LongCat 2.0 is Meituan's 1.6T-parameter MoE model with a native 1M-token context, trained entirely on domestic AI ASIC ...

Tensordyne says logarithmic computing could reduce AI inference costs and power demands, offering an alternative to conventional chip designs.

For the first time, a research team has demonstrated an artificial intelligence semiconductor technology that integrates the ...

As the Principal Scientist at Abu Dhabi-based QRDC, he has been paving the way for newer advancements in quantum photonics ...

The latest generation of AI models are such competent coders, engineers and (soon) scientists that many worry they may be ...

Microsoft's Copilot+ program for NPU notebooks launched in 2024. Now, GPUs are also allowed, deviating from the initial path.

In an effort to maintain US' technological edge, President Donald Trump has imposed tariffs and, at various points, tightened restrictions on the export of advanced AI chips ...

Researchers have demonstrated a ferroelectric memory chip that performs both random sampling and AI computation, paving the ...

The AI boom doesn't run on GPUs. It runs on a 12-layer tower of memory, and here's why that monopoly is now setting the price ...

Nvidia’s Grace CPU, the company’s bid to challenge Apple Silicon in energy-efficient ARM-based computing, sits roughly two ...

A supercomputer in Shenzhen was declared the world's fastest. It uses only standard microprocessors and not the special-purpose chips called graphics processing units ...

The U.S. ASIC Market was valued at USD 4.19 Billion in 2025 and is expected to reach USD 7.69 Billion by 2035, at a CAGR of 6.30%. Driven by Surging AI Accelerator Demand from Hyperscalers and Rising ...