ByteDance, Alibaba, Tencent Order Over $35 Billion in Nvidia Chips

China's three largest technology buyers — ByteDance, Alibaba and Tencent — have placed Nvidia chip orders worth more than $35 billion, according to an exclusive report relayed by WTAQ. The figure covers orders for Nvidia processors placed by the three companies, who together rank among the U.S. chipmaker's biggest customers worldwide.

The $35 billion-plus total lands at a moment when Nvidia's China business sits under layered U.S. export controls. Those restrictions cap which chips American vendors can legally ship to Chinese customers, and the rules have tightened repeatedly since 2022. That makes the size of these orders the story: three sanctioned-market buyers are still committing tens of billions of dollars to Nvidia silicon.

Who is buying, and what do we know?

The report names three companies:

• ByteDance — the parent of TikTok and its Chinese sibling Douyin, and a heavyweight in AI-driven recommendation and video processing.
• Alibaba — the e-commerce and cloud operator whose cloud unit, Alibaba Cloud, is one of China's largest AI compute providers.
• Tencent — the gaming and social-media giant behind WeChat, with a large and growing AI inference workload.

All three operate hyperscale data center fleets and have raced to build out AI training capacity since the generative-AI boom began. Nvidia's GPUs (graphics processing units, the accelerators that train and run large AI models) remain the industry's default hardware for that job.

The report does not break the $35 billion down by company, by chip model, or by delivery timeline.

What does this change for Nvidia and the China market?

It signals sustained Chinese demand despite the regulatory chokehold. Washington's export rules have already forced Nvidia to design cut-down chips specifically to stay within performance limits for Chinese customers. Each new order wave from ByteDance, Alibaba and Tencent tells investors that demand for compliant Nvidia silicon is holding, not collapsing.

For the buyers, the calculus is straightforward. Domestic Chinese alternatives — Huawei's Ascend line above all — are scaling, but supply remains tight and the software ecosystem still trails Nvidia's CUDA (the parallel-computing platform most AI code is built on). Paying up for Nvidia hardware is still the fastest route to compute.

For the broader market, an order book above $35 billion from just three customers underlines how concentrated AI infrastructure spending has become. A handful of hyperscalers on both sides of the Pacific account for a large share of Nvidia's revenue.

Why the number matters now

Nvidia has repeatedly flagged China as its largest single geographic opportunity that it cannot fully serve under current rules. Wall Street watches every signal from that market for evidence of either pent-up demand or genuine erosion.

A combined commitment of this size, from these three names, lands firmly in the first camp. It also pressures policymakers in Washington, where hawks argue that any Nvidia silicon sold into China ultimately strengthens rival AI capability, and industry argues the alternative is losing the market to domestic Chinese chips.

No delivery schedules, chip specifications or per-company allocations were disclosed. WTAQ attributed the figures in its exclusive report.