DIGITIMES ICT: Server, IPC, cloud computing, IoT news

Home Tech ICT Server, IPC, cloud computing, IoT

Wednesday 29 October 2025

Digital transformation specialist IISI set for November IPO, drives SaaS shift and global expansion

Digital transformation solutions integrator International Integrated Systems (IISI) held a pre-IPO performance briefing on October 28, announcing plans to list in November. The company aims to fuel its next growth phase through cloud-native SaaS offerings, transitioning toward a subscription-based revenue model built on long-term partnerships while actively expanding into overseas markets.

Wednesday 29 October 2025

Bloom Energy offers 800V DC power solutions for AI data centers

As AI data center energy consumption continues to surge, Nvidia has introduced a high-voltage direct current (HVDC) architecture aimed at reducing overall power usage in AI computing infrastructure. Meanwhile, US-based Bloom Energy has presented a supporting solution that uses fuel cells to supply 800V DC directly to servers, integrating carbon capture and combined heat and power (CHP) systems to recycle waste heat and boost efficiency, potentially cutting operational costs by about 15%.

Wednesday 29 October 2025

Commentary: With new AI chips, Qualcomm joins race to challenge Nvidia in the cloud

Qualcomm on Tuesday unveiled its latest AI accelerator cards, the AI200 and AI250, along with corresponding rack-level system designs, signaling a major expansion of the company's ambitions in cloud-based AI computing. The two products are slated for commercial release in 2026 and 2027, respectively, and Qualcomm pledged to launch new AI accelerator models every year going forward.

Wednesday 29 October 2025

AI server liquid cooling demand surges, heat dissipation firms race to expand capacity

Nvidia plans to launch its Vera Rubin architecture AI servers in 2026, significantly increasing the share of liquid cooling solutions. In response, heat dissipation manufacturers are aggressively expanding production capacity during this critical growth period. Industry sources note that Nvidia's new platform will drive much higher demand for key liquid cooling components such as quick connectors and cold plates.

Wednesday 29 October 2025

International firms join Taiwan supply chain to launch AI accelerator platform

To accelerate AI adoption across industries, the IPO Forum is partnering with multinational companies and local supply chains to launch a "Trusted AI Accelerator" platform in Taiwan. Terence Liao, general manager of Dell Taiwan, who serves as IPO Forum chairman, emphasized that since the emergence of generative AI, the industry has undergone extensive trial and error over the past two years, making 2025 a critical turning point. Liao expressed hope that the accelerator platform will leverage the strength of major players to empower smaller enterprises and enhance the resilience of Taiwan's industrial ecosystem.

Wednesday 29 October 2025

Huawei Cloud on the back foot in China's AI race

Huawei Cloud is navigating one of its most turbulent periods yet. Confronted by internal and external headwinds—from a disciplinary probe and sweeping restructuring to a shrinking market share—the company that once rode China's cloud boom is now at a decisive crossroads.

Wednesday 29 October 2025

Foxconn expands US AI server buildout, joins Nvidia-Stellantis-Uber robotaxi effort

Hon Hai Technology Group (Foxconn) used this week's GTC event in Washington to highlight two fronts of collaboration with Nvidia: expanding US-based AI server manufacturing and joining a new alliance with Stellantis and Uber to develop level-4 robotaxis.

Wednesday 29 October 2025

Alibaba Cloud's price cuts widen global reach, deepen China's AI cloud duopoly

As China's AI cloud competition intensifies, Alibaba Cloud has introduced new price cuts across overseas markets to accelerate its global expansion, extending the "dual-giant" rivalry beyond domestic boundaries.

Wednesday 29 October 2025

KES 2025: LG CEO shows interest in Giga Computing's liquid cooling

Facing accelerated AI server upgrades and soaring power consumption, Giga Computing is promoting next-generation cooling technology and one-stop data center solutions, with South Korea identified as a key market. The company recently participated in the Korea Electronics Show (KES 2025) to highlight its competitive edge to potential local customers.

Tuesday 28 October 2025

AMD MI355X to power US Department of Energy's new supercomputers

AMD and the US Department of Energy (DOE) have announced two next-generation supercomputing systems to be built at Oak Ridge National Laboratory (ORNL) — the Lux AI supercomputer and the Discovery supercomputer. The systems will expand the DOE's high-performance computing (HPC) and artificial intelligence (AI) capabilities as part of a combined US$1 billion public-private investment.

Tuesday 28 October 2025

Applied Digital secures US$15 billion lease to expand computing capacity at Polaris Forge 2 campus

Applied Digital, a firm specializing in data center design, development, and operations, has signed a US$15 billion infrastructure lease agreement spanning 15 years with an unnamed major US hyperscaler. The deal aims to add 200 MW of computing power at the company's Polaris Forge 2 campus in North Dakota, increasing its total leased capacity at both Polaris Forge 1 and Polaris Forge 2 to 600 MW.

Tuesday 28 October 2025

Foxconn commits US$1.37 billion to build AI supercomputing center and expand cloud infrastructure

Driven by the global surge in generative AI, Foxconn announced plans to invest NT$42 billion (approximately US$1.37 billion) toward procuring AI computing power and cloud business operational IT equipment. This investment aims to establish an AI computing cluster and supercomputing center, expand its cloud platform, and accelerate development in smart manufacturing, smart electric vehicles (EV), and smart cities.

Tuesday 28 October 2025

Trump turns a blind eye to US's AI power grid problem

AI data centers are putting tremendous pressure on the US's already fragile power infrastructure. In addition to Nvidia CEO Jensen Huang and Tesla CEO Elon Musk, multiple Wall Street investment banks and industry organizations have recently warned that the US power grid faces increasing risks of blackouts and rising electricity prices.

Tuesday 28 October 2025

Qualcomm launches AI200, AI250 to challenge Nvidia’s dominance in AI infrastructure

Qualcomm Technologies, Inc. has unveiled its next-generation AI inference solutions for data centers — the AI200 and AI250 chip-based accelerator cards and racks — marking its boldest move yet into the fast-growing generative AI infrastructure market. Designed to deliver high performance per dollar per watt, the new lineup leverages Qualcomm's expertise in neural processing unit (NPU) technology to offer scalable and energy-efficient AI inference for large language and multimodal models.

Tuesday 28 October 2025

Apple ships first AI servers from Houston to power US-made cloud and AI expansion

Apple has begun shipping its first batch of artificial intelligence (AI) servers from a new factory in Houston, Texas, marking a key milestone in its US$600 billion investment in the US. The servers will be deployed across Apple's data centres nationwide to power its Private Cloud Compute (PCC) and Apple Intelligence services.

Monday 27 October 2025

Topco accelerates overseas push to align with clients’ new fab investments

Geopolitical tensions are creating turbulence in global trade. According to TSMC founder Morris Chang, globalization is dead. Semiconductor material distributor Topco Scientific also reported sensing the trend toward localized supply chains.

Monday 27 October 2025

AI poisoning: Anthropic study finds just 250 malicious files can compromise the largest LLMs

Contrary to long-held beliefs that attacking or contaminating large language models (LLMs) requires enormous volumes of malicious data, new research from AI startup Anthropic, conducted in collaboration with the UK AI Security Institute and the Alan Turing Institute, has found otherwise. The study demonstrates that as few as 250 poisoned documents can implant backdoor vulnerabilities into LLMs, regardless of their size or training dataset volume.

Monday 27 October 2025

US and Malaysia strike deal to secure rare earth supply chain

On October 26, US President Donald Trump attended the ASEAN Summit in Kuala Lumpur, Malaysia, where the US and Malaysia signed a trade agreement and a memorandum of understanding (MOU) on critical mineral supply cooperation.

Monday 27 October 2025

Ablecom breaks ground on phase 4 plant in northern Taiwan to meet AI server demand

To capitalize on the surging demand for artificial intelligence (AI) servers, Ablecom, a leading supplier of server chassis and cooling solutions, has commenced construction of its fourth-phase factory in the Supermicro Science and Technology Park in Taoyuan, northern Taiwan. The new facility will cover 79,339 square meters) and is set to become the largest building within the park, as well as Ablecom's largest production line in Taiwan, with completion targeted for the end of 2026.

Monday 27 October 2025

As Huawei stumbles, ByteDance quietly rewrites China’s AI cloud order

Huawei Cloud CEO Zhang Ping'an and several top executives were recently demoted and had their salaries reduced, officially due to "performance falsification." The punishment, however, reflects a broader shift in the AI era: the inseparable link between large models, AI applications, and AI infrastructure. ByteDance's Volcengine has emerged as a key disruptor representing this integrated vision in China's cloud ecosystem.

Monday 27 October 2025

Taiwan firm turns to smart server locks as AI data centers demand tighter security

Electronic component distributor WFE Technology is set to list on the Taipei Exchange (TPEx) in mid-to-late November 2025, positioning itself to benefit from rising cybersecurity and environmental monitoring standards for data centers and AI servers.

Saturday 25 October 2025

OpenAI, AMD, Broadcom unite behind Ethernet to reshape AI infrastructure

The Ethernet for Scale-Up Networking (ESUN) Alliance officially launched at the 2025 Open Compute Project (OCP) Global Summit, uniting North America's AI heavyweights, including OpenAI, Meta, AMD, Nvidia, Microsoft, and Broadcom. Its goal is to create an open Ethernet interconnect standard to address AI chip connectivity fragmentation and redefine global AI infrastructure. The collaboration marks a rare alignment across major rivals in pursuit of interoperability.

Friday 24 October 2025

Micron begins sampling 192GB low-power DRAM module for AI servers

Micron Technology has started customer sampling of its new 192GB SOCAMM2 (Small Outline Compression Attached Memory Module), a low-power DRAM module designed for AI data centers.

Friday 24 October 2025

Taiwan shifts from assembly to architecture in the AI data center race

As global demand for AI computing power skyrockets, the data center landscape is undergoing a significant transformation — shifting away from standalone server performance and toward system-level, rack-based upgrades. At the OCP Global Summit 2025, the industry spotlight moved beyond individual servers to full racks designed with 800V direct current (DC) power, liquid cooling, and high-bandwidth network integration.

Thursday 23 October 2025

Samsung set to roll out Ballie AI robot in Vietnam with launch imminent

Samsung Electronics is reportedly set to start production of its AI-powered home assistant robot, Ballie, at its facility in Ho Chi Minh City, Vietnam, suggesting that the long-delayed device is finally nearing mass production.

More Server, IPC, cloud computing, IoT coverage

BIZ FOCUS

Oct 2, 10:32

SK Group partners with OpenAI to advance global AI infrastructure

MOST-READ
7 DAYS NEWS

Full list

MULTIMEDIA

Inside Korea’s CES: Flying Cars and Robots Are Here!

INDUSTRY GUIDEBOOKS

Geopolitics may bring dramatic shifts in semiconductor industry