AI Chips news and analysis on Web Pulse: 6912 curated articles tracking the latest AI Chips developments, tools, and research, updated continuously from vetted sources.
AI safety arguments that treat algorithmic progress as an exogenous factor with a fixed 3-10x annual rate are flawed because algorithmic progress is driven by the AI ecosystem itself, according to an analysis of AI slowd…
At the AI Infra Summit in Santa Clara on Sept. 16, 2026, Google distinguished engineer Dave Patterson said the push to make AI agents complete tasks autonomously is increasing demand for CPUs, while Nvidia vice president…
A prefill/decode disaggregation setup running DeepSeek-V4-Flash — a 284B total / 13B active model with 256 routed experts — bridged a 700,630-token cold prompt end to end in 11 minutes 26 seconds over plain 10 gigabit Et…
A Level1Techs forum user posting as boatman reported that USB RDMA on Macs is "just good software" and not real RDMA hardware offload, after being unable to obtain ib_read_lat, ib_write_lat or ib_send_lat latency numbers…
Qualcomm is redesigning its next Snapdragon flagship system-on-chip around High Bandwidth Compute (HBC), a near-memory computing architecture intended to address the "memory wall" that limits generative AI decode perform…
A new arXiv paper (2609.16161v1) presents an end-to-end integer-only quantization approach and a sparse dictionary-based KV cache compression strategy that enables LLM inference on Flash compute-in-memory devices, reduci…
The Trump administration has few effective options to slow China's rise as an AI superpower, according to Bloomberg, as it weighs export controls against Beijing's retaliatory leverage over rare earths. Anthropic PBC CEO…
Meta Platforms is preparing to deploy its third-generation in-house AI chip, the MTIA 450 (code-named Arke), in its data centers starting in the first half of 2027, after 12 units delivered by Taiwan Semiconductor Manufa…
DFlash, a block-diffusion speculative decoding drafter, delivered up to 5.02× single-request throughput on Qwen3.5-27B when run on AMD Instinct MI355X through vLLM on ROCm, according to the benchmark. DFlash beat Qwen3.5…
The cost of completing a given task with a machine learning model is falling sharply even though per-token prices for frontier models are not consistently declining, according to an analysis citing Epoch.AI data. GPU pow…
Nvidia CEO Jensen Huang said at Dreamforce on September 15, 2026 that choosing between AI innovation and safety is a false choice, arguing engineering solutions, market incentives, and independent evaluators should serve…
Cornelis Networks introduced its Active Compute Fabric (ACF) at the AI Infra Summit on Monday, an open architecture for scale-up and scale-out networking that integrates programmable compute into the fabric, as rivals ra…
Prefill, not decode, has become the dominant bottleneck in RAG and multi-agent workloads, according to an analysis citing NVIDIA's published figures of roughly 30x higher served-request counts for large MoE models on GB2…
The U.S. Commerce Department ordered prediction market Kalshi last month to take down its AI compute forward curves, citing national security concerns, according to a Semafor report published Tuesday. Kalshi launched the…
Cerebras Systems is hiring a Senior Frontend Engineer for its Developer Console in Sunnyvale, CA, a hybrid role owning platform areas including billing, usage tracking, quotas, request logs, and metrics built on Next.js,…
Elon Musk revealed Monday on the All-In podcast that he is living in an Airstream trailer in Memphis beside SpaceXAI's Colossus supercomputer complex, which has scaled to roughly 1 GW of AI compute capacity by mid-2026. …
AI agents are driving a resurgence in CPU demand as central processing units take on orchestration, tool calls and sandboxing workloads, according to a Forbes Innovation column by contributor Sandy Arnold. The shift carr…
NVIDIA CEO Jensen Huang said at the All-In Summit that annual infrastructure planning cycles at hyperscalers including Amazon, Microsoft, Google, and Meta are "almost always wrong" in the AI era, and predicted a wave of …
A developer argues that more efficient AI chips alone will not resolve the technology's energy problem, pointing to the broader electricity, water, and material demands behind chip production and operation. The analysis,…
MediaTek announced the launch of the Dimensity 9600 Pro, its flagship smartphone system-on-chip and the first chip announced on the 2 nm process node. The company said the SoC combines flagship performance and power effi…