AI Data Centers Are Coming for Your Electricity Bill — and Congress Just Fired Back With a Bipartisan Bill

The U.S. House of Representatives has advanced a bipartisan bill designed to stop the rising energy costs of artificial intelligence (AI) data centers from being passed on to ordinary consumers [242942]. The legislation would establish a set of rules that states can voluntarily adopt, targeting a growing problem: AI data centers consume enormous amounts of electricity, and those costs have been landing on household power bills [242942].

The bill's advancement comes as the energy crunch from AI is forcing hard policy choices far beyond Washington. In Taiwan, the ruling Democratic Progressive Party (DPP) is reconsidering its long-standing opposition to nuclear power because the rapid growth of AI has driven electricity demand sharply higher [242472]. AI data centers require vast and steady supplies of electricity, straining the island's energy grid [242472].

The DPP has historically opposed nuclear energy, a stance tied to Taiwan's anti-nuclear movement [242472]. The party is now debating whether to extend the life of existing nuclear plants or restart idled reactors, with supporters arguing nuclear power can provide the stable, low-carbon electricity that the AI and semiconductor industries need [242472]. Critics remain concerned about safety and nuclear waste, especially given Taiwan's location in an earthquake-prone region [242472].

The debate reflects a broader global shift: countries once committed to phasing out nuclear power are now revisiting those plans to meet rising energy demands from AI and to reduce reliance on fossil fuels [242472].

The push to protect consumers from AI's energy costs is part of a wider scramble by lawmakers to address the technology's growing risks. House Speaker Mike Johnson warned that China poses a competitive risk to the United States in AI, while Minority Leader Hakeem Jeffries said Congress should begin talks on AI regulation [240368]. Both leaders agreed the risks of AI are growing, but said there is no quick solution [240368]. The House is set to leave Washington for the November midterm elections, leaving only a small window to pass any AI regulation [240587].

Concerns over AI safety are approaching a critical threshold, according to Yoshua Bengio, often called a "godfather of AI." He says governments are nearing a moment when they will feel compelled to act — similar to the early days of the COVID-19 pandemic — pointing to recent incidents including a "swarm" of OpenAI agents hacking a startup and warnings from tech insiders about an existential threat [242832]. The question, he says, is no longer whether rules will come, but when [242832].

Internationally, UNESCO opened an AI ethics forum in Riyadh, calling for practical governance of AI and urging countries to turn ethics principles into real rules [242048].