bring Tariff Relief For Canada?
Following Democrats' 2026 midterm victories, Canadian investors anticipate potential tariff relief, but progress will likely be slow given President Trump's remaining two years in office and the complexity of trade negotiations. While some sectors like steel, lumber, and agriculture could benefit from reduced tariffs, Rust Belt Democrats and security concerns may limit the scope of relief.
Nvidia In Crisis: The Rise Of Jensen Huang’s Empire At The Center Of The AI Boom
This documentary examines Nvidia's rise to dominance in AI computing and the recent crisis surrounding founder Jensen Huang's resignation following allegations of affairs and potential involvement with a figure connected to Chinese AI chip diversion schemes. The narrative explores how Huang built Nvidia from a graphics card startup in 1993 into a $150+ billion company at the center of the AI boom, while also detailing geopolitical risks, export control complications, and questions about the company's future without its iconic founder.
How Democrats Won Back Union Voters
Democrats have dramatically reversed declining support among union voters, winning union households by 22 percentage points in the 2026 midterms compared to just 8 points in 2024. This turnaround was driven by aggressive union organizing efforts and increased voter enthusiasm, though questions remain about sustaining this advantage in future presidential elections.