Chinese AI firms are partnering with emerging-market nations to develop home-grown large language models tailored to local languages and industrial needs, according to a Xinhua report from Hefei on August 11, 2026. At Safaricom, Kenya's largest telecom operator, a digital human supported by a Chinese tech company demonstrates business services and interprets financial data. Experts warn that countries with less-used languages risk being sidelined by AI, as leading global models prioritize widely spoken languages. Chinese firm iFLYTEK has ramped up multilingual R&D; its speech model Spark supports speech recognition and simultaneous interpretation for more than 130 languages. In cooperation with south China's Guangxi Zhuang Autonomous Region, iFLYTEK launched the Spark ASEAN Multilingual Large Model Base, covering 10 languages including Malay, Indonesian, Vietnamese and Thai, with performance comparable to top international models at smaller parameter sizes. The company signed deals with partners in Laos, Malaysia and Thailand, and jointly set up an AI multilingual intelligent dubbing and translation center in Malaysia with Enjoy TV & Film Broadcasting Corporation, supporting translation for over 130 languages. "Mainstream English large models show clear shortcomings when serving many overseas markets. There is strong, genuine demand for large models customized for local tongues," said Dong Bin, iFLYTEK vice president. The company has open-sourced its model, allowing nearly 600,000 overseas developer teams to build custom products, and is pushing overseas roll-out of AI solutions for education