">

The Battle Among Tech Giants for AI Dominance: ByteDance Earns $20 Billion but Takes Out a $29.6 Billion Loan?", "ByteDance, Syndicated Loan, AI, Computing Power, Infrastructure

09/22 2026 513

ByteDance Borrows 200 Billion Yuan to Stockpile Computing Power

ByteDance Secures a 200 Billion Yuan Loan!

According to Bloomberg in September, ByteDance has secured a syndicated loan of up to $29.6 billion, equivalent to approximately 200 billion yuan.

Interestingly, ByteDance initially planned to borrow only $20 billion. However, once the news broke, nearly 30 top-tier banks, including Industrial and Commercial Bank of China, Bank of China, HSBC, and JPMorgan Chase, flocked to offer funds, resulting in the loan amount being increased to $29.6 billion.

This marks the second-largest dollar-denominated loan in the Asian market in 2026 and one of the largest offshore borrowings in the history of Chinese private enterprises.

According to a September 15 disclosure by The Information, ByteDance's revenue in the first half of 2026 reached approximately $120 billion, up about 30% year-on-year, with a net profit of around $20 billion, matching Meta's revenue.

ByteDance's revenue continues to grow rapidly, and its profits remain substantial. Founder Zhang Yiming has just become Asia's richest person, yet the company suddenly borrowed 200 billion yuan, leaving many wondering: Is ByteDance short on cash?

However, the answer becomes clearer when we consider how much ByteDance plans to spend next. This time, ByteDance is facing an increasingly capital-intensive AI battle.

Securing 200 Billion Yuan Without Collateral

Let's first examine the loan itself.

ByteDance's syndicated loan amounts to approximately $29.6 billion, or about 200 billion yuan, with an initial term of three years, extendable up to five years. The loan was arranged by institutions such as Citigroup and JPMorgan Chase.

The lineup of participating banks is impressive, attracting 28 banks from China, the United States, Europe, and Singapore. Fifteen Chinese banks formed the "rear guard," collectively committing about $18.9 billion, accounting for more than 60% of the total.

Among them, the Industrial and Commercial Bank of China alone committed $3 billion, while Bank of China and China Construction Bank secured $2.5 billion and $1.5 billion, respectively. HSBC and two other foreign banks also made significant commitments, each lending $1.5 billion.

This is the second-largest dollar-denominated loan in Asia in 2026, second only to SoftBank Group's $40 billion bridge loan in March for investing in OpenAI.

This is not a case of a single bank writing a massive check for ByteDance; instead, a group of banks is sharing the risk, which is the essence of a syndicated loan. In other words, when a company needs to borrow a large sum, but one or two banks are unwilling to bear the entire risk, dozens of banks contribute funds together.

For ByteDance, the biggest advantage is the ability to secure a large amount of capital at once. For the banks, it allows them to spread the risk across different institutions.

Moreover, ByteDance is not only borrowing a large amount but also doing so at a low cost. The initial interest rate spread for the loan is approximately SOFR + 68 basis points, 17 basis points lower than ByteDance's 2024 level of SOFR + 85 basis points, making it one of the lowest spreads obtained by a Chinese tech company in the offshore market.

Beyond the enormous amount, the loan's terms are also worth examining.

This is an unsecured credit loan. In the banking industry, it is almost unthinkable for a multi-billion-dollar dollar project to proceed without equity or physical assets as collateral. Banks' willingness to lend "unsecured" is a vote of confidence in ByteDance's strong cash flow and corporate credit, as they see sufficient cash-generating and debt-repaying capabilities when evaluating the company.

According to a September 15 disclosure by The Information, ByteDance's full-year revenue in 2025 reached approximately $200 billion, up about 29% year-on-year. By the first half of 2026, revenue had hit $120 billion, up about 30% year-on-year.

Moreover, growth did not rely solely on a single market. Overseas revenue from TikTok accounted for 30% of the total, up from 25% in 2024. TikTok Shop's global GMV reached approximately $50.3 billion in the first half of 2026, up 92% year-on-year.

Why would such a cash-rich internet giant voluntarily increase its leverage?

The answer lies in AI.

AI "Forces" ByteDance to Take on Massive Debt

While revenue continues to grow, ByteDance's profit margins have been declining for three consecutive years, with net profit margins falling from 26% in 2023 to 21% in 2024 and about 16.7% in the first half of 2026.

The primary driver of the declining profit margins is AI.

According to Caixin, ByteDance plans to invest nearly 160 billion yuan in AI in 2026, with more than half allocated to AI chip procurement, about 20 billion yuan for large models like Doubao and AI application iteration, and about 5 billion yuan for top AI talent and R&D. Meanwhile, the company plans to invest about 50 billion yuan in the construction and expansion of AI data centers, server clusters, and other computing power infrastructure at home and abroad.

ByteDance is also considering raising its capital expenditures to up to $70 billion in 2026, more than double the 2025 level.

In the protracted AI competition centered around large models, computing power is the "hard currency." From the 4.5 billion yuan Taihang Computing Center in Datong, Shanxi, to the planned multi-gigawatt smart computing cluster in Ulanqab, Inner Mongolia, and the "computing power enclave" being built in Malaysia, ByteDance is rapidly deploying AI infrastructure.

On the software side, the massive user base is also accelerating the depletion of computing power. As of June 2026, Doubao's monthly active users had surpassed 382 million, firmly occupying the top spot among domestic AI-native apps. The daily token usage of Doubao's large model on Volcano Engine exceeded 180 trillion, more than tripling in six months.

Behind every conversation, every image generated, and every video rendered lies soaring electricity bills and equipment depreciation costs.

Thus, ByteDance is choosing to match long-term assets with long-term funding. Infrastructure such as servers and data centers represents long-term assets with lifespans of several years to a decade. Relying solely on short-term operational cash flow to support them could easily drain the company's liquidity, something even ByteDance would struggle with.

Borrowing low-cost, long-term dollar debt with 3- to 5-year maturities during a low-interest-rate window to match long-term investments is a typical corporate finance practice.

Beyond cash management, another financial challenge arises: ByteDance does have cash on hand, but most of it is domestic and denominated in yuan.

Acquiring high-end chips globally and expanding overseas data centers all require U.S. dollars. Securing a low-cost offshore dollar loan overseas avoids exchange rate risks and regulatory processes while preventing additional financial friction costs.

Thus, ByteDance's $29.6 billion loan is about preparing ammunition for the "AI war" in the coming years to ensure stable output.

Financial Competition Among Tech Giants: Alibaba Raises Equity, Tencent Issues Bonds

Looking across the internet technology sector, ByteDance's "borrowing" is not an isolated case.

In 2026, China's internet industry is witnessing a "borrowing spree" among major players, with Alibaba, Tencent, and ByteDance leading the way. While their financing methods differ, the underlying logic is similar.

Alibaba opted for a share placement. In August, Alibaba announced the placement of approximately 80 billion Hong Kong dollars in new shares, yielding a net amount of about 79.7 billion Hong Kong dollars after deducting related fees. The purpose of these funds is clear: to invest in full-stack AI capabilities and strengthen AI infrastructure.

While Alibaba is not short on cash, its expenditures are substantial. As of the end of June 2026, Alibaba's cash and other liquid investments reached 474.5 billion yuan. However, its capital expenditures in the second quarter of 2026 hit 67.678 billion yuan, up 75% year-on-year. The previously announced plan to invest at least 380 billion yuan in AI infrastructure over three years had already seen approximately 190 billion yuan invested by the end of June 2026.

Thus, Alibaba's choice is to avoid depleting its cash reserves entirely and instead secure long-term funds through equity financing.

Tencent, meanwhile, chose to issue bonds. In June, Tencent issued $2.45 billion in notes and 15 billion yuan in dim sum bonds, including 10-year and 30-year maturities. According to Tencent's management, this financing is related to AI capital expenditures.

Tencent has stated that its AI investments in 2026 will at least double from last year. In the second quarter of this year, Tencent's capital expenditures reached 52.78 billion yuan, up 176% year-on-year. Tencent President Martin Lau said investors recognize (recognize) Tencent's deployment of AI capital expenditures.

It's not just these three Chinese giants; U.S. tech giants are also going all-in.

Alphabet, Amazon, Microsoft, and Meta are expected to spend a total of $725 billion on capital expenditures in 2026, up 77% from 2025. Goldman Sachs projects that capital expenditures by hyperscale cloud providers will rise to about 100% of operating cash flow in 2026, meaning they are investing nearly all their internal cash flow into AI infrastructure.

For today's internet technology companies, the biggest cost of AI is not just money but also time. Data centers need to be built in advance, chips procured ahead of time, talent secured early, and models continuously trained.

Waiting until everyone sees the demand before preparing may already be too late. This is the common logic behind Alibaba's share placement, Tencent's bond issuance, and ByteDance's loan:

They are not financing past operations but placing early bets on future AI infrastructure.

References:

"Why Did Cash-Rich ByteDance Suddenly Borrow 200 Billion Yuan?" Caijing WEEKLY

"ByteDance Secures $29.6 Billion Loan" Caixin

"Asia's Second-Largest Loan of the Year: ByteDance to Secure $29.6 Billion Syndicated Loan" Jiemian News

"China's Richest Companies Are Suddenly Borrowing Like Crazy" Huashang Taolue

"Alibaba Plans to Place 80 Billion Hong Kong Dollars in Shares as AI 'Cash Beast' Tests Big Tech's Wallets" Yicai

"This Year's Largest Dim Sum Bond Issuance! Tencent's Offshore Bond Issuance Targets AI" China Fund News

THE END",

Solemnly declare: the copyright of this article belongs to the original author. The reprinted article is only for the purpose of spreading more information. If the author's information is marked incorrectly, please contact us immediately to modify or delete it. Thank you.