Getting your news
Attempting to reconnect
Finding the latest in Climate
Hang in there while we load your news feed
Dell
Data center news, project activity, and monthly briefings for Dell.
Editor's picks
-
Building the UK’s Fusion AI Supercomputer
The UK government has announced a £45 million investment to build ‘Sunrise’, an AI-focused supercomputer for fusion research.
- Main announcement: The UK government (DESNZ) is funding £45 million to build Sunrise, a 1.4MW mission-focused AI supercomputer delivering 6.76 Exaflops of AI-accelerated modelling, targeted to begin operation in June 2026, located in the Oxbridge innovation corridor and dedicated to fusion energy research.
- Background and implementation details:Dell Technologies designed and delivered the compute infrastructure, working with AMD, Intel, WEKA, University of Cambridge, and the UK Atomic Energy Authority (UKAEA); Sunrise will support programmes including LIBRTI and STEP and uses dense, AI-optimised server configurations with rapid data storage to meet the 1.4MW power envelope constraint.
-
Fact of the Week: ASEAN Becomes the Middleman in US-China Tech Trade
Trelysa Long (ITIF) reports that U.S. trade policy shifts and tariffs have redirected high-tech manufacturing and final assembly away from direct U.S.-China trade toward Southeast Asia, with Vietnam playing a major intermediary role.
- Main finding: Over the past year U.S. trade policy toward China altered global trade patterns, with the U.S. trade deficit staying near $1.2 trillion in 2025 and China’s global trade surplus rising to roughly $1.2 trillion; companies including Lenovo, Apple, Dell, and HP moved portions of manufacturing and final assembly to Southeast Asia to reduce exposure to U.S. tariffs on Chinese goods.
- Background and mechanics: The shift reflects tariff avoidance and supply-chain diversification: only 1 percent of tech goods under HS code 84 coming from ASEAN faced tariffs compared to about 90 percent for those from China, and Chinese firms continued to supply upstream components to ASEAN-based manufacturing; rising demand for servers, networking equipment, semiconductors, and data centers (AI infrastructure) also increased ASEAN’s role.
-
NVIDIA GTC Taipei at COMPUTEX: Live Updates on What’s Next in AI
NVIDIA announced multiple wins at COMPUTEX 2026 Best Choice Awards and showcased new AI infrastructure products at NVIDIA GTC Taipei.
- Main announcement: NVIDIA’s Vera Rubin NVL72 won a Golden Award and the Sustainable Tech Special Award; Jetson Thor won a Golden Award; and Alpamayo won the Vehicle Technology and Smart Cockpit Category Award. The company is presenting at NVIDIA GTC Taipei (June 1-4 at COMPUTEX) and Jensen Huang will deliver a keynote Monday, June 1, 11 a.m. Taiwan time (Taipei Music Center).
- Details and technical/background facts:Vera Rubin NVL72 connects 36 NVIDIA Vera CPUs and 72 NVIDIA Rubin GPUs with a sixth-generation NVLink Switch, uses ConnectX-9 SuperNICs, Spectrum-X photonics switches and BlueField-4 DPUs; it claims up to 10x higher inference performance per watt and 10x lower cost per token, and up to 35x higher throughput per watt when paired with NVIDIA Groq 3 LPX. The system is 100% liquid-cooled at 45°C, has power shelves with 6x more onboard energy storage for power smoothing, and reduces compute-tray assembly time from two hours to five minutes. Jetson Thor offers up to 2,070 FP4 teraflops (configurable 40–130W) and Alpamayo includes 10-billion-parameter reasoning models, AlpaSim simulation framework and >1,700 hours of driving data in NVIDIA Physical AI Open Datasets.
-
Dell PowerRack Transforms AI Infrastructure with Scalable Compute, Networking and Storage
Dell Technologies announces PowerRack and PowerCool CDU C7000 at Dell Technologies World 2026.
- Announcement details: Dell announced PowerRack, a complete integrated rack-scale AI portfolio (compute, networking, Exascale Storage) at Dell Technologies World 2026, and introduced the PowerCool CDU C7000 that cools more than 220 kW in a 4U form factor and supports facility water inlet temperatures up to 40 degrees Celsius. The company says factory-validated racks arrive ready to run in just over six hours and that Dell AI deployment services deliver 84% faster time-to-value compared to self-assembly approaches. The announcement references IDC data that global AI infrastructure spending reached $82 billion in a single quarter of 2025.
- Background and supporting details: Dell positions itself as #1 rack-scale infrastructure provider in CY2025 (shipping more than twice the volume of the closest competitor). PowerRack is managed via Dell Integrated Rack Controller with OpenManage Enterprise (OME); PowerRack for Exascale Storage is built for 10+ PB environments and claims up to 6 TB/s per rack. Sources cited include IDC, Principled Technologies, ZestLab, and Morgan Stanley. The announcement is a product release/marketing post rather than a financial transaction or partnership agreement with implementation timelines beyond the product availability and performance claims.
-
Powering the AI factory: Building infrastructure for what comes next
Schneider Electric is showcasing AI-ready infrastructure at Dell Technologies World 2026.
Main announcement: Schneider Electric is presenting end-to-end infrastructure for AI at Dell Technologies World 2026, highlighting Motivair Coolant Distribution Units (CDUs), the MGX™ rack platform, 800VDC power architecture, and an AI Factory digital twin using NVIDIA Omniverse to support the full AI lifecycle (Design, Simulate, Build, Operate, Maintain).
- Event/publication details: Published May 15, 2026 on Schneider Electric blog; author Paul Tyrer; article references Dell Technologies World 2026 as the showcase venue.
Background and details: Emphasizes liquid cooling as essential for high-density AI deployments and the use of AVEVA™ digital twin technologies with NVIDIA Omniverse to validate power and cooling before deployment; cites collaboration with NVIDIA and Dell Technologies and positions infrastructure as a foundation for scalable AI.
-
How a Citizen-Owned Utility Built Infrastructure Residents Can Count On
Lafayette Utilities System (LUS) has standardized on Dell PowerStore as a unified storage platform and expanded Dell PowerProtect solutions to strengthen resilience and cyber protection.
- Main announcement: LUS standardized on Dell PowerStore for block and file workloads and, with Dell Services and a dedicated resident engineer, migrated 93TB block + 44TB file (137TB total)three times faster than previous efforts with no downtime, realizing ~50% better virtual performance, ~30% faster service deployment, backup windows down up to 75% (overall backups from ~20 hours to 4–5 hours, Exchange from 4 hours to 20 minutes), ~25% lower power use, and ~80% reduction in daily administration effort.
- Background and implementation details: The deployment included PowerStore features like Metro LUNs, synchronous replication, and file-level snapshots to achieve near-zero RPOs and fast failover; LUS also deployed PowerProtect Data Manager and PowerProtect Cyber Recovery with CyberSense for unified protection. The migration approach emphasized integration with existing tools (e.g., Active Directory) and operational simplicity; Dell engineers performed the migration and residency support during rollout.
Recent news
-
New Data Center Developments: June 2026
Data Center Knowledge has published a monthly roundup of global data center developments.
- Highlights include: CloudBurst breaking ground on a 1.2 GW flagship campus in Central Texas; Nvidia partnering with IREN to deploy up to 5 GW of global AI infrastructure with Texas’ Sweetwater as a flagship site; Prime Data Centers breaking ground on SMF02 (150,000 sq.ft, 18 MW IT load) in Sacramento; Applied Digital planning Delta Forge 1 — $3.6 billion, 300-acre AI campus in Boyce, Louisiana; Hive Digital/Buzz HPC planning an ~320 MW AI facility in the Greater Toronto Area.
- Additional concrete items and timelines: SoftBank plans up to €75 billion to develop 5 GW in France (targeting 3.1 GW by 2031); Ardian & Verne’s €5 billion digital campus (500 MW, with 200+ MW by 2030); TotalEnergies’ €100 million Pangea 5 supercomputer investment; Arcem’s Joroinen site delivering 60 MW by 2027 and 100 MW by 2029; CDC Data Centres’ 555 MW contract to be delivered with operations commencing in FY28 and FY29. All items are factual summaries from the article.
-
AI Server Market Update: Vendors Shift from Silicon to Services
Data Center Knowledge reports that server vendors are shifting toward software, professional services, and AIOps to win enterprise AI customers.
- Main announcement: Vendors including Dell, HPE, Lenovo, and Supermicro are emphasizing software management, professional services, AIOps, and liquid-cooling/packaged rack solutions to capture enterprise AI demand; IDC projects AI infrastructure spending to reach $487 billion in 2026 and surpass $1 trillion by 2029, while suppliers report large backlogs (e.g., Dell $43 billion AI backlog, HPE $5 billion AI systems backlog, Lenovo $15.5 billion AI server pipeline).
- Background & details: The article is an industry analysis citing interviews and earnings: IDC reported the global server market at $444 billion (2025); vendors report specific results such as Dell $9 billion AI-optimized server revenue (Q4 FY2027) and Supermicro $10.2 billion sales (FYQ3 FY2026); it highlights enterprise skill gaps, GPU/memory supply constraints, and differentiation via integration, delivery speed, power & cooling, and services.
-
Supermicro’s New AI Campus Embodies the Industrialization of AI Infrastructure
Supermicro announced the opening of its largest U.S. Data Center Building Block Solutions (DCBBS) campus near its San Jose headquarters on April 27, 2026.
- Main announcement: The new DCBBS campus spans ~32.8 acres and more than 714,000 square feet, becomes Supermicro’s fourth Bay Area location, expands the company’s regional footprint to nearly 4 million square feet, and will support advanced system design, domestic manufacturing, testing, service, and global distribution for Supermicro’s AI infrastructure portfolio. The facility includes 10 MW of on‑campus power capacity and is positioned as a rack‑scale, liquid‑cooled AI integration and validation hub.
- Background and related details: Supermicro frames this as a move from server manufacturing to rack-scale DCBBS integration, part of a global footprint spanning Taiwan, Malaysia, and the Netherlands; the company reported fiscal 2025 net sales of $22 billion (up from $15 billion the prior year), projected fiscal 2026 revenue of at least $33 billion, and in early May projected quarterly revenue of $11–$12.5 billion. On May 6, Supermicro signed a non-binding MOU with NANO Nuclear to explore pairing microreactor generation (KRONOS platform) with Supermicro’s liquid‑cooled AI systems (no commercial deployment timeline announced).
-
Data centre market set to hit USD $1.08 trillion by 2034
Polaris Market Research projects the global data centre market will reach USD $1,084.16 billion by 2034.
- Main announcement: Polaris Market Research forecasts the global data centre market will grow from USD $354.75 billion in 2024 to USD $1,084.16 billion by 2034, implying a CAGR of 11.50% from 2025 to 2034; key growth drivers named are cloud adoption, artificial intelligence, and edge technologies.
- Background & details: The report highlights hyperscale and edge expansion, increased demand from sectors such as banking, healthcare, technology, telecoms and government, a regional split with North America leading and Asia Pacific (India, China, Singapore, Australia) fastest-growing, and notes operational risks including high operating costs and supply chain constraints; named market participants include Amazon Web Services, Microsoft, Google, Cisco Systems, Dell Technologies, Equinix, NTT Global Data Centres, Schneider Electric.
-
Nutanix adds AI & cloud tools amid infrastructure push
Nutanix has announced additions to the Nutanix Cloud Platform including new AI, Kubernetes on bare-metal, expanded storage and cloud management capabilities.
Main announcement: Nutanix introduced Agentic AI (early access) and NKP Metal (early access), made Unified Storage 5.3 and Data Lens 2.0generally available, and released Nutanix Cloud Manager 2.0 GA; it also launched a Foundation Central appliance to simplify AHV deployment on servers from Cisco, Dell, Fujitsu, HPE and Lenovo and expanded synchronous DR support for Dell PowerFlex and integration for Everpure //c FlashArray. These features target AI workloads, bare-metal Kubernetes, air-gapped on-prem deployments, and multisite/multidomain cluster management.
Background and details: The updates address server and storage supply constraints and aim to broaden deployment options (on-premise, edge, public cloud) including AWS GovCloud support; other planned ecosystem support includes AMD GPU-accelerated servers, Dell PowerStore, NetApp ONTAP, Lenovo ThinkSystem, additional Cisco integrations, zero-copy migrations from VMware vSphere Virtual Volumes to AHV vDisks, and a certified integration between Nutanix Database Service and MongoDB Ops Manager.
-
Vertiv Targets AI Cooling Bottleneck with ThermoKey Deal
Vertiv has announced plans to acquire Italy-based heat-exchanger specialist ThermoKey.
- Main announcement: Vertiv announced plans to acquire ThermoKey, an Italy-based heat-exchanger specialist founded in 1991, to extend control across the full “thermal chain” (chip-level cooling through facility-level heat rejection). The company said the transaction is expected to close in the second quarter of 2026, subject to regulatory approvals.
- Background and details: ThermoKey brings dry coolers and microchannel heat exchangers, an active manufacturing footprint in EMEA, and OEM/system-integrator deployments; Vertiv frames the deal as expanding customer optionality for integrated liquid/air cooling and heat-rejection strategies, particularly in Europe where regulatory pressure on refrigerants and thermal approaches is intensifying.
-
Vertiv Targets AI Cooling Bottleneck with ThermoKey Deal
Vertiv has announced plans to acquire Italy-based ThermoKey.
- Main announcement: Vertiv will acquire ThermoKey, an Italy-based heat-exchanger specialist founded in 1991, to extend control across the full thermal chain (chip-level cooling through facility-level heat rejection); the transaction is expected to close in the second quarter of 2026, subject to regulatory approvals. The deal adds ThermoKey’s portfolio of dry coolers and microchannel heat exchangers and manufacturing capacity (EMEA-focused) to Vertiv’s offerings.
- Context and details: The move targets a growing heat-rejection bottleneck as AI workloads raise rack power densities; liquid cooling concentrates heat at the rack and shifts the burden to facility-level systems. Vertiv frames the acquisition around customer optionality, greater flexibility in refrigerant strategies and heat-rejection approaches, and responding to surging demand and tightening thermal-infrastructure supply, particularly in Europe.
-
Super Micro Indictment Highlights AI Infrastructure Supply Chain Risks
Super Micro Computer said co-founder and senior vice president Yih-Shyan “Wally” Liaw has resigned following a federal indictment that was unsealed on March 19 alleging a scheme to move systems containing Nvidia AI chips into China.
- Main announcement:Yih-Shyan “Wally” Liaw resigned from Super Micro Computer following a federal indictment unsealed on March 19 that alleges a scheme to transfer systems containing Nvidia AI GPUs into China; the company publicly acknowledged the resignation in response to the indictment.
- Background and context: The article documents how surging GPU demand, export controls, and supply-chain pressure intersect; analyst Matt Kimball highlights Super Micro’s historical positioning as a lower-cost, faster-to-market white-box vendor, raises concerns about governance and alleged ethical lapses, and notes that Nvidia was not implicated in the indictment.
-
Super Micro Indictment Highlights AI Infrastructure Supply Chain Risks
Super Micro Computer announced the resignation of co-founder and senior vice president Yih-Shyan “Wally” Liaw following a federal indictment unsealed on March 19 alleging a scheme to move systems containing Nvidia AI chips into China.
- Main action:Resignation of Yih-Shyan “Wally” Liaw after a federal indictment unsealed March 19 alleging an alleged scheme to transfer systems with Nvidia GPUs into China; the indictment document was linked in the article.
- Background and context:Export controls on high-end GPUs, surging GPU demand, and Super Micro’s historical positioning as a faster, lower-cost white-box vendor (compared to Dell, HPE, Lenovo) are central; analysts (Matt Kimball) flagged supply-chain integrity, vendor governance, and procurement diversification as immediate implications for data center operators and enterprise buyers.
-
Dell, HPE, and Others Unveil AI Innovations at GTC 2026
Dell announced advances to its Dell AI Factory, including the Dell Data Orchestration Engine and new storage and server roadmaps.
Main announcement:Dell announced the Dell Data Orchestration Engine (no-code/low-code) powered by Dataloop, plus Dell Lightning File System (now available) and Dell Exascale Storage for deployments exceeding 10 petabytes; Dell PowerEdge systems with Nvidia RTX PRO 4500 Blackwell Server Edition are shipping now and the Dell AI Factory with Nvidia Modular Architecture will begin in April 2026. Dell also plans availability timelines: PowerEdge R9822 and M9822 Vera CPU servers in September 2026 and the PowerEdge XE9812 (Vera Rubin NVL72) in the second half of 2026.
Background and other details:HPE announced HPE AI Grid and Private Cloud AI (air-gapped options) with Private Cloud AI scaling up to 128 GPUs, and timeline notes: HPE Vera Rubin NVL72 rack-scale available December 2026 and the HPE Cray Supercomputing GX240 Vega CPU blade in 2027. Storage vendors: NetApp launched NetApp AI Data Engine (AIDE) co-engineered with Nvidia; Everpure extended Evergreen//One to FlashBlade//EXA and announced a Data Stream beta launching later in 2026; VAST Data released pre-built open-source pipelines for Nvidia AI blueprints.
-
NVIDIA Debuts Vera CPU to Anchor Next Phase of AI Infrastructure
Nvidia has launched the Vera CPU, a processor designed to serve as the orchestration/control plane for agentic AI, unveiled at GTC 2026.
- Launch details: Vera was unveiled at GTC 2026; built on Nvidia Grace architecture with 88 custom Arm-based “Olympus” cores (each core can run two tasks via Spatial Multithreading), LPDDR5X memory delivering up to 1.2 TB/s bandwidth, and a second-generation Scalable Coherency Fabric; tightly integrated with Rubin GPUs via NVLink-C2C (up to 1.8 TB/s coherent bandwidth) and positioned as the host CPU for HGX Rubin NVL8 systems.
- Scale & ecosystem: Rack-scale Vera systems support up to 256 liquid-cooled CPUs and >22,000 concurrent CPU environments; ecosystem and early adopters named include Meta Platforms, Alibaba, ByteDance, Oracle Cloud Infrastructure, CoreWeave, Nebius Group, Lambda, and OEMs Dell Technologies, HPE, Lenovo, Supermicro; Nvidia and analysts claim lower power than x86 and a design optimized for agentic and reinforcement-learning orchestration.