Thomson Reuters Launches Proprietary LLM, 'Thomson', on Data Foundation

MarTech Series5h ago·1 min readStrategy & Trends

AI Summary

Thomson Reuters has launched its own proprietary large language model, 'Thomson', built on its extensive data assets and domain expertise. This in-house model was developed at a significantly lower cost than comparable frontier models by leveraging open-source foundations and a $40 million investment, allowing Thomson Reuters to maintain full ownership and control.

⚡ Marketer Insight

Thomson Reuters' approach demonstrates that building powerful AI doesn't solely rely on massive compute budgets; a strong data foundation and specialized training can yield cost-effective, controlled frontier models, challenging the 'bigger is better' narrative in AI development.

#proprietary llm#data assets#ai development#frontier models

Original article

MarTech Series

Read full article →