The Chinese AI StartupDPSEC has not yet determined the release time of its R2 model because CEO Liang Wenfering is not satisfied with its performance, citing two people with information about this information on Thursday.
Reuters reported earlier this year, the successor of the DiPsic’s wild R1 argument model, in May, was planned to release with the goals of creating better coding and reasoning in the languages beyond English.
In the past several months, Depsy’s engineers have been working to improve the R2 until Liang has given a green light for release.
However, the lack of NVIDIA server chips in China can be difficult to adopt R2 as a result of US export regulations, the report states that the employees of top Chinese cloud firms, referring to employees who offer dipseic models to enterprise users.
The report states that the possible increase in demand for R2 will overwhelm Chinese cloud providers, who need modern NVIDIA chips to run the AI model.
DiPsic did not immediately respond to Reuters’ request for comment.
The report states that DiPsic is in touch with some Chinese cloud companies, and provides them with technical specifications to guide the model hosting and distribution plans from their servers.
Information states that it is currently used in cloud users, the majority is running the model with H20 chips of NVIDIA.
Prevention of fresh exports imposed by the Trump administration in April has prevented NVIDIA from selling their H20 chips in the Chinese market – only AI processors that can legally export to the country.
© Thomson Writers 2025
(This story has not been edited by the NDTV staff and has been made auto from the Syndicate Fed.)
