Inference-optimized data center GPU for agentic AI, long contexts and high-concurrency serving.
Intel GPU D-series product page
Overview
What is Intel GPU D-series?
Memory-rich AI inference without HBM.
The Intel GPU D-series (D270) is an inference-optimized data center GPU built for agentic AI, long contexts and high-concurrency serving at practical power and cost.
Each card offers up to 480 GB of LPDDR5X memory for memory-rich AI inference without HBM. 32 Xe3P cores and 256 XMX matrix engines accelerate prompt processing (prefill) for more throughput per watt. The design targets production inference workloads: agentic AI, long-context RAG, enterprise AI assistants, coding assistants, document understanding, multimodal AI and high-concurrency serving. A standard PCIe 5.0 form factor with 350W air cooling fits existing data center racks with no liquid-cooling retrofit, and the platform is open by design with zero-day support for leading open models on frameworks such as vLLM.
Intel publishes no pricing on the product page; interested teams can sign up to learn more through Intel's contact channels.
Platforms and languages
Intel GPU D-series Availability
Platforms
Languages
Capabilities
Intel GPU D-series Key Features
480 GB LPDDR5X memory per card
Memory-rich AI inference without HBM, supporting larger models, longer contexts and more concurrent sessions.
Accelerated prompt processing
32 Xe3P cores and 256 XMX matrix engines accelerate prefill for more throughput per watt.
Inference-optimized for agentic AI
Tuned for agentic AI, long-context RAG and high-concurrency serving at practical power and cost.
Standard data center form factor
PCIe 5.0 form factor with 350W air cooling deploys in existing racks with no liquid-cooling retrofit.
Open by design
Zero-day support for leading open models, served on open frameworks such as vLLM.
Best for
Who uses Intel GPU D-series?
AI infrastructure architects, platform engineers and enterprise AI decision makers
Deploy production-grade AI inference for agentic AI and long-context workloads at practical power and cost.
Common questions
Intel GPU D-series FAQs
What is the Intel GPU D270?
An inference-optimized data center GPU designed for the agentic AI era, aimed at enterprise AI, SaaS, GPUaaS and sovereign cloud with large memory and high efficiency.
Which AI workloads does it support?
Production inference such as agentic AI, long-context RAG, enterprise AI assistants, coding assistants, document understanding, multimodal AI and high-concurrency serving.
Why does it matter for agentic AI?
Agents turn one task into many model calls, tool uses and retrievals; the D270 optimizes prompt processing, context size, memory capacity, concurrency and efficiency for that pattern.
What does inference-optimized mean?
The GPU is designed for running models in production rather than large-scale training: prompt processing, high-concurrency serving, memory capacity, efficient power and system-level performance.
How does long context help?
Long contexts need more memory for bigger prompts, retrieval-heavy flows and growing KV caches; the D270's large memory supports bigger models, longer contexts and more concurrent sessions.
Reviews
See what the community thinks and share your experience.
—
Based on 0 ratings
Rating distribution
Community reviews
No reviews
No written reviews yet.
