Semiconductor News & Analysis Feed

US AI startup Tensordyne claims 3nm Napier chip outperforms NVIDIA Blackwell by 13x in tokens per second The TDN72 rack houses 288 Napier chips, delivers 608 petaflops of FP8 compute, and claims to match the throughput of nine NVIDIA Rubin plus Groq LPX racks. VIEW GALLERY - 3 Hassam Nasir Tech Reporter Published Jun 16, 2026 12:15 AM CDT 1 minute & 45 seconds read time TL;DR: Tensordyne's 3nm N

US-based AI company, Tensordyne, has announced the successful tape-out of its Napier chip, which it claims to demolish NVIDIA's Blackwell & Rubin chips with leading token throughput and efficiency. The Napier chip will be the core component of the Tensordyne Napier TDN system, which is designed in collaboration with Broadcom and HPE Juniper Networks. The Napier platform has one goal: to unify AI

The company said it can consume an order of magnitude less power per token versus GPU alternatives.

Google's push to diversify its Tensor Processing Units (TPUs) supply chain is increasingly reaching into the foundry side, adding pressure on ASIC makers such as MediaTek. Recent reports indicate that Google is not only set to adopt Intel's embedded multi-die interconnect bridge (EMIB) packaging for its next-generation product, but is also planning to bring in Samsung Electronics for front-end waf

Broadcom has acknowledged it will not win all of Google's TPU orders, highlighting how tensor processing units have become a competitive arena for multiple chipmakers. For global readers, the shift signals a broader race for cloud AI supply, capacity, and influence, as more firms vie for a share of high-value custom chip demand.

Cloud-based artificial intelligence infrastructure startup TensorWave Inc. said today it has closed on a bumper $350 million Series B funding round as it strives to meet demand for an alternative to Nvidia Corp.’s graphics processing units. The round was co-led by Magnetar and AMD Ventures, the venture capital arm of the chipmaker Advanced Micro Devices Inc. Maverick Silicon, Nexus Venture Partner

Converting a quantized checkpoint into an NVIDIA TensorRT engine bridges the gap between model optimization and production deployment, enabling faster inference, higher throughput, and more efficient GPU utilization at scale. In a previous post, we produced a high-quality FP8-quantized Contrastive Language-Image Pretraining (CLIP) checkpoint with NVIDIA TensorRT Model Optimizer. This post picks

Tensormesh Inc. has hit upon a way to make artificial intelligence inference more efficient by eliminating the need for redundant computations, and its technology is so convincing that several of AI infrastructure giants are backing it with $20 million in funding. Today’s round saw the participation of Nvidia Corp., Advanced Micro Devices Inc. and CoreWeave Inc., as well as the venture capital fi

__fail__