Under One Massive “AI” Roof, Deploying Vera Rubin, 6th Gen EPYC & MI455X, and Diamond Rapids Among Many Others
Oracle Cloud Infrastructure unveiled its next-gen AI infrastructure that deploys NVIDIA, Intel, AMD and many other accelerators, bringing massive cloud capabilities to support core businesses and frontier-scale AI.
Oracle’s Acceleron Is a High-Performance Fabric Designed To Connect Its Massive Cloud Infrastructure As It Scales Compute Across A Multitude of Accelerators
During Oracle AI World 2026, the company announced that it is building up its next decade of Oracle Cloud Infrastructure by leveraging large-scale deployments of accelerators & hardware, spanning multiple vendors.
Oracle stated that OCI now spans more than 200 cloud services, including large-scale AI. During its Q1 FY2027 earnings, the company delivered an additional 850MW of data center capacity, and looking ahead, the company is going big.
First up, OCI is expanding its partnerships with Intel, NVIDIA, and AMD. On the NVIDIA front, OCI will deploy a multi-gigawatt scale of Vera Rubin NVL72 systems to support growing demand for AI training, inference, and agentic applications, with general availability starting now. NVIDIA Vera CPU servers and NVL8 “HGX” systems will also be deployed by OCI.
OCI Superclusters combine NVIDIA NVL72 rack-scale systems, Oracle Acceleron networking with NVIDIA Spectrum-X Ethernet networking featuring NVIDIA Spectrum-6 switches and NVIDIA BlueField-4 DPUs, and gigawatt-class power and cooling to train and serve the largest AI models. Each rack connects 72 GPUs through NVIDIA NVLink scale-up networking, while OCI’s resilient, multiplanar network scales across hundreds of thousands of GPUs.
AMD’s 6th Gen AMD EPYC CPUs will be deployed to power OCI’s Compute E7 instances. The heavy-weight Helios AI rack will feature the full array of AMD’s enterprise architecture that is going to leverage the latest Instinct MI455X GPUs.
What Do E7 Instances Deliver? OCI Compute E7 instances, powered by 6th Gen AMD EPYC processors, will provide customers with advanced performance and capabilities for agent orchestration, concurrent agent sandboxes, retrieval services and in-memory caches, as well as web, API and enterprise application tiers. E7 Acceleron Standard virtual machines will offer flexible compute and memory configurations.
Do note that both NVIDIA and AMD already power OCI’s infrastructure with deployments ranging from P100, V100, A100, H100, H200, GB200, GB300 on the NVIDIA front, & AMD MI300X, MI355X on the AMD front.
Intel’s chips are also being deployed massively by OCI with the next-gen OCI X14 platform using the upcoming Diamond Rapids “Xeon 7” chips. The chip will boost AI through dedicated general-purpose and AI-specific acceleration in addition to Intel AMX (with FP8, FP16, BF16 support) and Intel APX (Advanced Performance Extensions).
Intel and Oracle today announced plans for a multiyear collaboration intended to expand Intel Xeon-based computing on Oracle Cloud Infrastructure (OCI).
The collaboration is intended to meet the growing demand for enterprise applications, agentic AI, and inference, and is expected to span current and future Xeon platforms, including planned OCI X14 offerings based on Intel’s next-generation server processor codenamed Diamond Rapids.
Besides these, OCI also announced that it’s bringing Google TPUs to the platform. With this, OCI will become the first hyperscaler outside of Google to deploy TPUs, starting in 2027, and ramping to GW-scale over multiple years.
By pairing on-die Inter-Chip Interconnect (ICI) routing with Optical Circuit Switching (OCS), the control plane detects hardware or link degradations and dynamically routes optical paths around faults in real time without job restarts, sustaining over 97% goodput across thousands of chips on large-scale training runs and high availability for production serving. Increased on-chip SRAM, more HBM capacity, and a new low-diameter interconnect topology that cuts network hops by more than 50% allow the latest generation TPUs to deliver up to 80% better performance per dollar for low-latency serving.
And lastly, but just as significant, OCI will leverage Cerebras and SambaNova to drive AI inference faster at the lowest latency. The first Cerebras CS3 deployments are planned for 2027
“AI inference demands choice,” said Mahesh Thiagarajan, president, Oracle Cloud Infrastructure, Oracle. “As agents take on more complex work, customers demand compute that matches each application’s performance needs. Bringing Cerebras’ inference to Oracle CIoud Infrastructure will add another ultrafast and deterministic option to our broad compute portfolio, helping customers build agents that respond quickly and tackle demanding tasks.”
AMD is also playing a big role in powering Oracle’s Acceleron Fabric, which features Pensando data processing units (DPUs) for AMD CPUs and GPUs on OCI. Acceleron brings together dedicated network fabrics, converged NICs, zero-trust routing, and multiple network paths to improve performance, security, and resilience.
We’re now extending Oracle Acceleron with High-Performance Fabric for front-end RDMA connectivity and more NVMe block-storage attachment options across bare metal and virtual machines. Our new Oracle Collective Communications Library helps distributed AI workloads use healthier backend paths. Front-end multiplanar networking adds automatic path selection, failover, and congestion control. Together, these capabilities move data faster, keep workloads running when a path fails, and get more useful work from the compute you’ve deployed.
| Vendor | Key Deployments & Platforms | Scale / Timeline | Focus Areas |
|---|---|---|---|
| NVIDIA | Vera Rubin NVL72 Superclusters; Vera CPU servers; NVL8 HGX; Spectrum-X + BlueField-4 | Multi-gigawatt; GA now | Training, inference, agentic AI |
| AMD | 6th Gen EPYC (E7 instances); Helios racks (MI455X); Pensando DPUs | Ongoing expansion | General compute, AI racks, networking |
| Intel | Diamond Rapids “Xeon 7” (OCI X14); multi-year Xeon platforms | Upcoming X14 | Enterprise, agentic AI, inference |
| TPUs (first non-Google hyperscaler deployment) | Starting 2027; GW-scale | Large-scale training & serving | |
| Cerebras | CS3 systems | First deployments 2027 | Ultra-low-latency inference |
| SambaNova | Inference systems | Expanding | High-speed AI inference |
OCI is assembling one of the most open, multi-vendor AI platforms among hyperscalers—pairing specialized accelerators with its own high-performance Acceleron fabric—so customers can match the right compute to each workload while scaling efficiently from enterprise applications to frontier training and ultra-low-latency inference.
https://wccftech.com/oracle-cloud-infrastructure-brings-nvidia-amd-intel-google-cerebras-sambanova-broadcom-under-one-massive-cloud-roof/amp/?shem=aimgspe,
