Google Cloud announced that AI.AGG, a fully managed, multi-level semantic aggregation function natively embedded within BigQuery GoogleSQL, is currently in preview. The company said BigQuery has helped set standards for structured data aggregation for more than a decade, but relational SQL engines have historically lacked the native ability to aggregate, synthesize, and extract overarching semantic themes from millions of rows of text and multimodal data. Although BigQuery introduced row-level AI functions such as AI.GENERATE and AI.CLASSIFY, analyzing massive datasets holistically presents an architectural challenge due to the LLM context window, since millions of rows cannot be passed into a single prompt without severe truncation or hard execution failures. AI.AGG conceptualizes semantic aggregation as a distributed map-reduce problem, enabling data teams to run natural-language instructions over millions of rows using a single SQL query. The BigQuery engine orchestrates row-level validation, dropping or returning an error on payloads >10 MiB, context-safe batching, parallel inference via Vertex AI, and recursive synthesis until a single final aggregated string is produced per SQL GROUP BY clause. Because AI.AGG requires deep reasoning, it relies entirely on full-weight LLMs at every stage, differentiating it from row-level functions like AI.CLASSIFY, which can leverage distilled CPU proxy models.