Thomson Reuters launched its first in-house AI model, Thomson-1, built on Alibaba's open-source Qwen. CTO Joel Hron cited cost control and IP ownership. The move draws scrutiny as US officials warn about Chinese open-source security risks.
Thomson Reuters launched its first in-house artificial intelligence model on Monday, built on an open-source system from Chinese tech giant Alibaba. The model, called Thomson-1, is based on Snowdon, which the company created by “realigning” Qwen, an open-source model from Alibaba’s cloud division, Chief Technology Officer Joel Hron said.
Open-source models are free to download and modify. That lets companies avoid the per-token fees charged by proprietary AI labs like Anthropic and OpenAI. Hron cited cost as one of the main reasons for building internally. He compared the decision to renting versus buying a house.
“Renting a house, you still have a roof over your head, and somebody's taking care of it, and it's great,” he said. “You're not building any equity that compounds into something valuable for you long term.”
Thomson-1 will initially handle document review tasks that had been outsourced to Anthropic’s Claude. Hron said the new model isn’t meant to replace the company’s work with Anthropic or other labs entirely. Thomson Reuters expanded its partnership with Anthropic in May for its AI legal assistant, CoCounsel, which still runs mostly on Claude. “Our main objective is to make Thomson the model that powers more and more of CoCounsel's capabilities over time,” Hron said.
A joint team from Thomson Reuters and Imperial College in the UK spent several months adapting Qwen to build Snowdon. Hron said the team ensured the model was “ethically and politically de-biased and safe to use.” He added that “there’s nothing that necessarily ties us to Qwen.”
The decision also lands Thomson Reuters in the middle of a debate over Chinese open-source models. Anthropic has said Chinese labs are illicitly “distilling” the outputs of its AI models to train their latest chatbots, and has called for the US to impose restrictions. Senator Tom Cotton has raised concerns about US companies using Chinese open-source models, citing potential security threats including backdoors.
Anthropic and Alibaba didn’t respond to requests for comment.
For Thomson Reuters, owning its own model means building on its own intellectual property instead of paying outside AI companies repeatedly. Hron made the rent-versus-buy comparison to explain the shift toward internal development.
The stock page for TRI shows an Alpha Score of 28/100, labeled Weak. BABA carries an Alpha Score of 49/100, labeled Mixed.
Prepared with AlphaScala editorial tooling from the source reporting linked above. Indexable analysis may include a cited Alpha Score value. Publishing checks screen each story before release. Educational coverage, not personalized advice.