Tuesday, November 17, 2026
AI 인프라 · 뉴스 & 분석
홈 › 컴퓨트·클라우드 › 리포트
컴퓨트·클라우드 · 리포트

Tencent는 Oracle을 통해 100,000개의 GPU를 확보했으며, 수출 규제로 인해 중국 본토에서 사용할 수 없는 GPU 컴퓨팅 용량에 접근하게 되었다.

중국 하이퍼스케일러가 미국 클라우드 인프라를 이용해 AI 칩 수출 제한을 우회하고 있으며, 글로벌 컴퓨팅 시장의 구조적 분열과 지속적인 칩 접근 마찰을 나타냅니다.
업계 전문지Slicast · 2026년 10월 1일 14:28 UTC · 글로벌 · 출처: Data Center Dynamics
중요도 77

Tencent has signed a contract to access 100,000 GPUs via Oracle's cloud platform under a five-year lease spanning multiple Oracle data centers in Southeast Asia. The lease will enable Tencent to access Oracle's GPUs in regions where the hardware is unavailable in China due to US export controls, positioning the company to advance its AI models and agentic tools. According to sources familiar with the matter, the deal is valued at approximately $7 billion, with Tencent paying 30 percent upfront.

The timing of the agreement reflects Tencent's significant AI infrastructure investment cycle. The company reported a negative cash flow of RMB 13.8 billion ($2.06 billion) for its second quarter, marking its first negative figure in over a decade. CFO John Lo attributed this to "large AI infrastructure capex and AI-related prepayments," noting that excluding those prepayments, free cash flow would have totaled RMB 37.6 billion ($5.61 billion)—a difference of roughly $7 to $8 billion that aligns with the Oracle deal's scale.

Tencent president Martin Lau articulated the company's confidence in the capital commitment: "We're comfortable in making significant investments in AI because not only is there a substantial upside potential, there is also clear downside protection. The AI investments we're making are mostly in AI infrastructure, and in the worst case, which we do not believe would happen, we can choose to rent that infrastructure out at cost recovery or even better prices via Tencent Cloud if needed."

The immediate priority for the leased compute is training bigger and better Hunyuan models, followed by inferencing workloads and later deployment as bare metal rental or through Model-as-a-Service offerings. This deployment strategy aligns with Tencent's previously stated approach of prioritizing internal AI development. CSO James Mitchell explained in May 2026: "Looking through the rest of this year, as the supply of China-designed GPUs progressively ramps up, we'll be remedying that situation; we will be making more capacity available in Tencent Cloud, and consequently driving up Tencent Cloud's rate of expansion. That's where the trade-off has been made: we have been consciously late to monetize the AI opportunity through Tencent Cloud because we've been simultaneously supporting a number of AI initiatives internally."

원문 보기
Tencent는 Oracle을 통해 100,000개의 GPU를 확보했으며, 수출… · Slicast