Nvidia to Supply 1 Million AI Chips to Amazon Through 2027
Multi-year deal with AWS extends beyond GPUs to include networking gear, signaling sustained AI infrastructure demand
Nvidia Corp. has agreed to supply Amazon Web Services with one million artificial intelligence chips by 2027, in a multi-year deal that underscores the explosive demand for AI infrastructure and validates Chief Executive Jensen Huang's projection of a $1 trillion market opportunity.
The agreement, which expands a 15-year partnership between the two companies, encompasses not only graphics processing units but also Nvidia's Spectrum networking chips and ConnectX systems, according to detailed analysis of the deal. Shipments are scheduled to begin this year and continue through 2027, although financial terms were not disclosed.
The deal marks a strategic shift for Amazon's cloud division, which has traditionally developed its own networking equipment rather than relying on external suppliers. AWS's decision to integrate Nvidia's Spectrum-X Ethernet platform and ConnectX SuperNICs into its data centers reflects the growing complexity of AI workloads, where optimized networking performance is increasingly critical.
Nvidia's Spectrum-X platform is designed to enhance network performance for AI tasks by 1.6 times, while the ConnectX systems provide high-bandwidth connectivity between GPU servers. This integrated approach addresses what analysts identify as a growing bottleneck in large-scale AI deployments: the need for seamless communication between thousands of chips processing complex models simultaneously.
"This collaboration showcases the increasing demand from enterprises for optimized and integrated solutions," according to industry analysis. "As AI models become increasingly complex and distributed, networking can become a significant bottleneck, and Nvidia's vertically integrated stack offers optimized performance."
The announcement comes as Nvidia maintains a market capitalization of $4.42 trillion, with analysts maintaining an overwhelmingly positive outlook on the stock. Of 61 analysts covering the company, 58 rate it as a buy or strong buy, with a consensus price target of $267.54—roughly 50% above the current share price of $178.56.
Nvidia reported quarterly revenue growth of 73.2% and earnings growth of 95.6% in its most recent period, reflecting the surge in AI-related spending across industries. The company's profit margin stands at 55.6%, while its operating margin has reached 65%, underscoring the pricing power Nvidia commands in the AI chip market.
For Amazon, the deal represents a balancing act between developing its own custom silicon—the Trainium and Inferentia chips—and deploying Nvidia's proven technology for the most demanding AI workloads. According to AWS's official announcement, the collaboration aims to "accelerate AI from pilot to production" for enterprise customers.
The agreement also provides insight into the competitive dynamics of the cloud infrastructure market. While AWS, Microsoft Azure, and Google Cloud are all developing their own AI chips to reduce reliance on general-purpose GPUs, Nvidia's continued dominance suggests that its integrated hardware and software stack remains difficult to displace for the most advanced applications.
Analysts noted that the deal reinforces Nvidia's position as the leading provider of full-stack AI infrastructure, combining compute, networking, and software into a unified system. This vertical integration has become increasingly important as enterprises seek to simplify AI deployments and reduce the complexity of managing disparate hardware components.
The announcement follows recent developments in the semiconductor sector, including ongoing U.S. enforcement actions against companies accused of illegally exporting AI chips to China, highlighting the geopolitical tensions surrounding advanced semiconductor technology.
For Nvidia, the AWS deal represents another step in its evolution from a gaming-focused GPU manufacturer to a comprehensive AI infrastructure provider. The company's networking business, which generated $11 billion in revenue in its most recent fiscal year, is increasingly seen as a strategic asset that complements its dominant position in AI accelerators.
Looking ahead, analysts will be watching for similar announcements from other major cloud providers, as well as indications of how the massive scale of these deals might affect Nvidia's production capacity and pricing power. The company has signaled that it is ramping up manufacturing to meet demand, though supply constraints remain a challenge across the semiconductor industry.
The deal with AWS also provides validation for Nvidia's Blackwell and Rubin chip platforms, which Huang has identified as key drivers of the company's projected $1 trillion revenue opportunity. With cloud providers continuing to invest heavily in AI infrastructure, Nvidia appears poised to maintain its dominance despite intensifying competition from chipmakers including AMD and Intel.