ByteDance borrows $30B to invest in AI infrastructure

ByteDance just convinced nearly 30 banks to hand over $29.6 billion with no collateral attached.

The unsecured syndicated loan, coordinated by Citigroup and JPMorgan, was originally pegged at $20 billion. Then demand piled up past $30 billion, and ByteDance decided to upsize the facility to $29.6 billion. The deal represents the second-largest dollar-denominated borrowing in Asia for 2026, trailing only SoftBank’s $40 billion facility from March, which was tied to its OpenAI dealings.

Where the money is going

ByteDance plans to channel the proceeds into overseas AI infrastructure, with a particular focus on building data centers across Southeast Asia. The company is reportedly weighing capital expenditures of up to $70 billion in 2026, with projections stretching toward $100 billion in 2027.

Amazon, Alphabet, and Microsoft are collectively expected to spend a combined $725 billion on AI-related capital expenditure in 2026.

US export restrictions on advanced chips have made ByteDance’s path to AI dominance considerably more complicated. Building data centers in Southeast Asia serves a dual purpose: it puts infrastructure closer to fast-growing markets, and it creates supply chain flexibility around Washington’s technology chokepoints.

Lenders are betting big on ByteDance’s credit

Perhaps the most telling detail in this deal is the pricing. The loan opened at 68 basis points over the secured overnight financing rate (SOFR). For context, ByteDance’s previous loan in 2024 priced at roughly 85 basis points over SOFR, and that was for a significantly smaller sum.

Chinese banks took more than 60% of the total allocation. The remaining allocation was spread among international institutions.

The unsecured nature of the facility deserves emphasis. Most corporate borrowings of this scale come with assets pledged as security. ByteDance convinced lenders to forgo that safety net entirely.

The AI arms race gets a new entrant at scale

ByteDance has been steadily building its AI capabilities over the past two years. Its large language models power features across TikTok and Douyin, from content recommendation to generative tools. But this loan signals something more ambitious: a desire to compete at the infrastructure layer, not just the application layer.

The risk, of course, is execution. Spending $70 billion to $100 billion annually on capex requires operational discipline that few companies in history have managed at that scale. And ByteDance faces regulatory uncertainties on multiple fronts, from the ongoing TikTok divestiture saga in the US to evolving data sovereignty laws across Southeast Asia.

Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.