Innovations in Data Center Power and Cooling Solutions

Bergen Engines signs 750MW data centre deal

Bergen Engines, a Norwegian manufacturer of medium-speed gas and dual-fuel engines, has signed an agreement with Crusoe to provide up to 750MW of power generation capacity for AI data centre developments in the United States. The agreement comprises a 438MW contract and a further 310MW letter of intent, supporting Crusoe's expanding portfolio of AI infrastructure projects. Crusoe develops large-scale AI data centre campuses using a combination of grid power, natural gas generation, renewable energy, and battery storage. The company deploys both grid-connected and behind-the-meter power infrastructure to support the high energy demands of AI workloads. John Adams, Senior Vice President of Power at Crusoe, says, "The pace of AI infrastructure development demands builders who treat power as a first-class AI infrastructure layer. "Bergen’s gensets give us the reliable baseload power we need to energise large-scale campuses, deployable on our timeline. We’re building AI factories at record speed, and this agreement helps us maintain that pace." Under the initial contract, Bergen Engines will supply 27 gas-powered generating sets rated at 12.5MWe and 20 units rated at 5MWe. Additional units are included within the letter of intent, with deliveries planned across multiple US locations through 2027. On-site generation supports growing AI power demand The generators are intended to provide continuous baseload power for AI data centres operating around the clock. The systems will incorporate alternators from Marelli Motori and dynamic power stabilisation technology from Piller Power Systems. According to the companies, the technology is designed to manage rapid fluctuations in electricity demand associated with computing-intensive workloads. Dean Richards, CEO of Piller Power Systems, says, "AI workloads have a distinct power profile that demands purpose-built generation and stabilisation technology. "SHIELD-X is designed to manage those dynamics, protecting the generation assets and maintaining stable plant operation while ensuring consistent power quality for the data centre." As AI infrastructure capacity expands, developers are increasingly turning to on-site and behind-the-meter power generation where grid connections are unavailable or unable to support required capacity within project timescales. Theo Lorentzos, Vice President of Sales for Bergen Engines Americas, notes, "The pace of AI infrastructure development is unlike anything the power generation industry has seen before. "In this market, access to power determines how fast you can scale. Crusoe’s model is built around speed and stable power, and our solution is designed to deliver both." The agreement forms part of a wider trend towards dedicated power infrastructure for AI data centres, enabling developers to accelerate deployments while reducing reliance on traditional utility connection timelines. For more from Bergen Engines, click here.

Schneider Electric unveils Uniflair XCA chillers

Global energy technology company Schneider Electric has introduced the Uniflair XCA range of air-cooled and free-cooling chillers, designed for high-density, liquid-cooled data centres supporting AI workloads. The new portfolio comprises the Uniflair XCAC air-cooled series and the Uniflair XCAF free-cooling series. Both incorporate oil-free centrifugal compressors with magnetic bearing technology and variable-speed drives to support operation across varying thermal loads and environmental conditions. The chillers are available in six sizes, ranging from 1,200kW to 2,500kW, and utilise low global warming potential (GWP) refrigerants. Schneider Electric says the systems are designed to support elevated water temperatures commonly associated with liquid cooling deployments in AI data centres. Andrew Bradner, Senior Vice President, Cooling Business at Schneider Electric, notes, "Energy efficiency, adaptability, and reliability are essential components of liquid cooling systems for AI-optimised data centres, and we’ve designed the Uniflair XCA line with these most important design features at the forefront. "With adaptable water operating temperatures and versatile deployment options, the XCA line features a system-level approach that gives operators scalability, enhanced performance, and long-term peace of mind as data centre complexity continues to rise." Cooling infrastructure adapts to rising AI power densities As AI applications, GPU clusters, and liquid cooling deployments increase data centre power densities, cooling infrastructure is becoming an increasingly important factor in facility efficiency and reliability. The Uniflair XCA platform incorporates oil-free magnetic bearing centrifugal compressors, which remove the need for lubrication systems and are intended to reduce maintenance requirements and mechanical losses. The chillers also feature a spray evaporator combined with V-shaped microchannel coils, designed to improve heat exchange performance while reducing refrigerant volume and material usage. For free-cooling deployments, the XCAF models support water outlet temperatures of up to 33°C and are designed to operate in ambient temperatures ranging from -20°C to 52°C. Schneider Electric states that, in suitable climates, the free-cooling configuration can reduce energy consumption compared with mechanical cooling systems by extending free-cooling operating periods. The range can also be configured with a variety of electrical, hydraulic, acoustic, and performance options to suit different deployment requirements. Additionally, a quick restart capability is included, enabling systems to reportedly return to full operating capacity within three minutes of a power outage. New control features target operational efficiency The XCA range also introduces new firmware and control functions designed to optimise cooling performance. These include variable-speed pump algorithms supporting constant flow, constant temperature differential, and constant head pressure operation, alongside advanced fan control modes that can be adjusted according to temperature, load conditions, or scheduled operating periods. Additional monitoring capabilities include energy metering and real-time water flow measurement to provide greater visibility into system performance. According to Schneider Electric, these features are designed to reduce compressor cycling and improve long-term operational stability. The first Uniflair XCA chiller units are scheduled to begin shipping globally in June 2026. For more from Schneider Electric, click here.

Panduit expands fault managed power portfolio

Panduit, a manufacturer of electrical and network infrastructure hardware, has launched the second generation of its Fault Managed Power System (FMPS), introducing higher power density and additional deployment options for enterprise environments. Designed for centralised power distribution, the FMPS Gen 2 platform is intended for use across campuses, warehouses, and large distributed facilities. The system uses Class 4 fault managed power technology, which allows power to be delivered over longer distances using low-voltage installation methods. According to Panduit, the platform is UL 1400 listed and SIL 3 rated, enabling organisations to distribute power while reducing electrical hazards and simplifying installation requirements. The company also states that the system uses less copper than traditional power distribution methods and remains backward compatible with existing FMPS deployments, allowing infrastructure upgrades without replacing existing installations. A key feature of the platform is the consolidation of backup power systems. By centralising UPS infrastructure rather than deploying units within individual intermediate distribution frames (IDFs), organisations can reduce equipment requirements, maintenance demands, and space utilisation. New hardware targets enterprise and edge deployments The FMPS Gen 2 portfolio includes a new 2kW system comprising a 1kW transmitter, a 2kW power supply, and a 2kW receiver. The range also includes a 600W single-channel receiver designed for applications such as lighting and security systems. Additional updates include higher-density power delivery within the same footprint, expanded receiver options, and support for both PoE++ and DC-powered devices. The platform is designed to support a range of applications including enterprise networking equipment, security and surveillance systems, wireless and in-building cellular infrastructure, lighting, and smart building technologies. Mahmoud Ibrahim, Senior Business Development Manager at Panduit Ventures, says, "FMPS Gen 2 reinforces our commitment to making enterprise power safer, simpler, and more efficient. "By increasing power density and enabling true UPS consolidation, customers can place power where it’s needed, remove complexity from IDFs, and confidently support the growing demands of modern networks - all without introducing new risk." The platform also incorporates monitoring and management capabilities intended to provide centralised visibility of connected infrastructure and support future expansion. Tom Kelly, Chief Technology Officer at Panduit, explains, "FMPS is engineered and designed by Panduit as a complete power platform, integrating power, cabling, and physical infrastructure into a single, coordinated solution. "Drawing on the expertise that developed the first generation of UL-listed Class 4 power distribution products, Panduit has engineered a second-generation system that aligns with where the market is going while also meeting requests from customers and partners in the space. "We’re excited to see the market transformation taking shape, as Class 4 power distribution adoption grows." For more from Panduit, click here.

VIRTUS installs super-grid transformers at Berlin campus

VIRTUS Data Centres, a UK data centre owner-operator and part of ST Telemedia Global Data Centres (STT GDC), has completed the installation of two 185MVA super-grid transformers at its Wustermark campus in Berlin/Brandenburg, Germany. According to the company, the transformers are among the largest deployed at a European data centre and represent a key milestone in the development of the site. The Wustermark campus is expected to become the first data centre campus in the Berlin/Brandenburg region to connect directly to a 380kV transmission network. VIRTUS says this will give customers the option of operating without diesel generators while maintaining access to conventional backup generation where required. The transformers form part of the campus's initial 300MW capacity, with power supplied through a dedicated 500MW substation and dual direct connections to the 50Hertz 380kV network. VIRTUS says the integration with the 50Hertz Wustermark substation and the high-voltage transmission connections are designed to provide a resilient and stable power architecture for large-scale data centre operations. High-voltage design targets efficiency and resilience The company says the site has been designed to support both traditional generator-backed operations and a generator-free operating model. As with other VIRTUS facilities, the campus will operate using 100% certified renewable electricity. The site is also located close to regional renewable energy resources, including onshore wind generation. According to VIRTUS, the higher-voltage transformer design provides several operational benefits, including improved electrical efficiency, reduced transmission losses, increased system stability, and enhanced resilience for high-density computing environments. The company adds that the approach may also help reduce system usage charges and long-term energy costs. Mike Golding, SVP of Construction at VIRTUS Data Centres, says, “Delivering the Wustermark Campus has been one of the most ambitious engineering programmes VIRTUS has undertaken to date. “From the 380kV connections to the deployment of these super-grid transformers, every element has been designed to deliver levels of resilience and scalability that have not previously been available in this region. “This campus represents a new generation of infrastructure - one that supports AI-scale growth, reduces reliance on generators, and aligns with the future of renewable energy.” For more from VIRTUS, click here.

ChemTreat joins Dow coolant network for data centres

ChemTreat, a US provider of industrial water treatment chemicals and cooling system services, has joined materials science company Dow’s Coolant Care Network as a strategic US service provider for AI and liquid-cooled data centre environments. Under the agreement, ChemTreat becomes Dow’s only preferred service provider in Virginia, USA, and will provide national support for the company’s coolant management programme. According to Dow, the Coolant Care Network combines coolant supply, fluid testing, data analysis, and field support within a single framework for data centre operators. ChemTreat will provide on-site services including fluid sampling, mitigation, and coolant optimisation, working alongside Dow-qualified laboratories and technical specialists. The companies say the collaboration is intended to support data centres deploying liquid cooling systems for AI and high-density compute workloads. Ashour Khamis, President of ChemTreat, notes, “The data centre industry is under enormous pressure to scale liquid cooling environments to meet AI-driven workload demands. “Pairing ChemTreat’s proven service-focused approach with Dow’s decades of thermal fluid innovation and reliable global supply chain allows us to help customers quickly deploy mission-critical systems and maintain reliable cooling lifecycle performance.” Liquid cooling demand grows alongside AI workloads ChemTreat says its data centre offering includes water treatment technologies, monitoring systems, specialist chemistries, and support for direct-to-chip cooling loops and facility cooling infrastructure. Through the partnership, the company will also provide access to Dow’s DOWFROST LC and DOWFROST HD heat transfer fluids, alongside certified coolant testing services and technical support. Chuck Carn, Data Center Growth Platform Director at Dow, says, “This collaboration reflects Dow’s clear understanding of the operational complexity data centre operators face as cooling systems become more critical to performance and uptime. “Collaborating with experienced service providers like ChemTreat, who uphold rigorous technical and service standards, is key to helping customers run their operations smoothly and with confidence.” The companies say the partnership is designed to address increasing cooling requirements as AI infrastructure deployment continues to expand globally.

MPs warn grid failures could cost Britain the AI race

The All-Party Parliamentary Group (APPG) for Data Centres has published its Insights Paper, summarising findings from its inaugural 'Call for Evidence'. The group is a cross‑party group of UK MPs and Peers that fosters parliamentary understanding of data centre development, examines sector challenges (particularly planning, energy, resilience, and sustainability) and makes evidence‑based policy recommendations to support UK digital infrastructure and economic growth. Notably, respondents to its Call for Evidence signalled a substantial appetite to invest in UK data centre infrastructure. Operators including Ark Data Centres, Nebius, Pure DC, and VIRTUS collectively identified £11–12 billion in specific investment plans, while Microsoft's submission committed a further £22 billion to UK AI infrastructure. Despite this intent, respondents consistently described a set of interconnected structural barriers constraining the pace and location of development: • Grid access and energy supply ranked as the sector's top priority, with 52% of respondents placing it first. • Planning was placed in the top three by nearly four in five respondents (79%), while energy costs, sustainability, water use, and skills also featured prominently across submissions. • The APPG is particularly keen to hear further evidence from community representatives, local authorities, and organisations with experience of data centre development outside London and the South East. Parliamentary comments Chris Curtis MP, Chair of the Data Centres APPG, notes, "This Call for Evidence shows that while significant investment is ready to support the UK's expanding AI and digital economy, it remains constrained by grid access, energy costs, and planning inconsistencies. “The APPG will use this evidence over the coming year to work constructively with stakeholders and the Government to ensure that there is a well-informed view on how data centre infrastructure drives our national economic growth." Alison Griffiths MP, Vice-Chair of the Data Centres APPG, adds, “The submissions to this Call for Evidence make clear that the barriers to data centre development are not insurmountable. "They highlight gaps in the consistent application of planning policy by local authorities, as well as the need to ensure electricity cost competitiveness is felt across every part of the country. "It is clear there are practical steps the Government can take to strengthen the UK’s leadership in digital infrastructure [and] I look forward to exploring these issues further in our upcoming evidence sessions.” David Reed MP, Officer of the Data Centres APPG, highlights, "The submissions from academic institutions such as Exeter, Durham, and Oxford remind us that research computing infrastructure is increasingly cost-prohibitive for academia. This gap risks undermining the UK's long-term international scientific competitiveness. “As the APPG deepens its work, I look forward to hearing from a broad range of stakeholders in this vital debate and developing practical solutions that support a thriving data centre ecosystem.” The Rt Hon. the Lord (Philip) Hunt of King’s Heath OBE, Officer of the Data Centres APPG, concludes, "Sustainability is not a secondary consideration for this sector; it is central to its long-term viability and its licence to operate in communities across the UK. “The evidence on waste heat recovery is particularly striking: the UK is currently capturing just 3–5% of the heat generated by data centres, against a backdrop of a national housing and energy challenge that demands innovative solutions. The APPG will be pressing hard on what policy levers can unlock this opportunity."

Schneider to showcase AI infrastructure at Datacloud

Global energy technology company Schneider Electric has announced it will present its latest AI-ready data centre technologies during the Datacloud Global Congress 2026 in Cannes, France, from 1–4 June 2026. The company says it will showcase technologies designed to address increasing demands around power delivery, cooling, resiliency, and high-density compute as AI workloads continue to scale. According to Schneider Electric, the event will focus on infrastructure designed for AI deployments, including power architectures, liquid cooling systems, software platforms, and digital services for data centre environments. The company cites forecasts from Morgan Stanley and Gartner predicting significant growth in AI-related infrastructure investment and global AI spending over the coming years. Industry discussions to focus on AI growth challenges On 2 June, Frédéric Godemel, EVP of Energy Management Business at Schneider Electric, will join representatives from Oracle, DATA4, QTS Data Centers, and CBRE for a keynote panel examining AI infrastructure demand and the development of neocloud deployments. Later that same day, Thierry Chamayou, Vice President of Cloud and Service Providers in EMEA at Schneider Electric, will participate in a separate panel discussing energy investment strategies and utility collaboration for AI infrastructure projects. Participants include representatives from GreenScale, Trench Group, Kao Data, JSM Group, and Solar Turbines. Marc Garner, Global President of Schneider Electric’s Cloud and Service Provider Segment, says, “AI is fundamentally reshaping the future of digital infrastructure, creating new demands around power, cooling, and resiliency at unprecedented scale. “At Datacloud Global Congress, we will demonstrate how collaboration across the ecosystem is enabling the next generation of AI factories and helping organisations build scalable, resilient, and sustainable infrastructure built for the AI era.” Liquid cooling and AI data centre technologies on display Schneider Electric will also present technologies including its 800VDC architecture, liquid cooling systems from Motivair, high-density racks, and digital modelling platforms. The company says demonstrations will include NVIDIA Omniverse integrations, digital twin technologies, and NVIDIA reference designs for the GB300 NVL72 platform. Visitors to Stand 122 will also be able to view the company’s EcoStruxure IT DCIM and EcoStruxure Foresight software platforms, alongside its microgrid and data centre services portfolio. For more from Schneider Electric, click here.

Schneider supplies AI infrastructure for Lake Mariner campus

Global energy technology company Schneider Electric and Motivair, a provider of liquid cooling systems for data centres, owned by Schneider Electric, have completed phased delivery of more than $290 million (£215 million) in AI infrastructure systems for TeraWulf at its Lake Mariner data centre campus in New York, USA. The companies say the project is focused on supporting scalable AI and high-performance computing (HPC) infrastructure, with the Lake Mariner campus expected to support up to 750MW of power demand once fully developed. TeraWulf explains that the site is being redeveloped from a former industrial location into a data centre campus designed for AI, cloud, and HPC workloads, and that the project combines power infrastructure, cooling technologies, racks, monitoring systems, and engineering services from Schneider Electric and Motivair. The company adds that the campus benefits from access to existing power infrastructure and a regional electricity grid that is approximately 89% zero-carbon. Sean Farrell, Chief Operating Officer at TeraWulf, comments, “TeraWulf’s strategy is centred on delivering scalable, energy-efficient infrastructure capable of supporting the increasing intensity of AI and HPC workloads. "By working closely with industry leaders like Schneider Electric and Motivair, we are accelerating the development of AI-ready capacity at our Lake Mariner facilities while reinforcing the strong operational foundations needed to support long-term customer demand.” Project combines power, cooling, and monitoring systems Schneider Electric says the deployment included Galaxy VX UPS systems, lithium-ion battery systems, EcoStruxure IT Data Center Expert software, and NetShelter racks and enclosures. Meanwhile, Motivair supplied coolant distribution units (CDUs), in-rack manifolds, chilled door technologies, and associated support services. The companies state that the systems were delivered within a 12-month deployment timeframe as TeraWulf accelerated construction of its AI-ready data centre facilities. The Lake Mariner campus is also supported by long-term customer agreements with Core42 and Fluidstack, with the latter being backed by Google. Manish Kumar, Executive Vice President, Secure Power & Data Centers at Schneider Electric, notes, “As demand for AI infrastructure accelerates, ‘time to power’ has become a defining constraint on growth. "Operators need partners who can bring together advanced infrastructure, services, and expertise in energy technology to underpin large-scale AI data centre deployments at pace. “Our partnership with TeraWulf establishes a strategic blueprint for pairing on-site power, AI-enabled automation, advanced liquid cooling, and digital intelligence at a legacy industrial site. We are delivering resilient, efficient, and scalable data centre solutions at the speed and scale this AI era demands.” For more from Schneider Electric, click here.

LiquidStack expands GigaModular CDU capacity

LiquidStack, a US-based provider of liquid cooling technology for data centres, has expanded the scaling capabilities of its GigaModular CDU platform, with the system now validated for deployments of up to 14MW. The modular liquid cooling platform is designed for AI and high-density data centre environments, including infrastructure aligned with NVIDIA Vera Rubin specifications. LiquidStack says the architecture is intended to allow operators to expand cooling capacity incrementally without large-scale infrastructure redesigns. First introduced in June 2025, the GigaModular platform has now completed multi-module system integration and full-load testing. The company says the system has achieved ETL certification and has been released to manufacturing. The announcement comes amid continued growth in AI infrastructure demand and increasing pressure on data centre capacity. According to CBRE, the global weighted average data centre vacancy rate reached 6.6% during the first quarter of 2025. Modular cooling aimed at AI infrastructure growth LiquidStack says the platform has been designed to support phased AI infrastructure expansion through modular deployment and centralised controls. Key features of the GigaModular CDU platform include: • Centralised system controls intended to simplify operations and reduce infrastructure redundancy• Modular scaling designed to support phased AI deployment growth• Flexible cooling distribution architecture for changing rack densities and facility layouts• Support for high-density GPU environments and large-scale AI deployments• Global service support through Trane Technologies, LiquidStack's parent company Joe Capes, Vice President at Trane Technologies and General Manager of LiquidStack, says, “The challenge for AI infrastructure today is the ability to scale quickly and efficiently enough to keep pace with demand. "GigaModular was designed to remove the infrastructure constraints limiting AI growth through a centrally controlled, modular architecture built for system-level scalability, flexible deployment, and hyperscale AI expansion.” LiquidStack has also announced it will demonstrate the GigaModular platform at Datacloud Global Congress 2026, taking place from 2–4 June in Cannes, France. For more from LiquidStack, click here.

Bergen secures 500+ MW of orders from Liberty Energy

Bergen Engines, a Norwegian manufacturer of medium-speed gas and dual-fuel engines, has secured an order from Liberty Energy for more than 500MW of onsite power generation capacity to support AI data centre developments in the United States. The projects will use gas-powered generation systems capable of operating both independently from the grid and in parallel with utility infrastructure, supporting high-density AI computing environments. The developments are being delivered through Liberty Energy’s Liberty Power Innovations division, which focuses on distributed power infrastructure for AI-era data centres. Under the agreement, Bergen Engines will supply 45 gas generator sets, each rated at 11.2MWe, providing a combined installed capacity exceeding 500MW. The infrastructure combines Bergen Engines’ medium-speed engines with alternators from Marelli Motori and SHIELDX dynamic power stabilisation technology from Piller Power Systems. Power stability for AI workloads According to the companies, the SHIELDX platform is designed to manage rapid fluctuations in AI-related power demand by stabilising short-duration load variations. The flywheel-based system is intended to help maintain stable power delivery while reducing the need for oversized generation infrastructure. Ron Gusek, CEO of Liberty Energy, says, “AI data centres are fundamentally changing how power infrastructure can be designed and deployed, and this initial order with Bergen Engines reflects a shared commitment to providing reliable, high-efficiency power solutions to support critical data centre infrastructure growth. “Collaborating with Bergen Engines strengthens our power platform, serving as an important component of our broader integrated solution that includes LPI’s Forte modular power generation architecture and Tempo power quality system. "Together, we will be able to deliver essential power generation infrastructure to support the demanding requirements of next-generation computing.” Theo Lorentzos, Vice President Sales Americas at Bergen Engines, adds, “These environments require robust baseload generation and the ability to respond to rapid and significant load fluctuations. "By working in partnership with Liberty and integrating SHIELDX, we are delivering a solution that combines proven generation performance with the flexibility required for AI-driven demand profiles.” Dean Richards, CEO of Piller Power Systems, comments, “SHIELDX protects the generation assets from highly dynamic, sub-second load behaviour, enabling stable plant operation under extreme load fluctuations. This ensures optimal engine performance while delivering consistent, high-quality power to the data centre.” Deliveries for the projects are scheduled to begin during the second half of 2027.