A Miami-based startup claims to have solved a bottleneck holding back large language models for almost a decade. Subquadratic’s new model, SubQ, is faster, cheaper, and uses less energy than existing models. Independent evaluations suggest SubQ can process 12 times more text at once, matching the performance of top models like Google DeepMind and OpenAI. This breakthrough could change how LLMs are built, offering huge increases in speed at a fraction of the cost. Why it matters: If Subquadratic’s claims are validated, it could lead to a new era of efficiency in AI development, making LLMs more accessible and practical for a wide range of applications.