The Download: The startup that says it can stop lightning, and inside OpenAI’s Pentagon deal

OpenAI has reached a significant agreement with the US Department of Defense, reversing its previous prohibition and allowing the military to utilize its AI technologies in classified settings. CEO Sam Altman admitted the negotiations were "definitely rushed," initiated only after the Pentagon publicly reprimanded rival Anthropic for refusing similar terms. While OpenAI's blog post emphasizes built-in protections against autonomous weapons and mass surveillance, the deal raises immediate questions about the company's ability to enforce these safeguards as the military rapidly deploys AI in active geopolitical conflicts, such as recent strikes on Iran. This compromise tests OpenAI's internal culture, balancing commercial opportunity against employee concerns over ethical boundaries in military applications.

In parallel, startup Skyward Wildfire has secured millions in funding to accelerate development of its controversial technology designed to prevent catastrophic wildfires by stopping the lightning strikes that ignite them. The company's approach, suggested by online documents, involves seeding clouds with metallic chaff—narrow fiberglass strands coated with aluminum—a method first evaluated by the US government in the 1960s. However, researchers highlight major uncertainties regarding the technique's efficacy under varying conditions, the scale of material required, operational frequency, and potential secondary environmental impacts. The funding push underscores the urgent demand for wildfire solutions but advances a largely unproven and potentially ecologically disruptive intervention.