Intel D270 GPU AI Accelerator–Up To 480 GB LPDDR5X

TL;DR

Buying for a business?Offer from Amazon

Get business pricing on tech for your team

• Business-only prices and quantity discounts
• Tax-exempt purchasing
• Multiple users, one account, clear invoices

Create a free Business account →Free to register

As an affiliate, we earn on qualifying purchases.Hide

Intel has presented the D270 as an inference-focused data-center GPU with up to 480GB of LPDDR5X memory, 350-watt air cooling, and a standard PCIe 5.0 design. The company says it is built for long-context and high-concurrency AI serving; independent performance results, pricing, availability, and detailed configurations are not provided in the supplied material.

Intel is positioning its D270 data-center GPU as an accelerator for production AI inference, advertising up to 480GB of LPDDR5X memory per card and a 350-watt, air-cooled design. The specifications matter for operators serving long prompts and many concurrent AI sessions, but Intel’s published material does not give a release date, price, or independent performance results.

Intel describes the D270 as an inference-optimized GPU, intended to run deployed AI applications rather than train very large models at scale. Its listed hardware includes 32 Xe3P cores and 256 XMX matrix engines. Intel says the design allocates die area to compute instead of graphics and ray-tracing pipelines, but the supplied information does not include detailed benchmark results demonstrating the effect.

The card is specified with up to 480GB of LPDDR5X, a 350-watt power rating, air cooling, and a standard PCIe 5.0 form factor. Intel argues that the memory capacity can help keep larger models, long prompts, and inference cache resident on a GPU, while standard server integration could allow evaluation in existing data-center environments. The material does not specify which configurations reach the maximum memory capacity or provide bandwidth figures.

Intel lists agentic AI, long-context retrieval-augmented generation, enterprise and coding assistants, document understanding, multimodal workloads, diffusion models, and high-concurrency serving as intended uses. It emphasizes prompt prefill—the processing of input tokens before a model generates a response—as a target workload. Intel also describes support for open models and software frameworks including PyTorch and vLLM; the provided information does not establish the full software support matrix or benchmarked throughput.

Memory Capacity for AI Serving

The D270’s headline capacity addresses a practical constraint in inference: long prompts and concurrent sessions can require substantial memory for model weights and the key-value cache used during generation. More available memory could allow a serving system to accommodate larger contexts or more sessions on a card, depending on the model, precision, software, and memory bandwidth. Intel presents this as a way to improve deployment economics, but those benefits remain company claims until tested against specific workloads.

The design also speaks to data-center operators wary of expanding power and cooling requirements. Intel lists 350 watts and air cooling, and says the card uses standard PCIe infrastructure rather than proprietary networking or a liquid-cooling retrofit. That could simplify evaluation for some existing facilities. It does not, on its own, establish total system cost, performance per watt, or compatibility with every server configuration.

Intel’s Inference-Focused D-Series

Intel’s product page frames the D-series around changes in AI serving: agentic systems may make repeated model calls, retrieve information, use tools, and process long contexts before returning an answer. Those workflows can place demands on memory capacity, prompt processing, and concurrency, not just raw compute. The D270 is presented as a response to those requirements.

Intel says the GPU is intended for enterprise AI, cloud services, GPU-as-a-service providers, and government or sovereign-cloud deployments. Its claim that the card can support whole-context reasoning with fewer GPUs and without high-bandwidth memory is a vendor positioning statement; the supplied page does not provide workload-specific measurements to validate it. Intel’s accompanying performance-projection disclaimer says comparisons with the RTX Pro 6000 rely on public information and Intel internal analysis as of February 2026, and that results vary by use and configuration.

“Intel GPU D270 is an inference-optimized data center GPU designed for the agentic AI era.”

— Intel, in its D-series product information

Specifications Still to Come

The available material leaves several purchasing questions unanswered. It does not state when the D270 will ship, its price, which memory configurations will be sold, or whether the 480GB figure applies to a particular model or configuration. Intel also does not publish memory bandwidth, supported model-size limits, measured tokens per second, or power results for named workloads in the supplied information.

Software compatibility and availability require more detail as well. Intel names PyTorch and vLLM, while the supplied product copy also refers to SGLang and OpenAI Triton in guidance about messaging; that wording is not a complete, confirmed compatibility list. No independent evaluation or customer deployment evidence is included, so claims about cost, throughput, and performance relative to competing accelerators remain unverified here.

Availability and Testing Ahead

Intel directs prospective customers to contact the company for further information and updates. The next useful milestones would be confirmation of pricing, shipment timing, product configurations, and software support, followed by third-party or customer results on representative inference workloads.

For operators considering the D270, comparisons will need to account for the specific model, context length, concurrency, precision, and full server configuration. Until those details and reproducible measurements are public, Intel’s claims describe the product’s intended role rather than a demonstrated advantage over alternatives.

Key Questions

What is the Intel D270?

Intel describes it as an inference-focused data-center GPU for running production AI applications, including long-context and high-concurrency workloads.

How much memory does the D270 have?

Intel lists up to 480GB of LPDDR5X per card. The supplied material does not identify the available configurations or clarify which configuration reaches that maximum.

What power and cooling does Intel specify?

Intel lists a 350-watt power rating and air cooling. It also describes the card as using a standard PCIe 5.0 form factor.

Is the D270 available to buy, and what does it cost?

The supplied product information gives no availability date or price. Intel asks interested customers to contact the company for updates.

Has Intel published independent performance results?

No independent results are included in the supplied material. Intel describes intended benefits and references projections based partly on its internal analysis, but workload-specific performance remains to be established.

Source: hn