New instances with Oracle Acceleron provide powerful platform for cloud-native applications and AI inference

Oracle today announced a new variation of its Ampere® -powered A4 instance, featuring Oracle’s Acceleron technology. The compute shape, A4.Standard.Ax, delivers enhanced infrastructure capabilities specifically designed for most demanding cloud and AI workloads.

The introduction of A4.Standard.Ax represents the next phase of Ampere and OCI collaboration, enhancing the infrastructure surrounding the platform while maintaining the efficiency and predictable performance required to scale modern cloud environments.

As organizations scale cloud infrastructure to support modern applications and advanced AI services, the demand for efficient, high-performance and cost-effective compute platforms is paramount. This expansion of the A4 platform delivers enhanced infrastructure capabilities specifically designed for today's most demanding cloud and AI workloads to meet this demand.

Oracle’s Acceleron technology intelligently offloads and accelerates key infrastructure services. This, in concert with Ampere’s flagship AmpereOne® M, provides a powerful combination: AI-ready, resilient, scalable and efficient for distributed applications, data platforms, and a wide array of AI inference workloads in OCI. Organizations can leverage these capabilities often at a significantly lower cost compared to alternative compute types, particularly for AI inference workloads, and especially where consistent, responsive performance at scale is valuable.

Designed for Diverse AI and Modern Cloud Workloads

A4 instances have quickly become a powerful platform for cloud-native applications and AI inference. Organizations like Oracle Red Bull Racing and Uber rely on Ampere-powered infrastructure in OCI to run demanding services that require consistent performance and efficient scaling.

A4.Standard.Ax instances excel across a broad spectrum of AI inference use cases, including:

• Agentic AI/OpenClaw: Enabling agent-driven workflow automation and decision making
• Generative AI: Supporting Small Language Models (SLMs) and Large Language Models (LLMs) up to 15B active parameters
• Computer Vision: Image classification, object detection, and facial recognition
• Machine Learning (ML): Anomaly detection and time-series forecasting
• Recommender Engines: Powering real-time recommendation systems atural Language Processing: Voice transcription, content summarization, and language translation

These instances are particularly adept at allowing higher model density than GPUs for certain AI inference tasks, making them ideal for running enterprise models and agents. The inherent characteristics of AmpereOne M, combined with the underlying efficiency brought by Acceleron, ensure that even as these workloads scale, performance remains consistent and predictable.

“Customers running modern cloud applications need infrastructure that can move data faster, operate securely, and scale efficiently,” said Kiran Edara, VP Compute at Oracle Cloud Infrastructure. “Acceleron-enabled A4.Standard.Ax instances powered by AmpereOne M bring Oracle Acceleron SmartNIC to OCI Compute, delivering higher throughput, strong isolation, line-rate encryption, no-downtime network patching, and faster NVMe-based storage access for distributed applications and AI-driven workloads in OCI.”

“AmpereOne M was designed to support the scalable, efficient infrastructure modern cloud environments require,” said Sean Varley, VP and Chief Evangelist at Ampere. “By combining AmpereOne M with Acceleron in OCI’s A4.Standard.Ax instances, customers gain a powerful platform for running distributed applications and AI inference workloads with performance, efficiency and value at scale.”

Acceleron: Bolstering the A4 Platform with Accelerated Infrastructure

The A4 platform, already powered by AmpereOne M, is significantly enhanced by Oracle Acceleron, a dedicated infrastructure acceleration technology. Acceleron transparently offloads crucial infrastructure tasks from the main CPU, including networking, storage processing, data movement and foundational security functions. By handling these services in specialized components, Acceleron ensures the AmpereOne M CPU remains entirely focused on application workloads and executing AI inference tasks, thereby enhancing overall system efficiency.

This strategic offloading yields several key benefits for modern cloud and AI workloads. It improves overall system performance by accelerating data movement and network communication. It also enhances security by enabling clearer separation between infrastructure services and application workloads, which can reduce the attack surface and support stronger isolation between system layers. Crucially, Acceleron delivers predictable latency for application-critical operations, a vital characteristic for demanding AI inference workloads like real-time recommender engines, interactive NLP services and multi-modal generative AI, where consistent responsiveness is key for optimal user experience and operational efficiency.

Combined with the scalable compute provided by AmpereOne M processors, Acceleron strengthens the infrastructure foundation supporting A4 workloads and represents an important step in the continued evolution of the A4 platform in OCI.

Experience A4.A Capabilities and Value

To illustrate the capabilities of A4.Standard.Ax instances, a demo video is available showcasing various AI inference blueprints and production-grade applications running on this powerful platform. From agents to object classification and multi-modal Vision+LLM applications, these demos highlight the real-world value of the A4 platform for a number of AI workloads.

Prospective customers can explore these capabilities further at the Oracle AI Customer Excellence Center. Here, visitors can engage with live demonstrations, gain deeper insights into the instance architecture, and initiate Proof-of-Concepts to validate how A4.Standard.Ax can transform their specific workloads, particularly those where consistent, responsive performance at scale provides a competitive advantage.

Beyond raw performance, the A4 portfolio offers an exceptional value proposition for cloud and AI workloads. For a wide range of AI inference tasks, including SLMs and general machine learning, A4 and A4.Standard.Ax deliver efficiency and highly consistent responsiveness, often at a fraction of the cost of alternative compute types. This capability empowers organizations to significantly optimize their cloud spend while enhancing the quality and consistency of their modern cloud and AI services.

Continuing the Momentum of Ampere in OCI

Ampere and Oracle have worked closely to expand the availability of Arm-based cloud compute within OCI. Ampere-based instances have been widely adopted by a number of organizations seeking efficient and scalable infrastructure for modern workloads. OCI will continue expanding the A4 platform with A4.Standard.Ax instances powered by AmpereOne M and Acceleron, providing an enhanced foundation for organizations running modern cloud and AI workloads at scale with unparalleled efficiency, value, and strong, responsive performance.

Related: Announcing OCI Compute A4 Standard Acceleron Instances for AI and Cloud-Native Workloads