Skymizer HTX301: 700B LLMs at 240W on PCIe
Skymizer's HTX301 uses six 28nm chips and 384 GB LPDDR5 to run 700B-parameter LLMs on a single PCIe card at just 240W.
Articles Tagged "Inference"
Meta MTIA 450: 18.4 TB/s Inference Accelerator
Meta's MTIA 450 doubles HBM bandwidth to 18.4 TB/s and adds FlashAttention hardware acceleration for GenAI inference in 2027.
Meta MTIA 400: GenAI Inference at Scale
Meta's second-gen ASIC delivers 6 PFLOPS FP8 and 288 GB HBM for GenAI and recommendation inference inside Meta's data centers.
SubQ Review: 52x Faster, but Show Your Work
Subquadratic's SubQ claims the first linear-scaling LLM with a 12M-token window - but private beta access, self-reported benchmarks, and a 17-point MRCR gap make independent verification the only test that matters.
SubQ
SubQ is the first LLM built on a fully subquadratic attention architecture, achieving a 12M-token research context and 52x faster inference than FlashAttention at 1M tokens.
ZAYA1-8B: Open Reasoning Model Rivals Claude on AMD GPUs
Zyphra's ZAYA1-8B matches Claude 4.5 Sonnet on HMMT 2025 math benchmarks at just 760M active parameters, trained entirely on AMD Instinct MI300X GPUs under Apache 2.0.
SubQ Launches: 12M-Token Context on Sub-Quadratic AI
Subquadratic exits stealth with SubQ, the first frontier model built on a sparse-attention architecture, a $29M seed round, and a 12M-token context window that costs a fraction of Opus.
Nebius Buys Eigen AI for $643M to Own Inference
Nebius agrees to acquire 20-person MIT inference startup Eigen AI for $643M, betting that optimizing every token per Nvidia chip is the real moat in the AI infrastructure race.
Cerebras WSE-3 - The Wafer-Scale AI Engine
The Cerebras WSE-3 is the largest chip ever built - a TSMC 5nm wafer with 900,000 AI cores, 44GB SRAM, and 21 PB/s bandwidth. Now powering a $20B OpenAI deal and Amazon Bedrock deployments.
Google TPU 8i - Low-Latency Inference for Agent Era
Google's TPU 8i is a purpose-built inference chip with 10.1 FP4 PFLOPs, 288GB HBM3e at 8,601 GB/s, and a Boardfly topology that cuts collective latency 5x for agentic AI workloads.
Qualcomm AI250 - Near-Memory Computing for Inference
The Qualcomm AI250 applies near-memory computing to the same 768GB LPDDR5X design as the AI200, promising 10x higher effective memory bandwidth and lower power for LLM inference at rack scale.
Rebellions RebelRack - 64 FP8 PFLOPs at 5 Kilowatts
The Rebellions RebelRack packs 32 Rebel100 chiplet NPUs with 4.5TB HBM3E and 153.6 TB/s aggregate bandwidth into a rack drawing just 5kW - roughly 4x the compute-per-watt of an H100 DGX.
