ByteDance has secured a $29.6 billion loan from a global group of banks, giving the TikTok owner a huge pool of capital as it increases spending on artificial intelligence models and the infrastructure needed to run them.
Reuters reported that the three-year facility involves nearly 30 Chinese, U.S., European and Singaporean banks, with Citigroup and JPMorgan coordinating the financing. The company originally sought $20 billion but expanded the facility after strong demand from lenders.
Bloomberg likewise reported that ByteDance had increased the loan from an initial $20 billion target after receiving strong bank commitments. The borrowing is Asia’s second-largest dollar-denominated loan this year.
Lenders are making a major bet on ByteDance
Demand for the financing underscores how much capital banks are willing to provide to companies competing at the top of the AI market.
According to Reuters, Chinese banks subscribed to more than 60% of the total facility. The loan is unsecured, meaning ByteDance is not pledging shares or other assets as collateral.
One source told Reuters that such a large unsecured loan was extremely unusual and that lenders were effectively relying on ByteDance’s creditworthiness. The facility has an initial three-year term with options for another two years.
Bloomberg reported that the deal attracted more than $30 billion of orders before the commitment deadline, despite a weak period for Asian syndicated lending.
Only SoftBank’s $40 billion bridge loan, arranged in March to support investments in OpenAI, has been larger in Asia this year.
AI infrastructure sits behind the financing push
The company has been increasing investment across the AI stack. ByteDance was discussing purchases of AI chips from Shanghai-based Iluvatar CoreX for inference workloads and considering a similar transaction with Baidu.
Its international infrastructure ambitions are also significant. ByteDance is an offtaker for multiple data centers being built in Southeast Asia, meaning it commits contractually to buying portions of their computing capacity.
Lian Jye Su, chief analyst at Omdia, shared that ByteDance is competing with local hyperscalers in AI data centers while challenging global hyperscalers in multimodal AI models—both areas that require enormous investment.
AI competition is increasingly a financing competition
The scale of ByteDance’s borrowing shows that frontier AI competition is no longer determined solely by research talent or model quality.
Training and serving advanced models require chips, data centers, networking equipment and long-term computing commitments. Those requirements increasingly turn AI strategy into a balance-sheet question.
ByteDance’s previous global loan, raised in 2024, totaled $10.8 billion from about 20 lenders. Its latest financing is almost three times that size.
For founders and investors, the shift is significant: the largest AI players are increasingly competing not just over who can build better models, but over who can secure enough capital and infrastructure to keep scaling them.
