The Download: OpenAI’s turning point for math and a battery record

OpenAI recently announced that it has solved one of mathematics’ most important unsolved problems, the Navier-Stokes equations, which describe fluid motion. This feat was achieved using 10,000 AI agents working together for 88 hours. The equations, first formulated in the 1930s, are one of the seven Millennium Prize Problems, each carrying a $1 million reward for a correct solution. While OpenAI claims this is the first major math problem solved by AI, the announcement has sparked controversy. Critics argue that the company failed to properly credit researchers whose earlier AI-assisted work contributed to the solution. This breakthrough highlights the growing role of AI in advancing mathematical research, raising questions about the future of human mathematicians in a field increasingly dominated by machines. The achievement also underscores the significant computational resources required, which are often only available to a few frontier AI companies.

The United States installed 20.2 gigawatt-hours of new battery capacity in the second quarter of 2026, a new record. This amount of energy is sufficient to power the daily electricity needs of 600,000 homes. The surge in installations is driven by falling battery costs and the urgent need for more energy storage as renewable energy sources like wind and solar are integrated into the national grid. The growth in battery capacity is uneven, with significant differences between large-scale grid batteries and smaller residential systems. Analysts note that this trend is reshaping the energy landscape in the U.S., supporting the transition to cleaner energy while addressing the intermittency challenges of renewables. The record-breaking quarter puts the country on track for another record year in 2026.

Danijar Hafner, a 31-year-old entrepreneur, is developing AI agents capable of planning and navigating environments they were not trained on. His approach involves training these agents using world models, which simulate real-world environments, and testing them in popular video games. Hafner’s work has recently expanded from virtual simulations to physical robots, with humanoid robots hanging in his San Francisco office as part of his new venture. His goal is to enable AI systems to generalize their learning and adapt to new, unseen situations. Hafner is recognized as one of MIT Technology Review’s 35 Innovators Under 35 for 2026, highlighting the significance of his contributions to AI research. This work could have broad applications, from robotics to autonomous vehicles, by improving how AI systems understand and interact with the physical world.

Ukraine is leveraging data collected from drones operating on the battlefield to create a new marketplace for military and commercial applications. The country has made millions of data points from tens of thousands of drone flights available to defense contractors and companies, turning war zones into training grounds for AI models. This initiative aims to attract funding and partnerships but raises ethical and regulatory concerns. Experts warn that treating battlefield data like ordinary commercial material could lead to unchecked use without proper oversight. The practice highlights the dual-use nature of AI technologies, where innovations designed for defense can quickly spill into civilian applications. There is a growing call for regulatory frameworks to ensure responsible handling of such sensitive data.

The U.S. government has accused six Chinese AI companies, including DeepSeek, Moonshot AI, Alibaba, and Z.AI, of engaging in industrial-scale theft of American AI technology. Officials allege that these companies stole trade secrets, including through a technique called model distillation, where AI systems are trained using outputs from other models rather than original data. The accused companies are also reported to have targeted popular AI models like Claude, ChatGPT, Gemini, and Grok. The accusations come amid rising tensions between the U.S. and China over technology transfer and intellectual property. The U.S. government’s move reflects concerns about China’s rapid advancements in AI and its potential impact on global competitiveness and security.

The Pentagon has requested that OpenAI develop an AI model with minimal refusal rates, meaning it would rarely decline to comply with requests. This initiative is part of a broader effort by the U.S. military to integrate AI into defense operations, including training on classified military data. The request has sparked debate about the ethical implications of AI systems designed to obey commands without question. Critics argue that such models could lower safety standards and increase the risk of unintended consequences in military applications. The Pentagon’s push for AI adoption highlights the growing intersection of technology and defense, as well as the challenges of balancing innovation with ethical and security concerns.

Apple is expected to unveil a new $2,000 folding smartphone, marking the biggest design change to the iPhone since its launch in 2007. The device, which folds to a more compact size, will be the first major test for Apple’s new CEO, John Ternus. Folding smartphones have gained popularity in recent years, with competitors like Xiaomi and Huawei already releasing their own models. Apple’s entry into this market could significantly influence consumer adoption and industry trends. The high price point reflects the cutting-edge technology involved, including flexible displays and advanced engineering. This launch is closely watched as a potential turning point for the smartphone industry.

Meta has introduced a new AI agent called Muse that can autonomously use apps and websites on behalf of users, such as sending emails and making payments. Internal tests revealed that the agent could expose sensitive personal data, raising privacy and security concerns. Meta’s move reflects the growing trend of AI systems integrating more deeply into daily digital interactions. However, it also highlights the risks of AI agents operating without clear boundaries or safeguards. Experts caution that such agents could inadvertently share private information or perform actions without user consent. The launch of Muse underscores the need for robust privacy protections as AI becomes more autonomous and capable.

Google has announced that it will degrade the quality of its search results in Europe to comply with the European Union’s Digital Markets Act (DMA). The changes aim to give more prominence to comparison sites and reduce Google’s dominance in search results. This follows a €460 million antitrust fine imposed on Google by the EU in 2023. The DMA is part of a broader regulatory effort to curb the power of large technology companies and promote fair competition. Google’s adjustments to its search algorithm could lead to less relevant results for users, as the company prioritizes compliance over search quality. This move reflects the increasing influence of regulation on how tech companies operate globally.