DIGITIMES ICT news

Home Tech ICT

Aug 24

How Nvidia's whitelist was beaten: 74 B300 servers, Taiwan to China

Taiwan's Keelung District Prosecutors Office has indicted nine people over the illegal export of high-end AI servers built on Nvidia's B300 GPUs. The indictment lays out, step by step, how a compliance regime designed to track every unit was defeated from the inside.

Nvidia's strikes US$6B licensing deal with Poolside to accelerate Nemotron models

Aug 24

Nvidia reportedly weighs Perplexity investment as AI strategy expands beyond chips

Aug 24

Notebook brands shift production back to China as Southeast Asia costs rise

Aug 24

Interview: Inside Foxconn's playbook for the sovereign AI data center race

Aug 24

MORE NEWS

ICT MANUFACTURINGKaynes Technology and BOSGAME sign MoU to expand India market presence

Aug 25, 10:55

Kaynes Technology India and BOSGAME have signed a strategic memorandum of understanding to support BOSGAME's entry and expansion in India. The deal highlights how global hardware brands are increasingly looking to India for design, manufacturing, and customer support as the country's electronics ecosystem matures.

EMS WATCHFoxconn scales AI server production in US and Mexico, keeps Asia expansion diversified

Aug 25, 10:47

AI server demand and the continued regionalisation of global supply chains are keeping Foxconn's capital spending on an upward trajectory in 2026, with the company maintaining its target for full-year capex growth of more than 30%.

AIChina-based Guangyu Xinchen says 3D-stacked edge AI chip has entered mass-production deployment

Aug 25, 10:39

According to 36Kr, Shanghai Guangyu Xinchen Technology says its TC1000 series of 3D-stacked near-memory-computing AI processors has moved from silicon validation to mass-production deployment, marking a claimed acceleration in the commercialization of memory-centric computing for edge AI.

AICommentary: Nvidia's 15% hike piles pressure on the AI supply chain, lifts ASIC prospects

Aug 25, 10:25

Nvidia has reportedly notified major customers that server systems using its AI chips will rise by more than 15%, with the new pricing set to apply to models shipping from early 2027, including Grace Blackwell and the next-generation Vera Rubin platform. The move is expected to ripple through the chip and system supply chain, while industry figures say it could also lift the appeal of inference ASICs.

AIOpenAI narrows gap as Anthropic's Fable 5 loses enterprise momentum

Aug 25, 10:03

OpenAI has been gaining ground on Anthropic in enterprise AI spending, while Anthropic's flagship Claude Fable 5 has delivered less market traction than expected. The shift comes as Ramp analyzed usage across more than 70,000 US companies through its bill pay and corporate card products, according to TechCrunch and the Financial Times.

SERVER, IPC, CLOUD COMPUTING, IOTNovosense encoder launch highlights China's push into robot joint sensing

Aug 25, 09:58

Chinese analog and mixed-signal chipmaker Novosense Microelectronics has introduced the NSM350x series, a Hall-based magnetic encoder designed for high-precision angular measurement in robot joints and industrial motion-control systems, reports Sina Finance.

SERVER, IPC, CLOUD COMPUTING, IOTTaipower pushes 'information flow' strategy to ease northern Taiwan's AI data center power crunch

Aug 25, 09:18

As artificial intelligence chip power consumption continues to rise, server racks are moving toward the megawatt (MW) scale. Amid the "electrical power is computing power" trend, the importance of power electronics is increasing, while the overall power system also faces significant upgrade requirements.

SERVER, IPC, CLOUD COMPUTING, IOTNvidia Groq 3 LPX enters full production for faster agentic AI inference

Aug 25, 09:04

Nvidia said its Groq 3 LPX inference accelerator is now in full production, a move that could speed up agentic AI systems used worldwide for coding, reasoning, and other tasks. The company said the platform is designed to improve responsiveness, lower latency, and support growing demand for real-time AI applications.

SERVER, IPC, CLOUD COMPUTING, IOTSpaceXAI to use Nvidia Vera CPUs for faster agentic AI at scale

Aug 25, 09:02

SpaceXAI plans to deploy Nvidia's new Vera CPUs to speed up agentic AI workloads, a move that could influence how advanced AI systems are built and run worldwide. The partnership also extends toward orbital computing, signaling a broader shift in where future AI infrastructure may be located.

AIBroadcom's debt push signals AI compute is becoming Wall Street asset class

Aug 25, 08:06

The AI infrastructure race is entering a phase where access to silicon is no longer the only constraint. As compute demand expands into multi-gigawatt projects, the industry is increasingly confronting a second bottleneck: how to finance hardware and data-center capacity at the scale frontier-model developers now require.

AIZhongji Innolight's 1.6T surge shows AI networking is becoming next hyperscaler bottleneck

Aug 25, 08:06

The AI infrastructure boom has long been defined by a race to secure GPUs, memory and power. Yet as cluster sizes continue expanding, another layer of the system is beginning to emerge as an equally decisive constraint: the optical links responsible for moving data fast enough between increasingly dense pools of accelerators.

More coverage