Taalas Launches Hardcore Chip With ‘Insane’ AI Inference Performance

Taalas has launched an AI accelerator that puts the entire AI model into silicon, delivering 1-2 orders of magnitude greater ...
Morning Overview on MSN

Taalas swaps GPUs for hardwired AI chips at blazing 17,000 tokens per sec

Taalas, a Finnish AI company, has reportedly moved away from NVIDIA GPUs in favor of hardwired AI chips, claiming inference speeds of 17,000 tokens per second. The shift coincides with a broader ...
Forbes

The Inference Economy: How Sparse Computing And Model Optimization Are Reshaping Enterprise AI Deployment

The AI industry stands at an inflection point. While the previous era pursued larger models—GPT-3's 175 billion parameters to PaLM's 540 billion—focus has shifted toward efficiency and economic ...
The Next Platform

Taalas Etches AI Models Onto Transistors To Rocket Boost Inference

Adding big blocks of SRAM to collections of AI tensor engines, or better still, a waferscale collection of such engines, turbocharges AI inference, as has ...
Semiconductor Engineering

The On-Device LLM Revolution

Users running a quantized 7B model on a laptop expect 40+ tokens per second. A 30B MoE model on a high-end mobile device ...
CRN

Intel Inks ‘Multiyear’ AI Inference Deal With SambaNova After Acquisition Talks End

Intel plans to tap into its ‘enterprise, cloud and partner channels’ for a new ‘multiyear strategic collaboration’ it has ...

AI Inference Costs Shift Enterprise Spending as Training Expenses Escalate, Stanford Finds

Stanford's 2025 AI Index shows inference costs reshaping enterprise AI budgets as training expenses climb and returns remain limited.

Researchers baked 3x inference speedups directly into LLM weights — without speculative decoding

Researchers from the University of Maryland, Lawrence Livermore, Columbia and TogetherAI have developed a training technique that triples LLM inference speed without auxiliary models or infrastructure ...
VentureBeat

New transformer architecture can make language models faster and resource-efficient

Large language models like ChatGPT and Llama-2 are notorious for their extensive memory and computational demands, making them costly to run. Trimming even a small fraction of their size can lead to ...
ProVideo Coalition

AWS Elemental Inference: AI-powered vertical video in real time

AWS Elemental Inference enables video customers to adapt video content into vertical formats optimized for mobile and social platforms in real time. Today’s viewers consume content differently than ...