Why is Oracle doing so well?

On September 10th, 2025, Oracle's (NYSE: ORCL) stock price surged by nearly 40% in just one day, adding close to $240 billion to its valuation, after the company announced its remaining performance obligations (RPOs — a financial measure of total contracted revenue yet to be recognized for all performance obligations such as cloud subscriptions, support, and services). But what truly drove this spike?

What caused the spike?    

Based on Oracle's Backlog Summary Report, the surge is largely attributed to multi-billion-dollar AI/compute contracts, including a $300 billion deal with OpenAI(as reported by media sources) for computing power over the next five years, along with a continued agreement with TikTok/Project Texas and other hyperscale-style wins such as Microsoft selecting OCI(Oracle Cloud Infrastructure) for Bing conversational search. The true catalysts that drove up the company's valuation were its RPOs, worth $455 billion with $7.2 billion in cloud revenue.

Why is Oracle so significant today?

Oracle's AI Infrastructure offers GPU superclusters that scale up to 131,072 Nvidia Blackwell GPUs with high-bandwidth processing for model training and inference. This technological cohesion between Oracle and Nvidia enables faster time-to-train, lower cost-per-token, and simpler “PO-to-first-epoch” — time taken from contract signing to when customers can begin AI training — timelines for AI teams.

Oracle's AI supercluster comprises the following vital systems:

• Nvidia's GB200 Grace-Blackwell architecture GPUs - The GB200 is a multi-node, liquid-cooled, rack-scale system for the most compute-intensive workloads. 
• Bare Metal Servers - OCI provides clients with single-tenant physical servers dedicated to computation and the management of storage resources.
• Low Latency, High Bandwidth Cluster Fabric - Oracle uses fabrics from InfiniBand and RoCE for extremely high throughput.
• Nvidia's AI Enterprise Software Stack - Oracle integrates Nvidia DGX Cloud on OCI and allows its users to directly interact with Nvidia NIM microservices directly in the OCI console to optimize inference.
• AI Storage Options - The supercluster offers a choice of HPC(High Performance Computing) storage organized to feed high volumes of data to GPUs at lightning-fast speeds.

What lies ahead?     

We live in a world where every corporation, be it a multi-billion-dollar conglomerate or a local trucking company, wants to apply some form of AI to its workflow, even if it adds no significantly noticeable value. This AI boom isn't a passing trend; It's reshaping industries at their core. In this era, with deep partnerships and multi-billion-dollar contracts with tech giants and governmental bodies, Oracle isn't just trying to join the Trillion-Dollar club—it is building the computational backbone for the next generation of innovation. 

With their Zettascale supercluster, Oracle is a one-of-a-kind cloud infrastructure company. As AI adoption accelerates globally, Oracle’s role as the engine powering this transformation cements its place at the very center of the digital future.