Amazon Expands Nvidia GPU Fleet by 2 Million Despite Proprietary AI Chip Efforts

Yahoo Finance ·

Amazon is significantly expanding its hardware pipeline with Nvidia, even as the tech giant continues to advance its proprietary artificial intelligence chips. Amazon Web Services announced plans to deploy two million additional Nvidia GPUs between 2027 and 2028, featuring architectures such as Blackwell Ultra, Rubin, and Rubin Ultra. This massive procurement builds on a prior commitment to install one million Nvidia chips starting in 2026. The substantial rollout highlights Amazon's ongoing reliance on Nvidia for advanced infrastructure, despite its years of in-house processor development through Annapurna Labs. Furthermore, the two organizations are broadening their alliance to encompass dedicated federal government artificial intelligence facilities using 100,000 graphics processing units, alongside integrating Nvidia Nemotron models into Amazon Bedrock and SageMaker. Concurrently, Amazon Robotics will utilize Nvidia technology for warehouse automation. While proprietary chip initiatives remain active, including collaboration on NVLink Fusion for Trainium processors, Nvidia clearly remains an indispensable pillar of Amazon's ecosystem.

AI 시장 분석

Despite developing its own AI chips, Amazon has decided to additionally adopt 2 million NVIDIA GPUs between 2027 and 2028. This is a massive investment totaling 3 million NVIDIA hardware units built into its infrastructure, including 1 million previously announced. With this expanded partnership, both companies have reaffirmed a close alliance, broadening their scope to include building government AI factories and robotics collaboration. From an investor perspective, this suggests that NVIDIA's market dominance remains solid, while highlighting Amazon's accelerated infrastructure expansion.

상승 영향

DYAX 전담 분석

Amazon's massive GPU adoption announcement serves as a direct positive catalyst for NVIDIA's (NVDA) hardware revenue growth, increasing visibility for demand on next-generation chips such as Blackwell Ultra and Rubin systems. It proves a high dependency on the NVIDIA ecosystem despite parallel custom chip development.

The bullish scenario is that cloud demand surges, leading to continuous AI infrastructure investment and a sharp increase in financial performance for both companies, while the bearish scenario is that custom chip (Trainium) performance improves faster than expected, reducing NVIDIA dependency in the long term. Going forward, AWS AI revenue growth rates and next-generation chip shipment indicators must be monitored closely.

AI가 생성한 분석으로 투자 자문이 아닙니다.

DYAX Investor Sentiment

Bullish (Long) 36% · Bearish (Short) 64%

354 participants

Related News

원문 보기 — Yahoo Finance