Test report DSG-5589 · Rev D · tested November 4, 2026

Supply Chain & PolicyDevice under test

Oracle Drops $7 Billion Tencent AI Chip Lease Deal Under US Pressure

Oracle abandoned a $7 billion, five-year deal to lease Tencent up to 100,000 advanced AI chips via Southeast Asian data centres after US officials objected.

Read
2 min
Words
498
Node
14nm
Operator
Marcus Bennett

Spec summary

  1. Oracle withdrew from a $7 billion agreement with Tencent, reported November 4 by the Financial Times and cited by Reuters.
  2. The five-year deal would have let Tencent lease up to 100,000 advanced AI chips through Southeast Asian data centres.
  3. Nvidia has told cloud providers to direct at least 51% of services to non-Chinese customers.
  4. Chinese officials urged Tencent to support domestic chipmakers as US officials raised objections.
  5. ByteDance and Alibaba are among Southeast Asia's largest Chinese infrastructure clients.
Oracle Withdraws From $7 Billion Tencent AI Chip Deal Amid US-China Tensions - CXO Digitalpulse
Fig. AOracle Withdraws From $7 Billion Tencent AI Chip Deal Amid US-China Tensions - CXO Digitalpulse — blickpixel / pixabay

Oracle has withdrawn from a proposed $7 billion deal to lease advanced AI chips to Tencent, following pressure from the Trump administration. The Financial Times first reported the collapse on November 4, and Reuters cited the report.

The five-year agreement would have let Tencent lease up to 100,000 advanced AI processors through data centres in Southeast Asia. US export controls bar those chips from being sold directly to China. Routing the hardware through facilities outside China was meant to work around that restriction. It did not survive scrutiny from Washington.

Why did the deal fail?

Two governments pulled in opposite directions, and the deal snapped. US officials raised concerns about the arrangement, because it would have handed a Chinese company access to processors that American law keeps away from China. Chinese officials, meanwhile, urged Tencent to buy from domestic chipmakers instead.

That left the deal squeezed from both capitals. US technology firms now face growing scrutiny when they supply computing infrastructure to Chinese businesses — even when the servers sit outside China.

What does this change for Chinese AI firms?

The practical question is simple: where does Chinese AI training capacity come from now? Domestic chip supplies remain limited. Chinese-made processors cannot yet match the most advanced chips used for AI model training, which forces companies like Tencent to look abroad.

Tencent had specific plans for this deal. The company is developing its Hunyuan AI models and has increased its infrastructure spending to support them. The Oracle agreement would have expanded Tencent's AI computing capacity and strengthened its position in the AI market. That path is now closed.

The Nvidia precedent

The Oracle case fits a wider pattern of restrictions on how US cloud and technology providers serve Chinese customers. Nvidia has told cloud service providers to ensure at least 51% of their services go to non-Chinese customers, according to the report.

That requirement shows how Washington now polices compute access, not just chip exports. A Chinese firm cannot simply rent capacity from a US-linked cloud and get the same chips it cannot buy.

Southeast Asia as the contested middle ground

Southeast Asia has become the main destination for Chinese technology companies seeking computing infrastructure. ByteDance and Alibaba rank among the region's largest Chinese clients. Data centres in the region offer proximity, capacity and a legal position between the two powers' export regimes.

The Oracle-Tencent collapse signals that this middle ground is narrowing. Deals involving advanced processors, cloud computing and data-centre services now carry political risk that no contract clause can remove. Any arrangement that effectively transfers restricted computing power to a Chinese company invites the same objections that killed this one.

For US vendors, the message is blunt: a $7 billion revenue opportunity is worth less than a Washington objection. For Chinese buyers, the message is equally blunt: the most advanced training compute stays out of reach, and the domestic chip industry has a gap it must close alone.

via Google News: AI chip (Source)

Filed under

  • ai-chips
  • export-controls
  • oracle
  • tencent
  • ai-infrastructure
Share this article:

More from Marcus Bennett

Marcus Bennett

Show full bio

News editor covering regulation, policy and their impact on hardware layer of artificial intelligence at Die Signal.

591 articles

Same lot · LOT-360B

« Previous articleNext article »