How AMD, Nvidia, Broadcom Can Ride Oracle’s $455B Cloud Surge
News Summary
Oracle reported an astonishing 359% year-over-year surge in its cloud division's remaining performance obligations (RPO), reaching $455 billion, significantly up from $138 billion in the previous quarter. This revelation propelled Oracle's stock up nearly 36% in a single day. This RPO figure represents guaranteed, contracted revenue, and Oracle anticipates exceeding half a trillion dollars in RPO within months. To fulfill these obligations, Oracle has increased its capital expenditure projection for fiscal 2026 by 65% to $35 billion, primarily for AI infrastructure expansion. As Oracle does not develop its own AI chips, this expansion directly drives demand for GPUs from Nvidia and AMD. Nvidia, a key partner, supplies most of Oracle Cloud Infrastructure's GPUs, offering "bare metal" instances with high-end chips. AMD is expanding its presence with the deployment of MI355X GPUs, providing a cost-effective alternative and enhancing Oracle's negotiation leverage. Furthermore, Broadcom is poised to benefit from the shift towards AI inference workloads, which are expected to increasingly favor custom application-specific integrated circuits (ASICs) for their superior cost and energy efficiency compared to GPUs. Investors' anticipation of Broadcom's ASICs capturing a share of Oracle's AI workload led to a nearly 10% surge in its stock, outperforming Nvidia and AMD.
Background
Oracle is a global leader in enterprise software and cloud services, having actively transitioned into cloud computing to become a major cloud provider. Its Oracle Cloud Infrastructure (OCI) division competes with hyperscalers like Amazon AWS, Microsoft Azure, and Google Cloud. The explosive growth of AI, particularly generative AI, has dramatically increased demand for high-performance computing infrastructure and advanced semiconductors. Nvidia dominates the AI training market with its A100 and H100 GPUs, while AMD and Broadcom are actively vying for market share with their Instinct series GPUs and custom ASIC chips, respectively. Cloud service providers are continuously increasing capital expenditures to expand data centers and procure AI accelerators to meet the growing AI demands of their enterprise clients.
In-Depth AI Insights
Will Oracle's 'bare metal' strategy provide a sustainable competitive advantage in the AI cloud market? - Oracle's offering of "bare metal" Nvidia GPU instances, allowing direct access to physical hardware and eliminating virtualization overhead, is crucial for AI model training and inference requiring extreme performance. This differs from the typically virtualized environments offered by many hyperscalers, potentially attracting specific enterprise clients and AI startups with stringent performance and cost-efficiency demands. - However, this strategy could also increase operational complexity, and over-reliance on third-party chip suppliers (like Nvidia) might limit long-term cost control and customization capabilities. If hyperscale competitors intensify their in-house AI chip development, Oracle might face challenges in custom optimization, though the flexibility and performance advantages remain attractive in the short term. What does Oracle's diversified procurement strategy across multiple chip vendors signify for the AI chip market? - Oracle's collaboration with AMD and Nvidia, and potential pivot to Broadcom's ASICs, indicates that hyperscale cloud customers are actively pursuing supply chain diversification to reduce dependency on a single vendor and enhance negotiation leverage. For Nvidia, this could mean increased pressure on its pricing power and profit margins in the AI GPU market in the long run, despite its undeniable technological leadership in the short term. - For AMD and Broadcom, this signals significant external market opportunities beyond existing hyperscale customers' internal chip projects. This diversification trend will encourage greater innovation and competition, accelerating the evolution of AI chip technology and potentially leading to a more fragmented AI accelerator market with a wider array of specialized players. What are the strategic intentions behind Oracle's massive capital expenditure and its deeper implications for AI infrastructure build-out? - Oracle's 65% increase in fiscal 2026 capital expenditure to $35 billion is not merely about fulfilling existing RPO but represents a strategic move to capture market share in the global AI infrastructure build-out. This suggests Oracle likely anticipates sustained high AI demand and is willing to invest heavily to ensure its cloud services remain competitive in the AI era. - Such aggressive capital spending will further intensify the global scramble for resources like electricity, data center space, and AI chips, potentially leading to further supply chain tightness and cost increases. This could present new investment opportunities for companies providing critical infrastructure components (e.g., energy management, cooling technologies, and advanced packaging) and drive innovation in data center design and operational efficiency.