Open-Source AI Battle, Google Throttles Meta, Micron Margins Moon | Edward Coristine & Tai Groot, Chad Rigetti, Pim de Witte, Yadin Soffer, Jack Morris, Neil Movva, Jakob Diepenbrock, Chris Altchek

The rapid advancement of open-source AI, exemplified by the release of Zipu AI’s GLM 5.2, is narrowing the performance gap with U.S. frontier models and intensifying geopolitical competition. This shift occurs alongside significant infrastructure bottlenecks, where surging demand for high-bandwidth memory has turned chipmakers like Micron into trillion-dollar entities at the expense of AI model producers. Meanwhile, the startup ecosystem is pivoting toward specialized applications, including on-device privacy models, quantum-accelerated AI for data centers, and clinical automation for chronic disease management. As the industry moves toward background-task automation and continual learning, the focus is shifting from human-in-the-loop interactions to highly efficient, autonomous workflows. These developments underscore a broader trend where intelligence is becoming the primary bottleneck, prompting a convergence toward gaming-inspired action spaces and more efficient, specialized hardware architectures.