AMD’s networking chief: Never bet against Ethernet
Four decades in networking taught Soni Jiandani to back Ethernet, rather than on any single company owning the stack
Analysis - AI & HPC
In-depth analysis of networking, cloud, and AI infrastructure trends
-
IBM Think: 5 thoughts on tokens, AI governance, and quantum
And why the giant thinks all businesses will be neoscalers by decade's end
-
Some neoclouds hate their name. So what should we call them?
A search for consensus from the new infrastructure elite
-
How OpenAI made the network ‘dumb’ for AI
Dissecting MRC with OpenAI’s networking guru
-
6 things we learned from Xcelerated Compute Show London
All eyes on neoclouds, sovereignty and inference economics
-
Nvidia’s networking chief explains what the heck ‘scale-in’ actually means
Gilad Shainer unpacks the fifth pillar of AI infrastructure and why the front-end/back-end split doesn't hold up anymore
-
Purpose-built vs general-purpose: The rise and rise of ‘AI-native’ clouds
With a spot on Gartner’s debut AI infra Magic Quadrant, CoreWeave is vying to be the purpose-built alternative to hyperscalers
-
TurboQuant: Is compression all you need to beat the memory crunch?
Vector quantization, stock dips & RAMageddon – the real story behind how a memory compression algorithm went viral
-
Why KV cache is key to AI memory woes
Exploring key-value cache with Weka, Vast, and Red Hat
-
Cornelis Networks sees supercomputing success as Ethernet vision crystalizes
Intel spinout CEO makes the case for Omni-Path as it toasts Top500 supercomputer spot
-
Nvidia's networking chief: Deep buffer switches are 'a disaster' for AI networking
As scale across networking takes shape, Nvidia's Gilad Shainer says rivals are reaching for the wrong toolbox
-
Beyond HBM: The flash memory technology that could reshape AI infrastructure
Sandisk thinks a NAND-based architecture can smash the memory wall
-
NaaS goes metal: Megaport’s Latitude sees it feed the AI beast with 'staffless' data centers
Forget VMs: Automated bare metal is the only way to satisfy the AI agent hunger, says Megaport CEO
-
Beyond the big three: How neoclouds are reshaping networks for AI
New kids on the block race to become the new kings of AI compute
-
Cerebras spins Nvidia’s Groq tie‑up as proof its wafer‑scale bet was right
Why Nvidia’s GTC 2026 keynote was ‘the best thing ever’ for Cerebras
-
AMD strikes back: The two-front battle with Nvidia for AI software & CPU control
Chip rival looks to fight off encroachment into the CPU space while also breaking CUDA’s lock-in
-
How Cisco is rethinking compute in the network
Move over speeds and feeds; it’s all about programmability, platforms, and ‘velocity’
-
The year that was in networking: Ethernet on the up & Nvidia’s side hustle
Quantum network breakthroughs, outages & the new home of networking news
-
The 5 unlikely winners of the memory chip shortage
From faster networks to smarter controllers, these companies are turning the memory crunch into opportunity
-
Beyond CUDA: Inside the push to loosen Nvidia’s grip on AI computing
AMD, Intel, and startups like Spectral Compute line up to reduce hardware lock-in and snatch chip giant's computing crown
-
How AI is making (and breaking) data storage
Billion-dollar GPUs may dominate the headlines, but a strong sovereign storage foundation will prove the real star in eliminating AI bottlenecks
-
Will memory fail the AI boom?
How the memory crunch is reshaping the AI infrastructure market and pushing costs downstream
-
2025's biggest winners and losers: SDxCentral's year in review (Pt. 1)
The year's biggest wins, most spectacular flops, and everything in between
-
Nvidia’s democratization strategy: How CUDA Tile simplifies GPU programming for AI developers
CUDA Tile simplifies AI programming beyond ‘moats’ to deepen Nvidia's ecosystem advantage
-
Ethernet's exascale victory: HPE’s Slingshot continues to conquer the exascale interconnect battle
Latest Top500 list reveals stark divide between proliferation and performance in supercomputer interconnects