Google’s Ironwood TPUs represent a bigger threat than Nvidia would have you believe
Chocolate Factory's homegrown silicon boasts Blackwell-level perf at massive scale
AI + ML06 Nov 2025 | 4
Tobias is The Register's datacenter systems editor. His focus is primarily on the hardware and system software powering today's networks and compute resources. Tobias has previously worked as an editor covering SD-WAN and network security, and as a daily news reporter.
Perplexity shows how to run monster AI models more efficiently on aging GPUs, AWS networks
Some clever networking hacks open the door
AI + ML05 Nov 2025 | 2
Google imagines out of this world AI - running on orbital datacenters
Chocolate Factory's latest moonshot aims to put AI supercomputing cluster in sun-sychronous orbit
Off-Prem04 Nov 2025 | 26
NextSilicon Maverick-2 promises to blow away the HPC market Nvidia left behind
The one chip startup building accelerators for something other than AI boasts performance up 10x that of modern GPUs using a fraction the power
Supercomputing Month22 Oct 2025 | 2
Dropping Nvidia for Amazon's custom chips helped gene therapy startup Metagenomi cut AI bill 56%
It's not the size of your accelerator, it's how you use it
AI + ML22 Oct 2025 | 8
Nexperia drama intensifies as Dutch chipmaker denies ousted CEO's claims of Chinese split
The government in the Netherlands has taken control of the company
Systems20 Oct 2025 | 11
Nvidia still needs Taiwan even as TSMC ramps Blackwell production in Arizona
AI arms dealer relies on Taiwanese advanced packaging plants for top-specced GPUs
AI + ML20 Oct 2025 |
Meta convinces Blue Owl to cut $30B check for its Hyperion AI super cluster
Deal keeps debt off Social Media empire's books
On-Prem17 Oct 2025 | 23
Amazon spills plan to nuke Washington...with X-Energy mini-reactors
Now they just need to get regulatory approval
Networks17 Oct 2025 | 39
TSMC hurrying to bring advanced chip tech to Arizona fab
CEO C.C. Wei cites strong demand for AI products. Intel may also be a factor
Systems16 Oct 2025 | 3
Decomposed dinosaurs make Texas a top destination for AI bit barns
Datacenter hopefuls looking to cash in on AI craze are setting up shop in Lone Star State in search of cheap power
AI + ML16 Oct 2025 | 27
Meta sends Arm a friend request asking for help with Nvidia’s Grace CPUs
No custom Arm CPUs to speak of yet
Systems16 Oct 2025 |
Apple goes all in on AI acceleration with M5 MacBook, iPad, and Vision Pros
Oh and the CPU is up to 15% faster for those that could care less about articifically intelligent Apple products and more about getting work done
Personal Tech15 Oct 2025 | 52
18 zettaFLOPS of new AI compute coming online from Oracle late next year
New clusters to feature 800,000 Nvidia Blackwell and 50,000 AMD Instinct MI450X GPUs
Systems14 Oct 2025 | 7
DGX Spark, Nvidia’s tiniest supercomputer, tackles large models at solid speeds
hands on This relatively affordable AI workstation isn’t about going fast; it’s about doing everything well enough
AI + ML14 Oct 2025 | 12
Nvidia's GB10 workstations arrive with 1 petaFLOPS of compute, 128GB of VRAM, and a $3K+ price tag
Systems from Nvidia, Dell, and others available starting Oct. 15
Systems13 Oct 2025 | 13
Inside the belly of the beast: A technical walk through Intel's 18A production facility at Fab52
deep dive Now if Lip Bu Tan can just find a willing customer
On-Prem12 Oct 2025 | 24
Amazon's Quick Suite is like agentic AI training wheels for enterprises
Slow down there Andy; you wouldn't want to bump into any hallucinations
AI + ML09 Oct 2025 | 1
Intel's open source future in question as exec says he's done carrying the competition
Kevork Kechichian says x86 giant's contributions should benefit Intel first
Software09 Oct 2025 | 58
Panther Lake sets stage for Intel's 2 nm comeback, but many details still TBD
Notebook chip promises 8 to 16 cores and up to 180 TOPS of total AI performance when it hits shelves in January
Systems09 Oct 2025 | 7
CoreWeave bets on serverless agent builder to woo penny-pinching enterprises
Because what enterprises really love are vague consumption-based pricing models
AI + ML08 Oct 2025 | 2
Cisco’s new router unites disparate datacenters into AI training behemoths
With enough routers, Switchzilla says it can link bit barns 1,000 km apart and scale fabrics beyond 3 exabits per second
Networks08 Oct 2025 | 2
China is building a thriving semi industry off US leftovers, export controls be damned
House Committee on China wants more comprehensive ban on chipmaking equipment exports to Middle Kingdom
Public Sector07 Oct 2025 | 52
Cerebras CEO insists dinner-plate-sized chip startup will still go public
Inference service launched a month before IPO filing turns out to have been a much bigger business than initially thought
AI + ML06 Oct 2025 | 6
Stargate is nowhere near big enough to make OpenAI's tie-ups with AMD and Nvidia work
Comment Since revealing Stargate in January, Altman and friends have brought about 200 MW online - they'll need at least 16 GW to claim their red and green prize
AI + ML06 Oct 2025 | 19
Bezos plan for solar powered datacenters is out of this world… literally
Aspiring Bond villain believes the best place to train our AI overlords is in orbit
Offbeat03 Oct 2025 | 63
Microsoft CTO says he wants to swap most AMD and Nvidia GPUs for homemade chips
Pivot will hinge on success of next-gen Maia accelerator
Systems02 Oct 2025 | 26
Trump’s tariff‑shaped stick can’t beat reality on US chip fabbing
Comment The proposed 1:1 chip rule means nothing but pain for US tech until he’s out of office
Systems28 Sep 2025 | 67
Intel reportedly wants TSMC's help to end its reliance on ...TSMC
Chipzilla seeks investment from its top fab frenemy
On-Prem26 Sep 2025 | 3
X2 Elite is Qualcomm’s latest attempt to bring Apple’s M-series magic to the PC
Plus the Snapdragon 8 Elite turns 5
Systems25 Sep 2025 | 8
Intel reportedly courting ex-flame Apple to become its next investor
Chipzilla can't say it's changed much, but could be a handy backup to TSMC
Systems24 Sep 2025 | 9
Tree-hugging hippie datacenter runs entirely on green hydrogen and wastes zero water
Lambda's latest innovation with bit barn builder ECL only supports a handful of Nvidia racks, but it's a start
Systems24 Sep 2025 | 34
Zuck has the power! Meta applies to sell excess electricity
With new electricity sources for AI datacenters, the company will have some juice left over
Off-Prem19 Sep 2025 | 29
Sorry, but DeepSeek didn’t really train its flagship model for $294,000
Training costs detailed in R1 training report don't include 2.79 million GPU hours that laid its foundation
AI + ML19 Sep 2025 | 26
Intel and Nvidia sitting in a tree, NVLink-I-N-G
But still no hero customer for Chipzilla's Foundry biz
Systems18 Sep 2025 | 7
Nvidia GeForced out of China as Beijing demands tech titans embrace homegrown silicon
Huawei or another, we're gonna getcha off Nvidia
AI + ML18 Sep 2025 | 12
AMD tries to catch CUDA with performance-boosting ROCm 7 software
House of Zen promises 3.5x improvement in inference and 3x uplift in training perf over last-gen software
Systems17 Sep 2025 | 5
SK Hynix cranks up the HBM4 assembly line to prep for next-gen GPUs
Top AI chipmakers count on faster, denser, more efficient memory to boost training
AI + ML12 Sep 2025 | 1
Intel talent bleed continues as Xeon chip architect heads for the escape hatch
Ronak Singhal will be moving onto better and brighter opportunities at the end of the month
Systems11 Sep 2025 | 16
Arm wrestles away 25% share of server market thanks to Nvidia's home-grown CPUs
Still far short of the 50% market share Arm infra chief was hoping for
Systems11 Sep 2025 |
OpenAI reportedly on the hook for $300B Oracle Cloud bill
Tick tock Sam, just fifteen months before your first bill is due
Off-Prem11 Sep 2025 | 33
Cadence invites you to play with Nvidia’s biggest iron in its datacenter tycoon sim
Using GPUs to design better bit barns for GPUs? It’s the circle of AI
AI + ML10 Sep 2025 |
Nvidia's context-optimized Rubin CPX GPUs were inevitable
Analysis Why strap pricey, power-hungry HBM to a job that doesn't benefit from the bandwidth?
Systems10 Sep 2025 | 2
Intel shuffles executive deckchairs, tosses 30-year veteran chief overboard
Michelle Johnston Holthaus' tenure as Intel Products CEO lasted just ten months
Systems08 Sep 2025 | 11
Chip designer SiFive aims to cram more RISC-V cores into AI chips
Why reinvent the CPU wheel when you can spend your time engineering a way out of your dependence on Nvidia?
AI + ML08 Sep 2025 | 1
AI chip startup d-Matrix aspires to rack scale with JetStream I/O cards
Who needs HBM when you can juggle SRAM speed and LPDDR bulk across racks
AI + ML08 Sep 2025 | 1
If Broadcom is helping OpenAI build AI chips, here's what they might look like
Analysis Whatever happened to that Baltra thing Tan and crew were helping Apple cook up?
AI + ML05 Sep 2025 | 3
Boffins detail new method to make neural nets forget private and copyrighted info
Because nobody's going to spend billions to retrain a model built on dubiously legal content
AI + ML04 Sep 2025 | 21
No chips for you! Senator wants Americans to get first dibs on GPUs, restrict sales to others
+Comment We’ve got hungry American datacenters to feed, argued the lawmaker - a revival Nvidia dubs ‘doomer science fiction’
Public Sector04 Sep 2025 | 41
Absolutely fabless: Trump derails TSMC's China chip-building effort
World's largest foundry operator joins Samsung and SK Hynix, which recently lost their validated end-user status
Systems03 Sep 2025 | 16
Uncle Sam doesn't want Samsung, SK Hynix making memories in China
End of verified end user status means South Korean memory vendors will need licenses to bring restricted chipmaking tech into Chinese fabs
Systems29 Aug 2025 | 36
Alibaba looks to end reliance on Nvidia for AI inference
Chinese cloud provider reportedly joins the homegrown silicon party
Systems29 Aug 2025 | 7
AI arms dealer Nvidia laments the many billions lost to US-China trade war
China would be a $50 billion a year market for Nvidia if Uncle Sam would let us sell competitive products, says Jensen Huang
Systems27 Aug 2025 | 7
Nvidia details its itty bitty GB10 superchip for local AI development
Hot Chips Starting at $2,999, tiny doesn't mean cheap
AI + ML27 Aug 2025 | 3
Intel pitches Clearwater Forest as a consolidation play for all you hoarding ancient Xeons
Hot Chips Chipzilla's first datacenter part to use 18A process tech is another core-packed monster
Systems27 Aug 2025 | 5
Azure apparatchik shows custom silicon keeping everything locked down
Hot Chips From hardware security chips and trusted execution pipelines to open source Root of Trust modules
Cloud Infrastructure Month26 Aug 2025 | 7
Trump made Intel an offer it couldn't refuse
Comment 10 percent equity and maybe that $8.9 billion in CHIPS funds you're waiting on doesn't get lost in red tape
Systems25 Aug 2025 | 43
Tinker with LLMs in the privacy of your own home using Llama.cpp
Hands on Everything you need to know to build, run, serve, optimize and quantize models on your PC
AI + ML24 Aug 2025 | 7
Google games numbers to make AI look less thirsty
Datacenters' drinking habit exaggerated, claims report comparing apples to oranges
Cloud Infrastructure Month22 Aug 2025 | 10
DeepSeek's new V3.1 release points to potent new Chinese chips coming soon
Point release retuned with new FP8 datatype for better compatibility with homegrown silicon
AI + ML22 Aug 2025 | 4
Biting the hand that feeds IT
