If MCP is the USB-C of AI agents, A2A is their Ethernet
Tell me, Mr. Smith ... what good is an agent if it's unable to speak?
AI + ML12 Jul 2025 | 3
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.
Intel's leaders have stopped pretending – and it's about time
Comment Not even in the top 10, CEO Lip-Bu Tan reportedly tells employees
CxO10 Jul 2025 | 64
Swiss boffins just trained a 'fully open' LLM on the Alps supercomputer
Source code and weights coming later this summer with an Apache 2.0 bow on top
HPC10 Jul 2025 | 13
IBM boasts new Power11 chips are stingy on power usage
More efficient cores plus an optional energy saver mode in Big Blue's latest CPUs
Systems08 Jul 2025 | 8
Samsung predicts profit slump as its HBM3e apparently continues to underwhelm Nvidia
Analysis Investors advised to brace for massive fall from Q1 to Q2
Systems08 Jul 2025 | 3
CoreWeave's $9B Core Scientific acquisition is a bid for more power
All the GPUs in the world aren't worth much if you don't have a place to put them
On-Prem07 Jul 2025 | 4
Amazon built a massive AI supercluster for Anthropic called Project Rainier – here's what we know so far
deep dive It's almost like AWS is building its own Stargate
On-Prem04 Jul 2025 | 26
Amazon's latest Graviton 4 EC2 instances pack dual 300Gbps NICs
And no, that's not a typo
PaaS + IaaS02 Jul 2025 | 2
With OpenAI, there are no allegiances - just compute at all costs
Analysis Google's TPUs might not be on Altman's menu just yet, but he's never been all that picky about hardware
AI + ML01 Jul 2025 | 3
Oracle just signed one mystery customer that will double its cloud revenue in 2028
Could it be an AI model builder? A Chinese e-tailer? Perhaps a TikTok mass migration
Databases01 Jul 2025 | 5
Google to buy power from fusion energy startup Commonwealth - if they can ever make it work
Someday, my prince will come
Science30 Jun 2025 | 13
How Broadcom is quietly plotting a takeover of the AI infrastructure market
feature When AI is a nesting doll of networks, so why reinvent the wheel when you can license it instead
Datacenter Networking Nexus27 Jun 2025 | 5
Microsoft nuke power deal for Three Mile Island appears to be ahead of schedule
837 megawatt reactor now expected in 2027, energy CEO says
Off-Prem26 Jun 2025 | 5
Rack-scale networks are the new hotness for massive AI training and inference workloads
Analysis Terabytes per second of bandwidth, miles of copper cabling, all crammed into the back of a single rack
Datacenter Networking Nexus25 Jun 2025 | 2
Huawei's latest notebook shows China is still generations behind in chipmaking
Kirin X90 SoC made on two-year-old 7nm N+2 process
Systems23 Jun 2025 | 48
Nvidia bets on Gates-backed nuclear startup to keep its AI ambitions from melting down
$650M funding round aims to bring TerraPower's Natrium power plant in Wyoming online by 2030
Science18 Jun 2025 | 14
Voltron Data throws its weight behind AMD for GPU-accelerated SQL
Exclusive In case you forgot AI isn't the only thing GPUs are good for
Databases18 Jun 2025 | 7
Taiwan thumbs its nose at Beijing by blocking chip exports to SMIC and Huawei
A symbolic political move
Channel17 Jun 2025 | 7
Rack scale is on the rise, but it's not for everyone... yet
Analysis Still buying B200s and MI300Xs? Don't feel bad, Nvidia and AMD's NVL72 and Helios rack systems aren't really for the enterprise anyway
Systems17 Jun 2025 |
Google Cloud flexes as first to host Nvidia RTX PRO 6000 Server VMs
Baby got Blackwell GPUs
PaaS + IaaS12 Jun 2025 | 2
AMD bets on rack-scale compute to boost AI efficiency 20x by 2030
Who'd have thunk? The bigger the iron, the more efficient it gets
European Supercomputing12 Jun 2025 | 6
AMD preps rack-scale Helios systems to contend with Nvidia's Vera Rubin NVL144
House of Zen's biggest iron yet boasts 72 MI400 GPUs, 260 TBps of UALink bandwidth, and 2.9 exaFLOPS of FP4
European Supercomputing12 Jun 2025 |
AMD's MI355X is a 1.4 kW liquid-cooled monster built to battle Nvidia's Blackwell
And the House of Zen wants to put 128 of them in your rack
European Supercomputing12 Jun 2025 | 9
Jupiter, Europe’s most powerful super, takes maiden flight - but misses exascale orbit
ISC Full system expected to break the exaFLOP barrier - perhaps in time for SC25 this fall?
European Supercomputing10 Jun 2025 | 1
Omni-Path is back on the AI and HPC menu in a new challenge to Nvidia's InfiniBand
After a five-year hiatus, Cornelis' interconnect returns at 400Gbps, with Ethernet support next
European Supercomputing09 Jun 2025 |
Trump official warns they're putting the squeeze on CHIPS Act winners
Commerce secretary wants bigger bang for taxpayers' buck
Systems05 Jun 2025 | 16
If it can’t double our money, we’re not building it, Intel Products chief says
New products must show potential for 50% gross margin to get the greenlight
Systems04 Jun 2025 | 28
GlobalFoundries commits $3B more to US fabs in Trump tariff flex
Silicon photonics and gallium nitride a major focus
Systems04 Jun 2025 | 11
Broadcom aims a Tomahawk at Nvidia's AI networking empire with 102.4T photonic switch
Chip giant's latest ASIC promises 200GbE to up to 512 GPUs
Networks04 Jun 2025 |
CoreWeave signs megalease at Applied Digital's not-so-little house on the prairie
A big win for North Dakota
PaaS + IaaS02 Jun 2025 | 2
Musk's smog-belching Colossus datacenter slammed by civil rights group
NAACP claims that 'temporary' gas turbines were an attempt to get around environmental laws
AI + ML02 Jun 2025 | 9
Nvidia scores its first DOE win since 2022 with Doudna supercomputer
System promises a 10x increase in 'scientific output' - not necessarily performance
Systems30 May 2025 | 4
The House of Zen joins the co-packaged optics race with Enosemi buy
Analysis Light is faster than copper, which is important with rack scale architecture
Systems30 May 2025 | 2
Chip designers latest casualties in US-China trade war
Cadence Systems, Synopsys, and Siemens among those subject to export licenses targeting Middle Kingdom
Systems30 May 2025 | 31
Trump trade policy to cost Nvidia $10.5B in lost H20 GPU revenues
But don't worry, sales are still up 69 percent
Systems29 May 2025 | 2
Here’s what it’ll take for Nvidia and other US chipmakers to flog AI chips in China
Jensen be limbo, Jensen be quick, Jensen go under the Uncle Sam’s limbo stick
Systems28 May 2025 | 12
Oracle's $40B Nvidia hardware haul may be too hot for OpenAI's Abilene, Texas DC to handle
400,000 GB200s could be more than a 1.2GW datacenter can chew
Systems27 May 2025 | 21
Turns out using 100% of your AI brain all the time isn’t most efficient way to run a model
Feature Neural net devs are finally getting serious about efficiency
AI + ML25 May 2025 | 22
Nvidia ain't done with x86 as it taps Intel Xeons to babysit GPUs
Computex AI-optimized CPUs promise 4.6GHz clocks, at least for one in eight cores
Systems23 May 2025 | 4
AMD puts Intel in rear view mirror with Threadripper Pro 9000 high-end desktop chips
Computex RDNA 4 makes workstation debut with 32GB R9700
AI Infrastructure Month21 May 2025 | 36
Intel bets you'll stack cheap GPUs to avoid spending top dollar on Nvidia Pros
Computex Behold ‘Project Battlematrix’
AI Infrastructure Month20 May 2025 | 7
Nvidia opens up speedy NVLink interconnect to custom CPUs, ASICs
Computex One of the two just needs to be made by Nv
AI + ML19 May 2025 |
CoreWeave may have built a house of (graphics) cards
Comment An overdependence on hyperscalers and a mountain of debt could pull the rug out
AI Infrastructure Month16 May 2025 | 5
Plan to keep advanced chips from China with tracking tech gains support in Congress
Every shipment you make, every FLOP you generate, Uncle Sam will be watching you
AI Infrastructure Month15 May 2025 | 25
AMD’s first crack at Nvidia hampered by half-baked training software, says TensorWave boss
Interview Bit barn operator to wedge 8,192 liquid-cooled MI325Xs into AI training cluster
AI Infrastructure Month14 May 2025 | 5
US tech titans rejoice in $600B Saudi shopping spree
Prince Mohammed bin Bone Saw will take a few hundred thousand GPUs with his missiles and fighter jets
AI Infrastructure Month14 May 2025 | 46
Trump ends Biden-era dream to cap US AI chip exports
Ding dong, diffusion is dead
AI Infrastructure Month13 May 2025 | 11
AMD is Ryzen to the SMB occasion with a bundle of baby Epycs
The House of Zen’s low-end enterprise strategy is badge engineering at its best
AI Infrastructure Month13 May 2025 | 13
Nip chip smugglers by building trackers into GPUs, US Senator suggests
AI phone home
On-Prem09 May 2025 | 74
Tech titans: Wanna secure US AI leadership? Stop giving the world excuses to buy Chinese
Execs from AMD, Microsoft, and OpenAI tear into profit busting AI diffusion rules
AI Infrastructure Month09 May 2025 | 22
OpenAI drafts Instacart boss as CEO of Apps to lure in the normies
A B2B, API move this ain't, in our view
AI + ML08 May 2025 | 8
Elon Musk’s xAI to pull about half of its smog-belching turbines powering Colossus
Updated Newly completed substation will help bear the load
AI Infrastructure Month08 May 2025 | 41
Trump's trade war with China to cost AMD $1.5B in lost rev
But a multi-billion dollar contract with Oracle for a pile of Instincts and Epycs should take the edge off
AI Infrastructure Month07 May 2025 | 4
FYI: Most AI spending driven by FOMO, not ROI, CEOs tell IBM, LOL
Just 1 in 4 bets paying off so far
AI + ML06 May 2025 | 14
Cerebras CEO actually finds common ground with Nvidia as startup notches IBM win
Feldman calls US's AI Diffusion rules ‘bad policy’
AI Infrastructure Month06 May 2025 | 1
Meta blames Trump tariffs for ballooning AI infra bills
World War Fee Capex could jump by $7B to $72B, Zuckercorp says
AI Infrastructure Month02 May 2025 | 36
Trio arrested in £3M UK bribery probe over Microsoft datacenter build in Netherlands
Blu-3 accused of paying off former Mace Group associates
Legal02 May 2025 | 7
Anthropic calls for tougher GPU export controls as Nvidia's CEO implores Trump to spread the AI love
+Comment This couldn't possibly be about Chinese model builders taking some of the shine off US rivals, could it?
AI Infrastructure Month01 May 2025 | 10
Intel tweaks its 18A process with variants tailored to mass-market chips, big AI brains
Direct Connect If Lip Bu Tan can't sell you his LLM accelerator, he's more than willing to build yours
On-Prem30 Apr 2025 | 1
El Reg's essential guide to deploying LLMs in production
Hands On Running GenAI models is easy. Scaling them to thousands of users, not so much
AI + ML22 Apr 2025 | 11
Biting the hand that feeds IT
