5 carefully selected reads across AI, business, and investing.
Data from OpenAI and new open-model benchmarks reveal a massive intensification in AI utilization and technical capability. Meanwhile, global AI expansion faces headwinds from inflationary hardware costs and rising geopolitical pressure for sovereign infrastructure. These shifts suggest an increasingly fragmented landscape as European leaders and global markets navigate regulatory and capacity-based dependencies.
Internal data shows a massive spike in token utilization across OpenAI departments, with median usage in Research growing 56x and Engineering 27x since November 2025. This rapid deepening of AI integration serves as a key indicator for enterprise adoption trajectories.
Source: Latent SpaceThe GLM-5.2 model family is setting new standards in coding and agentic reliability, with GLM-5.2 Max scoring 1595 on Code Arena and demonstrating high agent reasoning capabilities. Significant hardware-level optimizations have further increased throughput to 392 tok/s.
Source: Latent SpaceRapid infrastructure expansion for AI data centers is driving a third wave of inflation, specifically through increased demand for memory chips. Economists are monitoring whether expected productivity gains will arrive fast enough to offset these rising costs.
Source: ChartbookAt the Cerebral Valley AI Summit in London, Cohere CEO Aidan Gomez warned against excessive reliance on American foundation models. Investors noted that European leaders are increasingly seeking sovereign AI capacity to move away from total dependence on U.S. infrastructure.
Source: NewcomerApple has confirmed that its 'Siri AI' features will not launch in Europe due to ongoing conflicts with the Digital Markets Act. This move highlights the growing bifurcation of the global AI product market as regulatory friction intensifies.
Source: Stratechery