Tencent Plans $3 Billion Dual-Currency Bond Issuance, Returns to USD Bond Market After Five Years to Prepare for AI Transformation

According to LightNode Analytics, Tencent Holdings is preparing to raise approximately $3 billion through a dual-currency bond issuance in US dollars and offshore Chinese yuan (Dim Sum bonds). The transaction could be priced as early as Tuesday. Tencent has reportedly received regulatory approval to issue up to $4.5 billion in offshore debt, but this issuance may not fully utilize the approved amount. If successfully priced, this will be Tencent's first US dollar bond issuance since 2021 and its second offshore Chinese yuan bond issuance following its debut last year.

Reportedly, Tencent has engaged investment banks to arrange this multi-tranche bond issuance, planning to issue 10-year and 20-year US dollar bonds, as well as 10-year and 30-year offshore Chinese yuan bonds. The proceeds will be mainly used for refinancing and general corporate purposes. This issuance comes as Beijing recently decided to allow some top-tier insurance companies to purchase Dim Sum bonds through an expanded Southbound Connect scheme, stimulating domestic demand for offshore Chinese yuan bonds.

In terms of debt repayment pressure, Tencent's management revealed during a roadshow conference call that the company has $3.5 billion in debt maturing this year, including bilateral loans, and $2.8 billion maturing next year. Bloomberg data shows that Tencent does not have any publicly issued offshore bonds maturing in 2026, but it has a $2.5 billion note maturing in 2028.

Behind the debt issuance and refinancing is Tencent facing accelerated pressure to transform into AI-first. Last month, Tencent reported its slowest revenue growth in six quarters. To reverse this trend, Tencent has pledged to at least double its AI investments this year, reaching over 36 billion yuan (approximately $5.3 billion). Currently, Tencent lags behind competitors such as ByteDance and Alibaba in user adoption and the development progress of large language models.