In response to the increasing energy demands driven by artificial intelligence (AI), major technology firms Amazon and Nvidia have indicated a practical shift in their energy strategies, openly considering fossil fuels alongside renewable sources to power their expanding AI data centres.
This development was brought to light during a recent gathering at the Hamm Institute for American Energy in Oklahoma City, where executives from both the tech and energy sectors met to tackle the pressing challenge of securing reliable electricity for AI infrastructure.
Despite their long-standing commitments to sustainability—Amazon is the world’s largest corporate purchaser of renewable energy, and Nvidia actively pursues clean energy solutions—the immediate energy requirements for AI have prompted a reassessment of available power options.
Kevin Miller, Amazon’s vice president of global data centres, acknowledged the necessity of a diversified energy mix in the short term. “To have the energy we need for the grid, it’s going to take an all of the above approach for a period of time. We’re not surprised by the fact that we’re going to need to add some thermal generation to meet the needs in the short term,” Miller stated.
While Amazon is investing heavily in advanced nuclear technologies and carbon capture initiatives, these innovations are not expected to be operational until the 2030s, leaving natural gas as a crucial transitional fuel to maintain power stability.
Nvidia’s senior director of corporate sustainability, Josh Parker, echoed this sentiment, emphasising the unrelenting demand for power amid AI’s expansion. “At the end of the day, we need power. We just need power.”
Parker further noted that client priorities vary, with some insisting on clean energy while others are less concerned about the source, showing the complexity of balancing environmental goals with operational realities.
The scale of AI’s power consumption is staggering. Industry estimates, such as those from Anthropic co-founder Jack Clark, suggest that AI will require an additional 50 gigawatts of electricity by 2027-roughly equivalent to the output of 50 nuclear reactors. Clark urged stakeholders to adopt a realistic stance on energy availability, while also recognising that AI’s growth could stimulate the development of innovative, cleaner energy technologies over time.
The political scene has further complicated the energy equation. Under the Trump administration, there has been a marked pivot towards boosting fossil fuel production, notably natural gas and coal, with the latter receiving renewed attention through executive orders aimed at increasing coal output citing AI demand.
However, industry leaders remain cautious about coal’s role. When questioned, Amazon and Nvidia executives refrained from endorsing coal as a primary energy source.
“You have a broader set of options than just coal. We would certainly consider it, but I don’t think I’d say it’s at the top of our list,” Clark stated. This indicates a preference for cleaner and more flexible fossil fuels like natural gas over coal, which is less favoured due to its environmental impact.
Looking ahead, Amazon’s strategy includes significant investments in small modular nuclear reactors (SMRs), which promise a smaller footprint and quicker deployment compared to traditional nuclear plants. These efforts align with Amazon’s ambitious goal to achieve net-zero carbon emissions by 2040, a decade ahead of the Paris Agreement target.
The tech industry is also recognising nuclear power’s critical role in meeting AI’s insatiable appetite for electricity. Nvidia’s CEO Jensen Huang has emphasised nuclear energy as essential for the future, stating that renewables alone cannot satisfy the growing energy demands of AI and data centres. Similarly, other tech leaders, including those at Google and Microsoft, advocate for nuclear power as a reliable, carbon-free energy source capable of supporting the sector’s expansion.