Tag: inference

Intel Unveils Crescent Island AI Accelerator With Up to 480 GB LPDDR5X Memory

Intel introduces Crescent Island, a high-performance AI accelerator designed for inference tasks, featuring up to 480 GB LPDDR5X memory and 350 W power draw.

Read More

Advances in AI Hardware Could Lower Inference Costs, But Consumer Prices May Remain Stable

New AI processors promise cheaper inference, yet rising infrastructure expenses keep consumer costs largely unchanged for now.

Read More

Harnessing Ambient Noise: Thermodynamic Computing Offers Energy-Efficient AI Processing

Thermodynamic computing proposes using environmental noise as a resource to reduce energy spent on AI model training and inference.

Read More

Alphabet Collaborates with Marvell on New AI Inference Chips

Alphabet and Marvell are negotiating to develop two specialized AI chips aimed at enhancing inference performance and data transfer speeds.

Read More

NVIDIA Introduces Groq 3 LPU: A Shift Toward Deterministic AI Inference

NVIDIA’s Groq 3 LPUs mark a departure from traditional AI accelerators, enabling deterministic inference for next-gen heterogeneous AI platforms.

Read More

Nvidia Develops Tailored Groq AI Chips for Chinese Market Amid Export Controls

Nvidia is preparing a customized version of Groq AI chips for China, addressing U.S. export restrictions while enhancing inference performance.

Read More

Nvidia Unveils Groq 3 LPU Chip to Boost AI Inference Performance

Nvidia introduces the Groq 3 LPU, a new inference accelerator chip designed to enhance token-level processing with high throughput and low latency.

Read More

Nvidia Highlights Significant AI Inference Cost Reduction with Blackwell Architecture

Nvidia reports that the Blackwell AI architecture has cut neural network inference costs by up to 10 times, leveraging both hardware and software advancements.

Read More

Nvidia CEO Clarifies Commitment to OpenAI Investment Amid Rumors

Nvidia’s Jensen Huang affirms ongoing plans to invest in OpenAI despite rumors of lost interest following hardware alternatives.

Read More