Table of Contents
ByteDance, the Beijing-based parent of TikTok, has arranged a $29.6 billion loan from a syndicate of nearly 30 banks to finance its artificial intelligence expansion, according to Caixin Global's reporting on the financing and PYMNTS' account of the deal. The three-year facility is unsecured, meaning ByteDance pledged no collateral against it, and was coordinated by Citigroup and JP Morgan Chase.
The company originally sought $20 billion but expanded the target after unusually strong demand from lenders. "It is very rare to see such a mega loan unsecured," one source involved in the deal said. "The banks practically are counting purely on ByteDance's name." The loan ranks as Asia's second-largest dollar-denominated borrowing of 2026, behind SoftBank's $40 billion commitment to OpenAI in March.
The proceeds will fund AI initiatives outside China, including capacity commitments at data centers across Southeast Asia. ByteDance has budgeted roughly $23.8 billion for AI spending in 2026, with more than half of that allocated to chips, on top of an additional $5 billion earmarked for computing infrastructure this year. The company's leadership has acknowledged that its large language models trail those of global leaders such as OpenAI and Google, even as it presses ahead with what it describes as a self-developed AI strategy.
The borrowing places ByteDance alongside Alibaba and Tencent in a broader race among Chinese technology companies to build AI infrastructure at a pace that increasingly mirrors US hyperscalers' capital spending. That spending has come at a cost: ByteDance's net profit fell by more than 70% in 2025, a decline the company has partly attributed to accounting changes but which also reflects the scale of investment required to compete in AI. The loan gives ByteDance additional runway to fund that buildout without diluting equity, at a moment when access to advanced chips and offshore computing capacity has become as strategically important to Chinese tech giants as the models themselves.
