Combining Dell’s hardware, software, services and ecosystem with AMD’s advanced CPUs and GPUs, CSPs can build a future-ready foundation to capitalize on AI
Communications service providers (CSPs) operate networks that power the world’s digital data flow. Now that the initial 5G buildout is largely complete, they are shifting their attention toward new operational paradigms designed to manage escalating network complexity while unlocking new revenue streams. Artificial intelligence (AI) is the connective thread—simultaneously driving internal efficiencies and OPEX reduction while powering connectivity-enabled services tailored to the needs of specific consumer and enterprise segments.
MeriTalk research, commissioned by Dell Technologies, underscores the moment: 48% of telecom executives view AI as the most transformative technology shaping the industry, yet 68% say their organizations are struggling to keep pace with rapid changes in technology and customer demand. The existing geographically distributed, purpose-built network infrastructure resists the rapid modernization needed for AI infrastructures to thrive.
A solid approach is to centralize and consolidate AI projects into a common, high-performing, CSP-owned fleet. This strategy enhances data security, control, and compliance while reducing risks associated with fragmented and often organizationally siloed deployments.
Dell Technologies and AMD are redefining the foundation for telecom innovation with advanced AI infrastructures.
The Dell PowerEdge XE7745 exemplifies this approach: a purpose-built, air-cooled 4U server designed for accelerated AI workloads, easily integrated into existing distributed environments. It supports up to 8 double-wide or 16 single-wide PCIe GPUs and up to 2×192 CPU cores, giving CSPs the flexibility to fine-tune core ratios and scale performance as requirements evolve.
CSPs can consolidate with Dell PowerEdge XE9785/XE9785L platforms, featuring 8x AMD Instinct™ MI350 or MI355X accelerators. CSPs can select between air- and liquid-cooled platform configurations, both of which are available with AMD EPYC™ CPUs and AMD Instinct GPUs with Infinity Fabric™, AMD Pensando™ AI-NICs, and Ultra Ethernet readiness for scale-out fabric topologies. These platforms provide:
- Investment protection: Air or direct liquid cooling with 8x GPU high-memory capacity per node (288 GB HBM3E per GPU, up to 8 TB/s bandwidth) for telecom-scale model models and inference efficiency.
- Facility efficiency: XE9785L direct liquid cooling enables higher density and lower cooling energy within existing footprints.
- Confidence: The extensive AMD ROCm™ software ecosystem supports 2M+ AI models out of the box without the need to port code.
- Operational value: Deploy a fleet of virtual operator assistants and intelligent network visibility brokers to automate insights/actions and reduce OPEX.
- Assurance & optimization: Chaos engineering with synthetic data inside a generative network digital replicas to continuously calibrate devices for optimal QoE and forecast failures from network equipment.
Transform with a scalable, durable, CSP-grade AI platform that drives revenue and delights customers with innovative services.
CSPs looking to advance their AI strategies can explore additional resources on how Dell Technologies and AMD are enabling the future of telecom:
