AI's growth faces energy crisis: Ex-Google CEO Eric warns of looming power shortage

As artificial intelligence (AI) continues to dazzle the world — writing code, diagnosing diseases, and composing music — a stark warning has emerged from one of tech's most influential voices.

Former Google CEO Eric Schmidt has sounded the alarm over an impending crisis: the world's power supply may not be enough to sustain the future of AI.

In a recent episode of the Moonshots podcast, Schmidt, who now chairs the pro-AI think tank Special Competitive Studies Project, delivered a sobering forecast. "AI's natural limit is electricity, not chips," he declared, reports The Economic Times.

This challenge isn't hypothetical. Schmidt told listeners that the United States alone could require an additional 92 gigawatts of electricity to meet its AI demands—a need equal to building 92 nuclear power plants. "We need energy in all forms… and we need it quickly," Schmidt reiterated during testimony before the US Congress.

The world's top tech firms — OpenAI, Microsoft, Meta, and Google—are racing toward artificial general intelligence (AGI), machines capable of human-level reasoning. But that progress comes at a steep environmental cost. Schmidt's warning highlights a growing concern: while chips may become faster, they will be useless without the electricity to power them. "Chips will outrun power needs," he said.

This is no longer a theoretical issue. Microsoft has already signed a 20-year nuclear energy deal aimed at reviving the shuttered Three Mile Island facility, and OpenAI CEO Sam Altman has heavily invested in Helion, a startup focused on fusion energy. Meanwhile, companies are aggressively acquiring water rights and power contracts to support the vast cooling systems needed for their AI data centres.

According to Quartz, Microsoft's 2023 environmental report revealed a 34% increase in water usage, totalling 1.7 billion gallons — primarily for cooling AI-driven infrastructure. Researchers estimate that by 2027, AI workloads could consume enough water to supply Canada for an entire year.

Environmental organisations such as Greenpeace warn that AI's insatiable hunger for resources could severely undermine global climate objectives. Yet, the promise of superintelligence — AI systems that could revolutionise medicine, law, defence, and science — remains irresistible to investors and innovators alike.

"We don't know what AGI or superintelligence will ultimately deliver," Schmidt admitted, "but we know it's coming. And we must plan now to make sure we have the energy infrastructure to support it."

The tension is unmistakable. AI is often seen as a solution to some of humanity's greatest problems. But its unchecked development could strain — and possibly overwhelm — the very systems it aims to optimise. As Schmidt aptly warned: without sufficient electricity, there will be no intelligence — artificial or otherwise.