Oracle Reportedly Scraps $7B Tencent AI Chip Lease After U.S. Pressure, Blocking 100,000 Chips
Chinese firms rent overseas capacity for Nvidia chips they cannot buy at home, and Nvidia now reportedly requires cloud partners to sell at least 51% of capacity to non-Chinese customers.
Oracle (ORCL) has reportedly canceled a five-year, roughly $7 billion deal to lease advanced AI chips to Tencent, after U.S. officials raised concerns about it. The agreement would have given the Chinese technology company access to up to 100,000 chips in Oracle data centers across Southeast Asia. Neither company nor the White House immediately commented, and the report has not been confirmed.
What was reported
| Item | Details |
|---|---|
| Deal | Five-year lease of AI computing capacity, estimated at about $7B, described as Tencent's largest overseas lease |
| Chips | Up to about 100,000 advanced AI chips that Tencent cannot buy in China under rules from Washington and Beijing |
| Location | Several Oracle data centers in Southeast Asia |
| Payment | Tencent was to pay about 30% of the value upfront |
| U.S. pressure | Officials raised concerns with Oracle even though overseas chip leasing remains permitted under existing rules |
| Chinese pressure | Officials separately urged Tencent to prioritize domestic chips |
| Comment | Oracle, Tencent and the White House did not immediately comment |
By our calculation, $7 billion over five years is about $1.4 billion a year, and a 30% upfront payment would be about $2.1 billion. Spread across 100,000 chips, the deal is worth roughly $70,000 per chip over the term, or less if Tencent used fewer than the maximum. These are our own arithmetic on the reported figures.
Why Chinese firms lease abroad
Chinese companies lease overseas computing capacity to train AI models on Nvidia (NVDA) chips that U.S. rules restrict from entering China. A separate report says Nvidia now requires its cloud clients to sell at least 51% of their services to non-Chinese customers. Tighter export controls and rising demand have raised overseas AI leasing prices, contract lengths and upfront payments, according to the report, which complicates China's efforts to expand AI computing capacity.
Who else is exposed
| Company | Position |
|---|---|
| ByteDance | One of the two largest Chinese data center customers in Southeast Asia; Oracle supplies it through several Malaysian facilities |
| Alibaba | The other largest Chinese data center customer in the region |
| Tencent | Had signed the lease; officials in its home province of Guangdong have urged it to back chips made in China |
What to watch
- Confirmation. Whether Oracle, Tencent or the White House confirms the cancellation.
- Other leases. Whether U.S. officials press on the ByteDance and Alibaba arrangements.
- Oracle's AI backlog. How the loss of a contract of this size affects Oracle's cloud build-out and funding plans.
- Nvidia. How cloud partners meet the reported 51% requirement.
- Related coverage. See our reports on Meta's $15 billion bond, AMD and the Nasdaq tech selloff.
Sources: This article is based on published reporting that cites unnamed people familiar with the matter, and the companies had not confirmed it when published. Items labeled by our calculation are our own arithmetic. This article is informational only and is not investment advice.
Readers also read
Meta Prices $15B Bond Across Five Maturities With Coupons From 5.85% to 7.50%, About $990M a Year
EarningsAMD Beats as Q3 Revenue Jumps 45% to $13.38B on 97% Data Center Growth and Q4 Guide Tops $17B
ResearchNasdaq's Worst Month Since March 2025: What the Tech Selloff Means and What to Watch Next Week
MarketCatalyst is not a registered investment advisor and does not manage client assets. Content on this platform is provided for informational and educational purposes only. It is not investment advice, and MarketCatalyst is not a stock-picking or trade-alert service. Trading stocks and options involves risk, including the possible loss of principal. Consider your own goals, time horizon, and risk tolerance, and consult a qualified financial advisor before making any investment decision.