Thomson Reuters Develops a Frontier Model Trained on its High-Quality Datasets

In August 2026, Thomson Reuters launched Thomson, which is the first proprietary large language model (LLM) developed by the company. This model has been developed on the company's premises. Frontier Labs has conventionally invested billions of dollars to develop compute technologies and infrastructure to reach the frontier.

Thomson Reuters adopted a different approach; they began from a robust open-source base and allocated USD 40 million in funds to train the model to perform the most important tasks. Thomson Reuters has full control over this model and does not incur any inference costs, which are conventionally involved in frontier models.

Thomson Reuters has significant experience in developing professional content and analysis for its clients. The company has trained Thomson on several years of patented information, technologies, and expertise. This comprehensive dataset allowed the company to build a model to Fiduciary-grade standards and at a much lower total expenditure than its counterparts.

According to Joel Hron, the Chief Technology Officer of Thomson Reuters, AI companies used scale as a milestone, for example, larger models, more powerful compute, and higher investment. Thomson Reuters demonstrated that there is an effective alternative pathway. He said that the company started model development with a robust foundation, then specialized it comprehensively according to the tasks it is intended to perform.

Such a developmental methodology enabled the company to build a model that is customized to excel in the required tasks and under full control of the company. According to Steve Haskers, the CEO of Thomson Reuters, the company has demonstrated the quality of an AI model that can be built when it is trained on several years of proprietary data and editorial prowess.

Impact on the AI Industry

According to Precedence Research, the launch of the Thomson AI Model is expected to positively impact the AI Industry. Thomson Reuters used its rich dataset of patented data, technologies, and editorial prowess to develop a frontier model at much lower cost than its counterparts. The company demonstrated how even non-AI companies can build their own AI models by leveraging their massive datasets. This development will prompt many other corporations to use their historical and current data and technologies to build their own frontier models. This helps them to have robust control over the model. 

Impact on the Cloud Computing Market

The global cloud computing market size was estimated at USD 912.77 billion in 2025 and is predicted to increase from USD 1,106.28 billion in 2026 to approximately USD 5,946.84 billion by 2035, expanding at a CAGR of 20.61% from 2026 to 2035.

According to Precedence Research, the introduction of the Thomson AI Model is expected to positively impact the cloud computing market. This model has been trained on decades of patented data and technologies and performs critical tasks. Such models require a significant amount of computing power to process a large amount of data. Companies avail cloud services from prominent technology corporations to avoid significant upfront capital investments in cloud infrastructure. They can also scale their operations easily and avail the latest data processing and storage technologies.

Impact on the Cybersecurity Market

The global cybersecurity market size was valued at USD 301.91 billion in 2025, calculated at USD 339.96 billion in 2026, and is expected to reach around USD 969.45 billion by 2035. The market is expanding at a solid CAGR of 12.37% over the forecast period 2026 to 2035.

According to Precedence Research, the launch of the Thomson AI model is expected to positively impact the cybersecurity market. This AI model has been trained on patented data and technologies and hence will require advanced cybersecurity systems to prevent any cyberattacks.

The loss of data privacy can result in significant monetary and reputational losses for the company. It often causes a loss of trust among its clients, which usually takes years to build. Such models are usually hosted on hybrid cloud setups. Critical operations and data are hosted on a private cloud, and non-sensitive data is hosted on a public cloud.

About Thomson Reuters

Thomson Reuters is a prominent media company that provides products and services in various sectors, including legal, tax and accounting, risk and fraud, trade and supply, and news and media. The company also provides comprehensive technological expertise and content in the field of AI. Thomson Reuters simplifies complicated subjects and has been trusted by its clients for more than 150 years. The company is also committed to decreasing its environmental impact by using efficient technologies and responsible sourcing from its Global Print segment. Green Teams in the company undertake regional environmental drives across the world.