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
AAEON has announced the MIX-PTLWV1, a Mini-ITX motherboard built around Intel Core Ultra Series 3 processors and aimed at edge-AI work such as machine vision, robotic automation and industrial inspection. The important bit is not another processor family name in a press release. It is the attempt to…
Read the report & discussion ↗Watch Desk
Positron AI has raised $875 million to develop Asimov, an inference chip designed to tackle one of AI’s least glamorous but most expensive problems: moving enough memory quickly enough to keep models fed. The startup says its approach could make large-scale AI serving more efficient, but the headline…
Anthropic Watch
Anthropic has signed its first Australian data-centre lease, securing planned capacity of 2.16 gigawatts near Brisbane. Reuters says the facility is expected to begin coming online in 2027 and will handle inference, the rapid processing used to serve AI responses, rather than model training.
Cohere Watch
Cohere and Germany’s Aleph Alpha have signed a definitive merger agreement that would combine Cohere’s models with Aleph Alpha’s compliance and orchestration platform on Schwarz Group’s STACKIT cloud. The proposed Toronto-Berlin business is a serious attempt to give regulated buyers an AI stack outside US cloud jurisdiction, not…
Watch Desk
An AI system has reportedly designed, verified and deployed a small inference-chip design from a written specification, with Architect Labs saying the flow took under two weeks. The practical significance is not that a chatbot has magically become a semiconductor foundry, but that hardware and software optimisation may…
TSMC Watch
Micron has confirmed that TSMC will manufacture the base dies for its next-generation high-bandwidth memory, according to the supplied BigGo Finance report. That makes TSMC part of the HBM4-era supply chain for all three major memory makers: Micron, SK Hynix and Samsung.
Watch Desk
Nscale, the London-based cloud provider that exists to house and run AI models, has filed to go public in the United States. The filing pairs one of the steepest growth rates the AI boom has produced with a loss that dwarfs it, and it takes both numbers to…
Watch Desk
Nscale has filed to go public. The Nvidia-backed AI cloud company, all of two years old, disclosed on Friday that revenue in the first half of 2026 ran ten times higher than a year earlier, while losses mounted alongside billions of spending on data centres and the AI…
Watch Desk
Ramin Hasani has put a number on the intelligence Mercedes-Benz plans to build into its cars: about 600 megabytes, small enough to reach the whole fleet as an ordinary over-the-air update. The figure came up on the Latent Space podcast, and its significance is less benchmark than business…
TSMC Watch
TSMC's hold on AI chip packaging looks set to run through 2028. TrendForce's latest industry study projects that CoWoS-L, the advanced packaging behind Nvidia and AMD's accelerators, remains the mainstream technology for AI chips, with Intel's EMIB-T closing in on cost but not on maturity. The decisive factors…
ASML Watch
ASML is planning a new generation of chipmaking machines with larger masks, built so the industry can print processors as big as today's largest AI datacentre chips. A pilot line is pencilled in for 2031 and high-volume production for 2033, which turns the physical ceiling on AI silicon…
Zhipu AI / Z.ai GLM Watch
Zhipu's chief scientist Tang Jie says the lab has reached a "minimal closed loop" in recursive self-improvement: an agent driven by its GLM-5.3 model helped build and optimise the inference service that serves GLM-5.3-Flash. If the numbers hold, a model is now doing engineering work on the system…
Zhipu AI / Z.ai GLM Watch
Zhipu AI has launched GLM-5.3-FlashX on its BigModel platform, with the API live and a headline claim of up to 200 tokens per second of output. The pitch is refreshingly narrow: the new model keeps the intelligence and pricing of GLM-5.3-Flash and simply generates faster, which is precisely…
NVIDIA Watch
Huawei has pulled forward the launch of its next generation of AI chips, with the Ascend 960DT now planned for early 2027, roughly three quarters earlier than previously expected, and the 960PR following in the third quarter of that year. Nvidia's shares gained about 2.5% to $219.24 on…
Watch Desk
China's AI models may be winning users by the million, but they are earning a tenth of what OpenAI and Anthropic make combined, according to estimates the Rhodium Group published on Thursday. The sharper point sits in the valuations: the research firm thinks investors are paying far more…
Watch Desk
Crusoe has confirmed the size of one of the AI boom's biggest cheques: a $3.9 billion Series F that values the data centre developer at $30.9 billion, barely ten months after it raised $1.38 billion at $10 billion. The valuation has tripled while the paint dried. The more…
AMD Watch
AMD fell more than 4% on 14 September after executives at Anthropic, OpenAI and xAI publicly warned about the pace of AI development, then rebounded to roughly $549 by 17 September, up more than 7% on the day. The panic lasted about a day. Wall Street's price targets…
AWS AI Watch
Amazon has joined the AI safety argument at last, and it has picked a careful spot to stand in. The company says models should be released only when they are safe and fully tested, but it has stopped short of backing the slowdown call that swept up its…
Watch Desk
The AI industry has quietly changed what it does for a living: less building models, more running them. That shift from training to inference is now forcing a rethink of the chips and memory everything else depends on, and it is redrawing who buys silicon from whom. The…
