Sam Altman: Please Let OpenAI Into the Energy Grid

OpenAI CEO Sam Altman has held meetings with executives from Duke Energy, Exelon, Southern Co. and NextEra Energy since late July 2026, pitching OpenAI's Daybreak security model as a defense against AI-powered hackers targeting the U.S. electrical grid, Politico reported. OpenAI head of global energy policy John McCarrick told Politico that utilities are "different from the big banks" because regulations limit rate increases, and the outreach continued at an Edison Electric Institute meetup in Colorado Springs, Colorado this week. The pitch comes as 90% of state government CIOs named cyberattacks on critical services a top concern, according to a report from the National Association of State Chief Information Officers and General Dynamics Information Technology, and after electricity prices rose 267% in areas near AI data center hubs between 2020 and 2025, per a Bloomberg/DC Byte analysis.

OpenAI CEO Sam Altman has held meetings with officials from major power companies about electrical grid security—and surprise, surprise, his proposed solution to AI-powered hackers is more AI, Politico reported https://www.politico.com/news/2026/09/10/sam-altman-pitches-utilities-on-ai-grid-defense-01070425 . The meetings between Altman and industry execs began in late July 2026 and are ongoing, according to Politico, including a discussion at an Edison Electric Institute meetup in Colorado Springs, Colorado this week. The list of companies present to hear Altman pitch OpenAI’s Daybreak security model https://gizmodo.com/daybreak-openais-answer-to-anthropics-project-glasswing-has-arrived-2000757349 included Duke Energy, Exelon, Southern Co. and NextEra Energy. Many critical infrastructure entities across the country rely on operational technology that is decades old, and was never designed with cybersecurity in mind. Modern systems often don’t fare much better https://www.fortinet.com/blog/operational-technology/while-ot-security-is-maturing-risk-is-not-slowing-down . If those systems are reachable via the internet, whether via modernization or accident, they can become juicy targets https://www.wsj.com/articles/cyber-threats-against-heavy-industry-intensify-c27a157a for hackers. A recent report from the National Association of State Chief Information Officers and General Dynamics Information Technology emphasized https://statescoop.com/states-confront-growing-cyber-gaps-as-critical-infrastructure-threats-rise-report-finds/ that many state and local utilities lack the scale to have dedicated security staff. Ninety percent of state government CIOs said cyberattacks on critical services are one of their top concerns. OpenAI official John McCarrick, head of global energy policy, told Politico the company realizes infrastructure entities are “different from the big banks,” because regulations limit rate increases. Utilities are generally mandated to sell electricity at fixed rates with no markup for profit, meaning they can’t jack up the price to pay for arbitrary tech expenditures. On the flip side, whenever demand outstrips supply, they can punt the increased cost directly to customers. OpenAI’s own industry is a key driver of this. Electricity prices managed to rise 267% in areas near AI data center hubs over the years 2020 to 2025, according to a Bloomberg/DC Byte analysis https://www.bloomberg.com/graphics/2025-ai-data-centers-electricity-prices/ . AI-powered cybersecurity products often market themselves as the only solution to AI-powered threat actors note the circular relationship . Altman is perhaps not the wisest pick of vendors, given just weeks ago a swarm of OpenAI bots escaped a sandbox and hacked competitor Hugging Face https://www.cbsnews.com/news/openai-hugging-face-hack-ai-risks/ . Frontier models like Anthropic’s access-limited Mythos have allegedly demonstrated abilities stretching beyond vulnerability detection into uncanny territory like chaining exploits https://www.darkreading.com/application-security/anthropic-exploit-writing-mythos-ai-safe in novel ways, so AI boosters might have a point in this instance. That said, giving experimental AI systems designed by a cartoon villain https://www.newyorker.com/magazine/2026/04/13/sam-altman-may-control-our-future-can-he-be-trusted access to the power grid sounds like the kind of decision you make right before Skynet kill-bots burst through the wall.