Infrastructure & Hardware
Infrastructure and hardware comprise the physical compute systems, specialized chips, and data center components that power artificial intelligence workloads. Truncated covers chip architectures, supply chains, energy requirements, and the hardware deployments that enable large scale computing.
Qualcomm Demonstrates HBC Memory With 18X Better Bandwidth
Here's how much RAM cost in 2020 vs. 2026
Engadget — RAM is wildly expensive and will likely stay that way for a while. Comparing prices from just a few years ago shows how dire the situation is.
DCD Studio: Building at scale, with Michael Nudelman, IREN
datacenterdynamics.com — Seb Moss speaks with Michael Nudelman of IREN about vertically integrated AI cloud infrastructure, securing 5 gigawatts of power, scaling challenges, and strategic acquisitions and partnerships.
‘Bragawatts’ and why all data-center investing isn’t created equal
semafor.com — EQT, backed by Sweden’s most powerful business family, believes in the AI buildout’s fundamentals, but has stayed away from private credit
Does your model need a Rolls Royce or a bicycle?
The Stack — Nutanix CTO: “When you pay for infrastructure, it is much more predictable than paying per token”
Nebul deploys Nvidia GPUs at Datacenter United campus in Belgium
datacenterdynamics.com — European AI cloud leasing capacity in Machelen
While Americans turn against data centers, Brazil offers tax breaks and cheap power, representing a potential ‘release valve’ for U.S. constraints
Fortune — The government is cutting taxes on data center equipment as Google, Amazon and Microsoft expand in a country already hungry for AI services.
AWS Capacity Largely Sold Out Through 2027, With On Demand Headroom Reserved
InfoQ — Amazon CEO Andy Jassy says AWS’s large committed cloud capacity is largely sold out through 2027, with part of 2028 already allocated, while the company continues to reserve headroom for On Demand, EC2 and Bedrock. Practitioners discussing the remarks report difficulty obtaining On Demand capacity for ordinary EC2 instances, and similar constraints at other providers.
Elon Musk says he is taking over an old Spam factory and a gun range to expand his Colossus 2 AI complex
Business Insider — Elon Musk named the new AI facilities at the Colossus 2 campus "Spam" and "Gun Range."
National Grid to site biggest substation on Teesside, near planned AI datacentre
the Guardian — New electricity generating capacity would enable north-east to attract investment for major datacentre projects
Iole Inc. Sells Entire 151 BTC at a Loss to Fund AI Data Centers
tftc.io — Iole Inc. voted to liquidate all 151.21 BTC for ~$11.7M, absorbing a ¥254M loss to fund AI data centers, while keeping its Hyperliquid position intact.
Modders create automated tool to unlock AI performance on NVIDIA GeForce RTX 20, 30 and 40 series GPUs
Nvidia quietly gave RTX 40-series owners a new frame generation mode, and I measured it on my 4070 Ti Super
XDA — Does the frame generation "performance" mode live up to its promises?
Samsung Warns AI Memory Shortage Will Worsen Through 2027
PYMNTS.com — Samsung expects the demand for memory to “far exceed” supply in 2027, the company said in a Thursday (Oct. 29) press release announcing its third quarter
The fastest delta weight syncs in the world
Baseten — 770GB of weights previously took ~36 seconds to sync on average; now they take 15, with a sub-second request pause, 7.5x faster than previous SOTA.
Bringing the MONAI Model Zoo to AMD Instinct™ GPUs: Validated Inference on ROCm
ROCm Blogs — Learn how to run validated inference for five MONAI Model Zoo bundles on AMD Instinct™ GPUs using a ROCm overlay.
A ‘hello world’ milestone for Project Suncatcher 🚀 After launching our prototype satellite a few weeks ago, we’ve now…
PowerBank plans 2MW modular data center in Ontario, Canada
datacenterdynamics.com — Signs first lease with AI cloud partner
Netflix Treats GPU Reliability as a Distributed Systems Problem
InfoQ — Netflix has detailed how it is building a reliability platform for large-scale GPU training and inference, using layered health checks, automated remediation, and fault-tolerant training to keep jobs running despite hardware failures.
Dismiss email signup
Subscribe to 
The top AI news in your email delivered each morning. Choose categories