/vnd/media/agency_attachments/bGjnvN2ncYDdhj74yP9p.png)
/vnd/media/media_files/2025/09/26/vnd-banner-2025-09-26-11-20-57.jpg)
/vnd/media/media_files/2025/09/29/intel-gaudi-2025-09-29-11-31-38.png)
As demand for scalable, cost-effective, and efficient AI infrastructure continues to grow, Dell Technologies has introduced the PowerEdge XE7740 server, featuring Intel Gaudi 3 PCIe AI accelerators, which are available from today. Dell is the first to offer an integrated server configuration with the newly released Gaudi 3 PCIe AI accelerator.
The PowerEdge XE7740 is a 4U server designed to support intensive AI workloads through high flexibility, performance, and integration capabilities. It combines the form factor and integration ease of the PowerEdge R-Series with the scalability and performance of the PowerEdge XE-Series. With support for advanced networking and a range of accelerator configurations, it is aimed at meeting the evolving infrastructure requirements of AI workloads.
Infrastructure-focused design for enterprise AI
The XE7740 can be configured with up to eight double-width PCIe AI accelerators. It supports both individual accelerator configurations and bridged modes, offering two groups of 4-way RoCE v2 bridged accelerators. Such bridging is essential for scaling larger AI models and memory spaces, enabling more complex AI workflows.
The server also offers up to a 1:1 accelerator-to-NIC ratio, supporting high-throughput networking via eight full-height PCIe slots and an integrated OCP networking module. Designed to operate efficiently in data centres with typical 10kW rack constraints, the XE7740 supports high-performance AI workloads without requiring significant infrastructure changes.
The XE7740 is engineered for seamless integration into existing air-cooled racks. With Dell Smart Cooling technology and support for up to eight Intel Gaudi 3 PCIe accelerators, the server facilitates the deployment of AI models such as Llama3, Llama4, Deepseek, Phi4, Qwen3, and Falcon3, among others. This compatibility allows enterprises to adopt AI at scale without significant changes to their physical infrastructure.
Enterprise use cases
The PowerEdge XE7740 is well-suited to a wide range of enterprise AI workloads across multiple sectors. It supports Large Language Model (LLM) inferencing and fine-tuning, enabling organisations to adapt foundation models for specific use cases. In the area of multimodal AI, the server is capable of handling complex tasks such as image and speech recognition, which are essential for applications involving natural user interaction or content analysis.
In healthcare, the XE7740 can support data-intensive applications like genome sequencing and diagnostic analysis, where high-throughput processing and model accuracy are critical. The server is also applicable to financial services, where AI models are used for fraud detection and risk modelling to improve security and decision-making. Additionally, in retail and e-commerce, the XE7740 enables real-time personalisation and recommendation systems, helping businesses enhance customer experience through data-driven insights.
Intel Gaudi 3 PCIe accelerators are optimised for widely used platforms such as PyTorch and Hugging Face, enabling efficient deployment and fine-tuning of complex models.
Infrastructure scalability and efficiency
The PowerEdge XE7740 is designed to support long-term scalability and operational efficiency in on-premises environments. It offers a favourable performance-to-cost ratio, making it a viable option for organisations that need to scale AI capabilities while managing budget constraints. Its modular architecture, combined with optional accelerator bridging, allows enterprises to expand their infrastructure gradually in response to evolving AI workloads.
In addition, the XE7740 is engineered for compatibility with standard enterprise data centre environments, operating effectively within typical power and cooling limits. This eliminates the need for costly infrastructure upgrades or retrofitting, ensuring that organisations can scale efficiently without disrupting existing systems.
By deploying infrastructure such as the XE7740, organisations can mitigate variable cloud costs and avoid data egress charges, factors that often limit the feasibility of large-scale AI projects in cloud environments.
Preparing for evolving AI workloads
The PowerEdge XE7740 issaid to not only be equipped to support current AI models but is also built to accommodate future developments in AI technology. It provides the adaptability required for enterprises to refine their infrastructure strategies as new models and use cases emerge.
Venkat Sitaram, Senior Director & Country Head, Infrastructure Solutions Group Specialty Sales, India, Dell Technologies, commented,“With AI adoption accelerating across industries, enterprises need infrastructure that is both powerful and practical. The Dell PowerEdge XE7740 with Intel Gaudi 3 PCIe accelerators delivers the right balance of performance, scalability, and cost efficiency. It integrates seamlessly into existing environments, helping organisations overcome infrastructure barriers. This innovation makes enterprise AI more accessible, enabling businesses to scale confidently and unlock new possibilities.”