In a significant move poised to redefine enterprise artificial intelligence, Oracle and NVIDIA have announced an expanded strategic alliance. This collaboration, unveiled during Oracle AI World, focuses on delivering more accessible, potent, and practical AI services for businesses. The partnership encompasses cutting-edge hardware innovations and deeply integrated software solutions, positioning AI at the operational core of corporate data infrastructures.
Ian Buck, NVIDIA's VP of Hyperscale and High-Performance Computing, highlighted the essence of this initiative, stating that this renewed partnership with Oracle is carving new paths in accelerated computing. The goal is to streamline database AI pipelines, expedite data processing, empower diverse enterprise use cases, and facilitate easier deployment and scaling of AI inference on the Oracle Cloud Infrastructure (OCI).
A centerpiece of this announcement is the introduction of the OCI Zettascale10 computing cluster. This formidable platform, driven by NVIDIA GPUs, is meticulously engineered to handle demanding AI training and inference workloads that would overwhelm conventional server systems. The OCI Zettascale10 boasts an impressive peak AI compute performance of 16 zettaflops, underpinned by NVIDIA's Spectrum-X Ethernet. This specialized networking fabric prevents GPUs from idly awaiting data, enabling organizations to efficiently scale to accommodate millions of processors.
Beyond sheer computational strength, the true impact of this collaboration lies in its software integrations. These integrations are designed to infuse AI capabilities into every facet of a business’s operations. Mahesh Thiagarajan, Executive VP of Oracle Cloud Infrastructure, underscored this by noting that OCI Zettascale10 offers multi-gigawatt capacity for the most complex AI tasks, leveraging NVIDIA's next-generation GPU platform. Furthermore, the native integration of NVIDIA AI Enterprise on OCI equips joint customers with a premier AI toolkit, readily available alongside OCI's extensive suite of over 200 cloud services, fostering a broad spectrum of customer innovation.
A cornerstone of this innovative strategy is the Oracle AI Database 26ai. Traditionally, the approach has been to transport data to AI models. However, Oracle is pioneering an inversion of this paradigm, advocating for the more secure and efficient method of bringing AI directly to the data. This latest database iteration embodies the 'AI for Data' philosophy, making AI integration intuitive and effective.
Juan Loaiza, Executive VP of Oracle Database Technologies, emphasized that by co-designing AI and data, Oracle AI Database makes 'AI for Data' straightforward to learn and use. This empowers customers to effortlessly generate dependable AI insights, foster innovation, and enhance productivity across all their data, whether residing in operational systems or analytical data lakes. A key feature is the capability to execute agentic AI workflows within the database itself. These AI agents can address intricate queries by intelligently combining a company’s private, sensitive data with public information, all without requiring the private data to leave its secure environment. This is facilitated by advanced functionalities such as Unified Hybrid Vector Search, which allows AI to contextualize information across various data types, including relational tables, JSON files, and spatial maps.
Oracle is also prioritizing long-term security with the new database, which incorporates NIST-approved quantum-resistant algorithms for data both in transit and at rest. This measure acts as a safeguard against 'harvest now, decrypt later' attacks, where malicious actors seize encrypted data today with the intent of decrypting it using future quantum computing capabilities. Holger Mueller, VP and Principal Analyst at Constellation Research, remarked that effective AI demands high-quality data, and Oracle AI Database 26ai delivers both. It serves as a centralized, secure repository for business data, making it the optimal environment for applying AI without data relocation. To simplify and accelerate AI adoption, AI Database 26ai includes sophisticated new AI features extending beyond AI Vector Search. A notable highlight is Oracle's embedding of agentic AI into the database, enabling customers to construct, deploy, and manage their own in-database AI agents through a no-code visual platform equipped with pre-built agents.
The new database is engineered for seamless interoperability with NVIDIA’s comprehensive toolset. Its programming interfaces can now directly integrate with NVIDIA NeMo Retriever, a suite of microservices designed to manage the complex underlying architecture of modern enterprise AI. This integration significantly simplifies the implementation of techniques such as retrieval-augmented generation (RAG). In essence, RAG allows a language model to access pertinent information from company documents before formulating an answer, thereby enhancing the accuracy and utility of its responses. Furthermore, the Oracle Private AI Services Container is slated for a GPU-powered upgrade. This container enables businesses to run AI models within their own secure ecosystems. Soon, it will be capable of offloading the resource-intensive task of creating vector embeddings—a critical component for AI search—to powerful NVIDIA GPUs using the cuVS library. This enhancement is expected to dramatically reduce the time required to prepare data for AI applications.
The partnership extends beyond database enhancements to streamline the entire AI pipeline, aiming to make enterprise AI more accessible. The Oracle AI Data Platform now offers a built-in NVIDIA GPU option and the NVIDIA RAPIDS Accelerator for Apache Spark. This is a game-changer for data scientists and engineers, allowing them to accelerate data processing and machine learning workflows with GPUs, often without altering their existing code. All these tools and capabilities are being consolidated within the Oracle AI Hub, providing organizations with a centralized platform to build, deploy, and manage their AI solutions. From this hub, users can deploy NVIDIA’s NIM microservices—pre-packaged AI functionalities—via a user-friendly, no-code interface. To further reduce barriers to entry, the full NVIDIA AI Enterprise software suite is now natively available within the OCI Console. This means developers can launch a GPU instance and enable all necessary NVIDIA tools with minimal clicks, bypassing separate procurement processes. This seemingly minor adjustment significantly expedites team onboarding and project initiation.
This collaboration underscores a concerted effort to address the practical challenges businesses encounter when integrating AI. By consolidating hardware, data, and software tools into a unified ecosystem, Oracle and NVIDIA are strongly asserting that the era of practical, secure, and scalable enterprise AI has definitively arrived.
