From the archive
AI hardware provides the computing power, memory and networks used to train and run artificial intelligence models. Follow news on GPUs, AI accelerators, data centres, chip design and energy demand. Read how changes in infrastructure affect performance, availability and cost, and distinguish measured results from vendor claims about what the next generation of computing will deliver.
Lead story
AWS has published guidance for governing SageMaker HyperPod when several machine-learning teams share a cluster. Its central point is practical: project membership alone is not a security boundary, so access, compute allocation and operations need separate controls.
Read the report & discussion ↗Watch Desk
The AI industry may need as much as $6 trillion in revenue by 2031 to sustain its global infrastructure build-out. Bain’s projections put revenue from AI products at $1.2 trillion to $1.8 trillion by then, leaving a possible $4.2 trillion gap between the expected sales and the scale…
Watch Desk
Paul Krugman argues that the AI investment boom is already diverting money and resources from other parts of the economy, rather than posing only a future risk. His case points to rising interest rates, data-centre construction and pressure on supplies such as electricity and memory chips.
Watch Desk
South Korea has selected Doosan Robotics to lead two government-backed projects with a combined research budget of about $74 million, spanning AI-enabled collaborative robots and automated welding. The plans put locally processed AI into industrial machines, with commercial robot products targeted for 2031.
Watch Desk
Xeal plans to use spare capacity at its US EV-charging sites to build a distributed network for AI inference, with a target of more than 100,000 Nvidia GPUs. If the plan works, it could bring computing closer to users without waiting for new data centres to connect to…
Watch Desk
AI data centres are becoming a campaign issue in several closely watched US midterm races, where candidates are arguing over who should pay for the electricity and grid upgrades they require. Axios identifies contests in Michigan, Ohio, Colorado and Pennsylvania where those costs, and the politics around new…
Watch Desk
AI investment is booming, but the productivity data cited by Yahoo Finance does not yet show a clear lift in the US economy’s total factor productivity. That measure is slightly below zero, according to Apollo chief economist Torsten Sløk’s analysis, while the more upbeat output-per-hour figure may have…
Watch Desk
Google and Constellation Energy have agreed a 20-year power-purchase deal intended to add 890 megawatts of nuclear capacity in the US. The arrangement links growing data-centre demand with upgrades to existing reactors, rather than relying solely on new power plants.
OpenAI Watch
OpenAI has announced plans for a $20bn data centre campus in Effingham County, Georgia, with a peak electrical load of 3.2 gigawatts. The scale, power agreement and local hurdles make this a major piece of AI infrastructure news, not just another big number in a server-room arms race.
AMD Watch
AI agents may be shifting some of the computing workload back towards CPUs, giving AMD and Intel a larger role alongside the GPUs that have dominated the AI boom. The change matters because agents need processors to keep their virtual computers and tasks running, not just chips to…
NVIDIA Watch
NVIDIA’s head of sustainability argues that growing electricity demand, including demand from AI, strengthens the case for a US nuclear expansion. The essay also sets out the company’s view of where computing could help nuclear developers, and where policy needs to change.
OpenAI Watch
OpenAI and Synopsys are developing a specialised AI model that can operate chip-design software, with agents set to run and refine engineering workflows for human review. It is a substantial step towards putting AI to work inside semiconductor design, though there is no launch date or pricing yet.
TSMC Watch
TSMC and IC-Link by imec are expanding access to advanced chip-design infrastructure through a cloud service for customers developing silicon. The notable twist is support for mixed-partner electronic design automation flows, letting teams use their preferred tools while IC-Link manages the hosted environment.
Watch Desk
Chip-simulation startup Vinci has raised $250 million at a $1.5 billion valuation, with plans to expand its team and increase customer pilots from two to 20. Its software uses AI foundation models to simulate how chip and hardware designs behave before they are built, including the heat challenges…
GMI Cloud Watch
GMI Cloud says it has arranged $668 million in financing to expand its AI infrastructure across the United States, Taiwan and the wider Asia-Pacific region. The important distinction is the funding mix: $223 million in Series B equity and a $445 million credit facility, not a $668 million…
Cohere Watch
Cohere and Thales say a new task under a Canadian naval support contract will bring agentic AI into maintenance and support activities for Royal Canadian Navy vessels. The planned system is aimed at routine work, with personnel remaining in control and able to focus on more complex tasks.
Watch Desk
AI-related goods are driving more than half of export growth in most East Asian and Pacific economies, according to the World Bank. The bank has raised its regional growth forecast, but warns that a reversal in AI spending could remove a key support for growth.
Microsoft AI Watch
Microsoft has added real-time streaming transcription and expanded language support to its MAI speech models. The practical draw is faster, more flexible speech processing, although Microsoft’s documentation says the models are in public preview and not recommended for production workloads.
Google DeepMind Watch
Google DeepMind staffer Paige Bailey demonstrated Gemma 4 running locally in a web browser, with no API call or data sent off the device. The talk also laid out a range of model sizes and on-device tools, putting the practical case for open, smaller AI models on display.
Watch Desk
KoBold Metals uses AI and geological data to search for critical minerals, but the company’s high-profile claims about what that approach has found remain unverified. A Berkeleyside investigation describes a company valued at nearly $3 billion and operating across 14 countries, while asking how much of its promise…
AMD Watch
AMD has released ROCm 10.1, a substantial update to its open-source software ecosystem for AI and high-performance computing. The release focuses on moving data between storage, memory and GPUs, while adding tools and hardware support that could make AMD systems more useful to developers and operators.
