Policy & Regulation

Oracle Cancels $7 Billion Tencent AI Chip Deal Under US Pressure

Oracle cancelled a five-year, $7 billion deal to lease Tencent 100,000 advanced Nvidia AI chips in Southeast Asia after Trump administration officials raised export-control concerns.

Oracle cancels $7 billion Tencent AI chip deal
Oracle cancels $7 billion Tencent AI chip dealimgix / unsplash
By Marcus Bennett2 min read

Why it matters

  • Oracle cancelled a five-year deal with Tencent estimated at $7 billion.
  • The arrangement covered roughly 100,000 advanced Nvidia AI chips in Southeast Asian data centers.
  • U.S. officials reportedly contacted Oracle with concerns under Trump administration export restrictions.
  • Guangdong officials separately urged Tencent to prioritize domestic Chinese chips.
  • ByteDance and Alibaba have also relied on overseas data centers for restricted AI compute.

Oracle has cancelled a five-year, $7 billion agreement to lease advanced AI computing capacity to Tencent after the Trump administration raised concerns that the deal undermined U.S. technology export restrictions, according to the Financial Times and other outlets citing people familiar with the matter.

The contract would have given the Chinese technology giant access to approximately 100,000 advanced Nvidia AI chips installed in data centers across Southeast Asia. Neither Oracle, Tencent nor the White House immediately responded to requests for comment on the reported cancellation.

The deal's collapse matters well beyond one contract. It signals that Washington now treats overseas cloud leasing — not just direct chip sales — as a enforcement target under export controls, and that the roughly $7 billion in question sits at the center of an escalating fight over who gets access to frontier AI compute.

Why did the deal draw scrutiny?

U.S. export controls restrict Chinese companies' access to some of the most advanced AI accelerators, including high-performance processors from Nvidia. Leasing computing capacity outside China offers another route to that hardware, but overseas arrangements can still attract scrutiny from U.S. authorities — and this one did.

U.S. officials reportedly contacted Oracle to express concerns about the agreement. The performance difference between leasing 100,000 top-of-the-line GPUs and owning them is negligible; leasing let Tencent access computing resources it could not obtain in China because of restrictions imposed by both U.S. and Chinese authorities, without building and operating the underlying infrastructure itself.

The pressure came from both directions. Officials in Guangdong province, where Tencent is headquartered, separately urged the company to prioritize domestic Chinese chips, according to the reports.

What does Oracle lose?

For Oracle, abandoning the contract represents a potential loss of substantial business. The company could, however, seek other customers for the computing capacity originally allocated to Tencent.

That capacity is not trivial. Approximately 100,000 advanced Nvidia chips were earmarked for the arrangement across Southeast Asian data centers. Oracle operates data centers in Malaysia and maintains an established cloud relationship with ByteDance, which underscores how much regional infrastructure is exposed to tighter scrutiny.

How wide does the precedent reach?

The issue extends beyond Oracle. Chinese technology companies including ByteDance and Alibaba have also relied on overseas data centers to secure computing resources restricted at home.

That pattern now carries regulatory and commercial risk on both sides. For Chinese AI companies, access to overseas cloud capacity is becoming less certain. For U.S. providers serving those customers, each arrangement invites the kind of intervention that killed the $7 billion Tencent deal.

The bottom line: access to AI computing power is becoming a geopolitical issue as much as a technology and infrastructure challenge. Whether Oracle can redirect capacity originally intended for Tencent — and whether other U.S. cloud providers retreat from similar Chinese contracts — will define the next phase of compute-driven export enforcement.

Source: networkworld.com

Share this article:

More from Marcus Bennett

Marcus Bennett

Show full bio

Senior reporter covering companies, deals and strategy in artificial intelligence and machine learning at AI In Context.

526 articles

Related articles

  1. NVIDIA and AMD Enter January With Growth, Export Risk
  2. Nvidia Trades 18% Down as Washington Adds a Governance Discount
  3. Smuggled Chips and Server Loopholes: How China Gets Nvidia GPUs
  4. Report: Older Nvidia Chips Cost Oracle Its Microsoft Deal
  5. Moonshot AI Sought Nvidia Blackwell Capacity in Europe, Sources Say

« Previous articleNext article »