Urgent.News

What's breaking now, across thousands of outlets.

AI

US government pressure… Oracle cancels $73 billion AI chip rental agreement with Tencent

Translated from Korean Read in Korean

The US government's pressure led to the cancellation of a $7.4 billion (around 9.4 trillion won) artificial intelligence (AI) chip leasing agreement between US cloud giant Oracle and Chinese tech giant Tencent. According to a Financial Times source, Oracle terminated the five-year AI chip leasing deal with Tencent last Thursday.

The agreement allowed Tencent to utilize up to 100,000 advanced AI chips in various data centers across Southeast Asia. This was one of the largest foreign operation leasing agreements signed by a Chinese company. Tencent, a tech company operating popular domestic messaging platforms like WeChat, has been heavily investing in developing its own AI models.

The cancellation of the deal is believed to be due to simultaneous pressure from both the US and Chinese governments. US government officials reportedly contacted Oracle directly to express concerns over the Tencnet deal. Officials in China's Guangdong province, where Tencent's headquarters is located, are said to have urged the company to expand the use of domestic chips.

Chinese tech giants have been entering into large leasing agreements with Southeast Asian data centers to secure the necessary computational capacity for training high-performance AI models due to US export controls limiting the import of advanced chips from US companies like Nvidia. The US-China tech control extends beyond semiconductor exports to cover AI operating services provided abroad.

While such AI operating leasing agreements are allowed under US regulations, US companies are increasingly facing pressure to reduce their dependence on Chinese customers. Nvidia has recently also reportedly requested cloud customers to sell at least 51% of their services to non-Chinese companies.

Written by urgent.news from Hankyoreh's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at hani.co.kr →

More in AI

Agentic AI vs Generative AI Examples: Cost Verdict

These agentic ai vs generative ai examples show that one-shot generative tasks are the cheapest way to produce a given amount of text, while an agent loop re-reads its context every step and can cost…

  • Generative AI models charge fixed rate per token for input and output.
  • Agentic AI models incur higher costs due to context re-reading and repeated processing.
  • Mixed pipelines combining both layers can reduce overall costs significantly.

More from Thursday 5 November →