45k user reviews & 1 million votes, over 29k web hosting brands & 85k plans since 2004!
X
Menu
WHTop → 📋 News → Wistron Expands AI Server Capacity as Demand Outruns Supply
Wistron Expands AI Server Capacity as Demand Outruns Supply
📅 - Wistron is putting fresh money into AI server rack capacity across California, Taiwan, Vietnam, and Malaysia, a manufacturing expansion that says more about enterprise infrastructure pressure than supplier optimism. The Taiwanese contractor is following demand from Nvidia-linked systems and global cloud operators, while warning that CPUs and other components remain tight enough to slow deliveries for now.
The board action gives a cleaner view of what the AI infrastructure buildout looks like when it leaves keynote slides and hits factory planning. Wistron approved roughly $150 million for WisLab EMS, its California subsidiary, taking total investment in that unit to about $212.1 million. [...][... Check source for end of article ...]
📅 - Lenovo Targets AI Cost Issue with Hybrid Inference Systems - Lenovo is expanding its Hybrid AI Advantage portfolio with new inferencing and agentic AI systems built with NVIDIA, Intel, Red Hat and Canonical, arguing enterprises need cheaper token economics and local control as AI spending shifts from experiments to always-on production workloads running across PCs, workstations, data centers and clouds under tighter governance pressure globally now constraints.
The claim is straightforward enough: enterprise AI is moving out of the lab, and the bill is starting to look ugly. Training models was expensive, visible and centralized. Inference is different. It happens constantly. Search requests, service desk agents, code assistants, retail kiosks, HR [...]
📅 - OpenAI and Broadcom Unveil Jalapeño Inference Chip - OpenAI and Broadcom have unveiled Jalapeño, a custom inference accelerator designed around large language models, claiming early lab results show better performance per watt than current leading systems, as OpenAI extends its infrastructure ambitions from models and products into silicon intended for gigawatt-scale data centers beginning with partners by late 2026 and, eventually, more generations of hardware.
The chip is being framed as OpenAI's first Intelligence Processor, which is grand language for a simple industrial reality: the company does not want to remain fully dependent on other people's accelerators as inference demand becomes the expensive center of the AI business. [...]
📅 - QCi Acquires NHanced to Advance Quantum Chip Manufacturing - Quantum Computing Inc. has completed its acquisition of NHanced Semiconductors for $73.1 million in cash and stock, with up to $72 million more tied to performance targets, giving the quantum optics and integrated photonics company a larger U.S. manufacturing base as it tries to convert prototypes into commercial photonic and quantum hardware at scale for real customers.
The deal is about control. Not the abstract kind. Process control, packaging control, testing control, customer delivery control. QCi has been talking about moving from research-driven innovation and prototyping into scalable commercial production, and NHanced gives it a more substantial answer to the obvious question [...]
📅 - Colt DCS Reports 27% Emissions Cut as AI Data Center Load Rises - Colt DCS says it cut greenhouse gas emissions 27% against its 2019 baseline, a rare hard number in a data center market being stretched by AI demand, power scarcity, and customer scrutiny. The hyperscale operator also reported full renewable electricity coverage for Scope 2, largely through certificates, while expanding a new global design template for lower-carbon builds.
The figure lands in a sector now being asked to do two difficult things at once. Build faster. Pollute less. Do both while customers ask for denser racks, regulators examine grid pressure, and investors want growth without the reputational baggage that increasingly follows energy-intensive infrastructure.
Colt Data [...]
📅 - KAYTUS Debuts Modular AI Factory Builds And Ops Platform - KAYTUS used ISC 2026 to present three pieces of its AI infrastructure strategy: factory-built liquid-cooled data centers, a management platform for rack-scale operations, and a broader full-stack European offering, all aimed at operators trying to deploy dense GPU capacity faster than conventional construction, procurement, integration, financing, staffing, and operations models can realistically absorb without introducing new risk.
The company is positioning itself against a very specific infrastructure problem. AI capacity is no longer just a server procurement exercise. It is land, grid power, switchgear, chillers, coolant loops, network topology, software control, commissioning labor, [...]
📅 - South Korea’s $576B AI Chip Plan Faces DRAM Lawsuit Shadow - South Korea is tying its next industrial plan to memory chips, AI data centers and regional development, even as Samsung, SK Hynix and Micron face a U.S. class-action lawsuit accusing them of restricting DRAM supply. The two stories sit awkwardly together: national capacity building on one side, alleged market control on the other, and pricing risk too.
President Lee Jae Myung on Monday announced semiconductor and AI projects framed around more than $576 billion in spending over several years, with Samsung Electronics and SK Hynix executives joining the broadcast. The political signal was blunt enough. South Korea wants to protect its memory-chip advantage before AI infrastructure demand [...]