OpenAI and the Ukrainian government announced a partnership at OpenAI's New York office that gives Ukrainian cybersecurity officials access to advanced models through the company's Daybreak programme, together with a pledge of more than US$1 billion in subsidised tokens. Ukrainian officials described the intended uses as cybersecurity automation — incident response, threat triage, login analysis, system inventory, code analysis and vulnerability validation — with the stated purpose being speed rather than substitution: the volume of logs defenders must read is the constraint, and officials were explicit that the aim is to make the cyber defender act faster, not to replace them. Ukraine reports roughly 6,000 cyberattacks a year, about 15 a day, and its critical infrastructure has been under sustained Russian attack across twelve years, including the 2015 grid shutdown and the 2025 railway attack, which is why the programme is framed around keeping the country running rather than protecting computers. For other operators, the signal is that subsidised defensive AI at national scale is now a live procurement model: US Senate staff told the panel there is broad bipartisan support for similar adoption by US critical-infrastructure operators, with regulation the remaining obstacle.

• Source article