News

Zumtobel upgrades lighting at London data centre

Zumtobel, an Austrian company specialising in professional indoor and outdoor lighting, has completed a lighting upgrade at Global Switch’s London East data centre campus in Docklands, supporting the site’s ongoing refurbishment programme for AI and high-performance computing (HPC) workloads. The project covered multiple floors across the facility, including data halls, plant areas, offices, and a liquid cooling demonstration suite. The refurbishment programme is focused on improving flexibility, operational resilience, and energy efficiency as demand for AI-ready infrastructure continues to grow. Zumtobel worked alongside consultants including Hilson Moran, Burns & McDonnell, and AFK Studios, while Datalec Precision Installations carried out installation works. Lighting designed for AI-ready infrastructure The lighting installation was designed to improve visibility within the high-density data halls while supporting energy efficiency and long-term operational requirements. Zumtobel deployed its TECTON continuous-row lighting system across the halls, using split-lens optics to improve vertical illuminance at rack level for maintenance and operational tasks. Emergency lighting was integrated with the eBOX monitoring platform, providing automated testing and reporting functions designed for mission-critical environments. Plant areas, offices, and shared spaces were fitted with AMPHIBIA luminaires, selected for durability in technical environments, while the LITECOM lighting management platform enables centralised monitoring and control. Future refurbishment phases on levels eight and nine are expected to include TECTON II lighting, which supports faster installation through a modular 'plug-and-play' design. Ken Knight, Head of Data Centres - UK & Ireland at Zumtobel Group, comments, “Data centre environments place very specific demands on lighting, from vertical illuminance at rack level to reliability and energy efficiency. "Our role was to translate those requirements into a scalable solution that could be implemented across multiple floors while supporting Global Switch’s ongoing expansion and innovation strategy.” Matt Perrier Flint, Director - UK & Ireland at DPI, adds, “Delivering a project of this scale required close coordination between all parties. The modular design of the Zumtobel lighting system simplified installation and helped maintain programme certainty, while the collaborative approach ensured that technical requirements were clearly understood at every stage.” Lighting a liquid cooling demonstration suite Level 10 of the facility includes a liquid cooling demonstration suite designed by AFK Studios, showcasing technologies intended to support higher-density AI and HPC deployments. The lighting scheme was developed to support visibility, operational safety, and flexibility within the technical demonstration environment. According to Global Switch, the upgraded lighting infrastructure supports safer rack maintenance, lower energy consumption through LED technology and intelligent controls, simplified future upgrades, and improved emergency lighting monitoring. Derek Allen, Group Operations Director at Global Switch, notes, “Across our global portfolio, operational resilience and flexibility are fundamental. The lighting strategy implemented at our London data centre supports safe, efficient operations while giving us the adaptability required to meet evolving customer demands. "It forms part of the wider infrastructure platform that enables us to support increasingly complex AI and high-performance computing deployments.” Emily Clark, Global Switch, explains, “As our London data centre continues to evolve to support the demands of the most powerful AI and high-performance workloads, it was important that the supporting infrastructure could match that pace of innovation. "The lighting solution delivered by Zumtobel provides the performance, flexibility, and reliability we require across both operational data halls and demonstration spaces.” For more from Zumtobel, click here.

'DC construction enters a new era of delivery pressure'

In this exclusive article for DCNN, Dave Wagner, VP of Product Marketing at Newforma, examines how AI-driven demand, compressed timelines, and constantly evolving designs are forcing construction teams to rethink how data centre projects are coordinated and delivered: Why construction teams are rewriting the playbook The data centre boom has pushed construction into unfamiliar territory. Demand keeps climbing, driven by cloud computing, AI workloads, and real-time digital services. Analysts expect the global data centre market to pass $500 billion (£369 billion) within the decade. That growth sounds like opportunity; on the ground, it feels like pressure. Project teams face a new reality: Designs shift mid-build, stakeholders span continents, precision requirements leave no room for error, timelines shrink, and the old workflows do not hold up under these conditions. The solution isn’t just to work harder; it’s to work differently. The golden thread is under strain The “golden thread” promises a clear, traceable record of decisions from design through to delivery. In data centre projects, that thread gets pulled in every direction. Designs evolve while construction is already underway and a change in server density drives new cooling requirements. That then triggers updates across mechanical and electrical systems, while documentation must reflect those changes in real time or the thread breaks. Carl Veillette, Chief Product Officer at Newforma, sees this first hand, stating, “On data centre projects, the golden thread is not a static record; it is a live system. If it falls out of sync with reality, the risk compounds fast.” When information lags, teams build off outdated assumptions. That leads to rework, delays, and finger-pointing. Maintaining continuity of information is no longer a compliance exercise, but a delivery requirement. Design does not sit still Traditional construction relies on a stable design phase. Data centres ignore that sequence as technology advances too quickly. A facility planned around one generation of hardware often needs to support the next before completion. GPU-heavy AI workloads increase power density while liquid cooling replaces air in certain zones and redundancy strategies evolve. Each shift forces coordination across disciplines:• Electrical systems must handle higher loads.• Cooling infrastructure must adapt to new methods.• Structural layouts must support revised equipment footprints. These are not minor tweaks; they affect core systems. Carl puts it plainly, “You are designing for a future state that keeps changing. The teams that succeed are the ones that accept that volatility and build processes around it.” That means parallel workflows. Design, coordination, and construction happen at the same time, whilst decisions move faster, often with incomplete data. Teams need immediate visibility into the latest information to stay aligned. Precision, security, and uptime raise the stakes Data centres operate under strict conditions. Downtime is not tolerated and systems must perform on day one. This drives extreme precision: • Redundant power systems must function without failure.• Cooling must maintain exact environmental conditions.• Security measures must meet strict standards.• System integration must be flawless. At the same time, security concerns limit information access. Teams must share data widely enough to stay aligned while also restricting sensitive details. The margin for error disappears. According to Uptime Institute, over 60% of data centre outages cost more than $100,000 (£74,000), with a growing share exceeding $1 million (£739,000). That risk shapes every decision. Teams cannot afford mistakes caused by poor coordination or outdated information. Speed to market is the new benchmark The race to bring capacity online has compressed schedules across the industry. Hyperscale operators push for faster delivery to meet demand, and delays translate into lost revenue and competitive disadvantage. This pressure reshapes project timelines: • Design cycles shorten.• Construction phases overlap.• Procurement accelerates.• Commissioning windows tighten. There is no buffer for inefficiency. Rework becomes expensive and miscommunication becomes costly. Carl highlights the impact, noting, “Speed to market is not a goal anymore; it is the baseline. The only way to hit it is to remove friction from how teams share and act on information.” A shift towards structured collaboration The common thread across these challenges is information flow. Projects succeed when the right data reaches the right people at the right time. That requires a shift in how teams manage project information: • Centralised access to current documents and models• Clear tracking of RFIs, submittals, and decisions• Real-time updates across all stakeholders• A complete audit trail for accountability and risk management This is where platforms like Newforma play a role. They support the golden thread by keeping project information connected, traceable, and accessible. The impact shows up in reduced risk and faster delivery. Teams spend less time searching for information and more time acting on it. Coordination improves and errors drop, whilst projects move forward with fewer disruptions. The new standard Data centre construction has set a new standard for the industry. It demands speed without sacrificing precision. It requires flexibility without losing control. It depends on collaboration at a scale most projects never reach. These conditions will not ease, as demand will keep rising and technology will keep evolving. The teams that adapt their workflows to this reality will keep pace. Those that do not will fall behind. The playbook has already changed. The only question is who is still using the old one.

Lightpath announces new long-haul US fibre route

Lightpath, a New York-based fibre network and connectivity provider, has announced plans to build a new long-haul fibre route in the US, connecting Columbus, Ohio, and Chicago, Illinois. The approximately 392-mile (630-kilometre) route will include 327 miles (526 kilometres) of newly constructed underground multi-conduit fibre infrastructure spanning three US states. According to Lightpath, the project will be delivered in phases, with full end-to-end completion targeted for the end of 2028. The company says the route will connect two rapidly growing North American data centre markets and will incorporate eight LightCube Data Centers facilities, including seven new sites. Lightpath states that the infrastructure will support services including conduit access, dark fibre, colocation, high-capacity wavelengths, and connectivity services for hyperscale, carrier, and enterprise customers. New long-haul route for AI infrastructure growth Chris Morley, CEO of Lightpath, comments, “The Columbus-to-Chicago corridor reflects sustained hyperscale demand for high-capacity, long-haul fibre built to production-grade standards. “Connecting two of the world’s fastest-growing data centre markets on our own, purpose-built infrastructure positions Lightpath to support the next generation of AI and cloud workloads at scale.” According to the company, the new route builds on existing fibre infrastructure in Columbus and follows its acquisition of a 323-mile (520-kilometre) fibre system between New York and Ashburn in late 2024. The southern section of the route, connecting Columbus and South Bend, is expected to be the first phase brought online. Tim Haverkate, Chief Commercial Officer at Lightpath, suggests, “Building this corridor from the ground up - on a timeline driven by a real anchor award - demonstrates what our team is capable of delivering.” Lightpath says it is also assessing additional in-line amplifier locations along the route to support future capacity expansion. For more from Lightpath, click here.

Airsys opens global HQ campus in South Carolina

Airsys, a provider of data centre cooling systems, has today opened a new global headquarters campus in Woodruff, South Carolina, USA, to expand manufacturing and engineering capacity for AI and data centre cooling technologies. The 60-acre (24.28-hectare) site will act as the company’s global hub for high-efficiency cooling systems supporting AI, edge computing, and digital infrastructure applications. According to Airsys, the development represents a $60 million (£44 million) investment and is expected to create 215 jobs in the region. Manufacturing operations at the site are scheduled to begin during the first quarter of 2027. The company says the facility will support the production and development of air, liquid, and hybrid cooling technologies designed for high-density computing environments. Facility to support AI cooling demand Yunshui Chen, founder and CEO of Airsys, says, “Today marks a major milestone for Airsys as we establish our global headquarters here in Woodruff, South Carolina. “This investment reflects our commitment to advancing cooling innovation across the entire spectrum of mission-critical infrastructure.” The company states that the new campus has been designed to support growing demand for thermal management systems linked to AI infrastructure and large-scale data centre deployments. Airsys says the site will also support work related to Power Compute Effectiveness (PCE), a framework focused on maximising the proportion of data centre power available for IT workloads after cooling and electrical overheads are accounted for. The project team for the campus included Choate Construction as general contractor and LS3P as architect. According to the company, localising manufacturing and engineering operations in South Carolina is intended to improve supply chain resilience and support North American customers more directly. For more from Airsys, click here.

Delta opens Germany R&D centre for AI power systems

Delta Electronics, a Taiwanese manufacturer of power electronics, energy management, and smart green solutions, has opened a new research and development centre in Soest, Germany, expanding its engineering capabilities for AI data centre technologies across the EMEA region. The new facility will focus on high-efficiency power infrastructure for AI and high-performance computing (HPC) data centres, alongside electric vehicle (EV) power systems. According to Delta, the site includes 7,500m² of laboratory space and 2,500m² of office space, with capacity for up to 250 employees. The company says the centre will support development work related to 800V DC server power supplies and EV power systems, as demand for higher-density AI infrastructure increases. Facility targets LEED Gold certification Peter Ide, Managing Director of Delta Energy Systems, comments, “Soest has been a vital innovation engine for Delta in EMEA for decades, and this expansion reflects our commitment to strengthening localised R&D capabilities in Europe. “It enables closer collaboration with customers and faster development of advanced and sustainable solutions for data centre and automotive applications.” Delta currently employs around 450 staff at its Soest site, which acts as a regional hub for power and charging technologies. The company says the new building has been designed to meet LEED Gold certification standards and incorporates on-site renewable energy and energy storage systems. These include a photovoltaic installation with more than 1MW of installed capacity, battery energy storage rated at 2.5MW and 5MWh, heat pumps, and energy management systems. According to Delta, waste heat generated during product testing will also be reused to heat the building. The site additionally includes a DC fast charging station and more than 60 AC EV charging points for employees. Delta says the Soest operation continues to work with universities across Germany, the Netherlands, and Denmark to support engineering recruitment and research collaboration. For more from Delta Electronics, click here.

Kao Data acquires site in Park Royal, West London

Kao Data, a data centre developer and operator, has acquired a 4.7-acre (19,020m²) industrial site in Park Royal, West London, for the development of a new data centre facility. The site, formerly part of the Frogmore Industrial Estate, was acquired from Reassure Limited, part of Legal & General, in March 2026. It is located within the Park Royal area of West London, one of the UK’s largest data centre and cloud computing markets. According to Kao Data, the planned facility will be designed to support AI and advanced computing workloads while targeting high environmental and energy efficiency standards. Plans for the development are currently being prepared in consultation with the Old Oak and Park Royal Development Corporation (OPDC), local authorities, and community stakeholders. Data centre planned for 2029 Kao Data says the new facility is expected to support computing infrastructure used across sectors including life sciences, healthcare research, artificial intelligence, and financial services. David Bloom, founder and Executive Chair of Kao Data, comments, “Today’s data centres are the engine rooms of the digital age, but it’s vital that new developments work hand in hand with local stakeholders and are developed responsibly, with the community front of mind. “Our acquisition of this former industrial site in Park Royal demonstrates our longstanding commitments to sustainable redevelopment, and we’re excited to work closely with the OPDC to continue our expansion in West London.” The company also states that community engagement activities linked to the development will include education and local support initiatives, building on programmes already established at its Harlow, Slough, and Stockport sites. Detailed proposals for the Park Royal development are expected to be submitted during the coming months as part of the formal planning process. Kao Data says the facility is expected to be operational in 2029. For more from Kao Data, click here.

Siemon's DACs, AOCs, and Transceivers: Why validation matters

Siemon's comprehensive range of Direct Attach Cables (DACs), Active Optical Cables (AOCs), and optical transceivers delivers the high-speed interconnects modern data centres demand. Supporting speeds from 10G through to 400G and beyond, these pluggable solutions offer the flexibility and density needed to keep pace with today's AI-driven workloads and mission-critical applications. However, deploying these components without proper validation is a costly risk. Malfunctioning transceivers can cause high bit-error rates, loss of connectivity, slow network performance, and reduced signal strength - yet many discarded transceivers are in perfect working condition, presumed faulty without proper diagnosis. This is where the EXFO FTBx-88480 proves indispensable. Powered by EXFO's iOptics application, it assesses transceiver health in under three minutes, supporting OSFP, QSFP28, QSFP56, SFP28, AOC, and DAC form factors. The result: validated Siemon infrastructure, reduced waste, and lower operational costs. Networks Centre is a distributor for both EXFO and Siemon, and the company says its technical support team would be happy to advise on the right products to meet your requirements. Find out more at Networks Centre's website or get in touch via email at enquiries@networkscentre.com or by phone on +44 (0)1403 754233. For more from Networks Centre, click here.

Navigating AI’s infrastructure surge

In this exclusive interview, DCNN speaks with Lottie Westerling (pictured above), Head of Product at techoraco, about the structural pressures emerging across digital infrastructure, the industry’s shifting priorities, and the debates set to define the next phase of AI-driven growth: Power, talent, and the road ahead DCNN: AI is accelerating demand for digital infrastructure at an unprecedented rate. From your perspective, is the industry genuinely keeping pace, or are we starting to see structural gaps emerge? Lottie: The pace of growth across digital infrastructure is unlike anything the industry has experienced before. AI has accelerated demand dramatically, and we’re seeing a surge of activity from both established players and new entrants looking to capitalise on the opportunity. However, this rapid expansion is also exposing clear structural gaps - most notably around access to power. The challenge is no longer just about building capacity, but about how quickly that capacity can be energised. From the resurgence of gas and the resulting pressure on turbine supply to increasingly long grid interconnection queues, the strain on energy infrastructure is becoming more visible. In many ways, demand is now outpacing the systems that support it - permitting, power delivery, and supply chains alike. Addressing these constraints will be critical for organisations that want to remain competitive in an AI-driven landscape. DCNN: Events like Datacloud Global Congress Cannes bring together a wide cross-section of the ecosystem. What recurring themes are you hearing most often from industry leaders today? Lottie: Several themes are consistently coming to the fore in conversations with industry leaders. As already mentioned, speed to power remains a dominant concern, but it’s closely followed by a broader shift in how data centres are being designed. As density requirements increase, we’re seeing a growing focus on new architectures, with liquid cooling becoming central to future-ready design strategies. At the same time, financing continues to be a key topic - not due to a lack of capital, but because of questions around risk allocation and the long-term bankability of large-scale projects. Talent continues to dominate as another major area of concern. The rise of AI-driven “gigafactories” is intensifying demand for skilled workers, and the shortage of talent is becoming just as critical as constraints in equipment or infrastructure. Finally, community engagement is rising on the agenda. Public perception and planning friction are increasingly influencing project timelines, making it essential for the industry to communicate its value more clearly and responsibly. DCNN: The industry is often described as highly collaborative, yet also competitive. How important is collaboration in addressing some of the sector’s biggest challenges, such as energy access or skills shortages? Lottie: Collaboration is fundamental to solving the industry’s most pressing challenges. Issues such as energy access extend far beyond the data centre sector; they sit at the intersection of grid planning, regulation, power generation, and infrastructure design. As a result, meaningful progress depends on close coordination between the energy ecosystem and digital infrastructure stakeholders. The same principle applies to talent. Addressing the skills gap will require a collective approach, from developing shared training pathways to increasing visibility into career opportunities across the sector. By working together, the industry can make these pathways more accessible and attractive to a broader, more diverse workforce. DCNN: Talent continues to be a critical issue across digital infrastructure. What changes are needed to attract and retain the next generation of talent into the sector? Lottie: One of the biggest challenges is awareness. The value proposition of a career in digital infrastructure is not always well understood, particularly among younger audiences. There is a clear need to better communicate the scale, impact, and long-term opportunity that the sector offers. This means investing in more structured entry points such as graduate programmes, apprenticeships, and industry-led initiatives that make it easier for people to find and pursue careers in the space. It also involves creating clearer career pathways and showcasing the diversity of roles available, from engineering and operations through to sustainability and innovation. Ultimately, attracting the next generation will depend on making the industry more visible, more accessible, and more aligned with the priorities of emerging talent. DCNN: From the conversations you’re helping to shape across the Global Congress community, are you seeing a shift in priorities? Lottie: There is a growing sense of cautious optimism across the industry. While demand remains strong, there is an increasing focus on ensuring that growth is both resilient and sustainable over the long term. Leaders are placing greater emphasis on the fundamentals: reliability of power supply, sustainability of water usage, and alignment with evolving regulatory frameworks. There is also a stronger focus on future-proofing assets, ensuring that infrastructure built today will remain relevant as technologies continue to evolve. This suggests a shift from purely rapid expansion towards a more balanced approach that prioritises durability, efficiency, and long-term viability. DCNN: Looking ahead, what topics or debates do you think will define the next 12–24 months in the data centre and digital infrastructure space? Lottie: Over the next 12 to 24 months, several key debates are likely to shape the direction of the industry. At the forefront is how to meet the enormous power requirements of AI at scale. This includes discussions around alternative energy pathways, the role of nuclear, and the viability of behind-the-meter solutions. Risk allocation will also be a central issue, particularly in how responsibility is distributed between investors, operators, and tenants in increasingly complex projects. At the same time, more forward-looking topics are beginning to gain traction. The potential for data centres in space, while still nascent, is generating discussion, as is the longer-term impact of quantum computing on infrastructure requirements. Together, these conversations reflect an industry that is not only responding to immediate pressures, but also actively shaping its future trajectory.

RETN expands Balkans fibre backbone

RETN, an independent global network service provider, has launched a new backbone route connecting Drobeta, Bucharest, Iași, and Chișinău, creating an additional connectivity path across Romania and into Moldova. Delivered as a single, end-to-end backbone route, the expansion provides an alternative physical connectivity option to existing regional IP transit corridors in Eastern Europe. The new route forms part of RETN’s wider strategy to expand its optical network infrastructure across Central and Eastern Europe. By linking Romania and Moldova with the company’s existing Balkans corridor between Budapest, Timișoara, and Sofia, the project introduces additional geographical route diversity across the region. The backbone also enables alternative routing into Ukraine via Moldova and into the Balkans through Bulgaria, offering additional routing options for regional and international traffic flows. New routing options across Eastern Europe Olena Lutsenko, Business Development Director at RETN, comments, “This project is an important step in strengthening connectivity resilience in Romania. “Bucharest and Iași are rapidly developing hubs for business, education, and technology, and demand for resilient, high-capacity infrastructure is rising fast. “By delivering a direct route from Timișoara to Bucharest and onward to Chișinău, we are enabling faster, more scalable access to the region from the Balkans, Ukraine, and Central and Eastern Europe in general - for operators, ISPs, enterprises, and international customers.” The expansion comes amid continued fibre infrastructure growth across Romania and Eastern Europe. According to the National Authority for Management and Regulation in Communications of Romania (ANCOM), Romania recorded 6.9 million fixed broadband connections in mid-2025, with 37% capable of gigabit speeds. The regulator also reported continued growth in demand for ultra-high-speed internet services and increasing fixed broadband traffic per capita. Internet adoption across Romania was estimated at approximately 94% of the population in late 2025. For more from RETN, click here.

Deep Green partners with Zendo on renewable data centres

British digital infrastructure company Deep Green has partnered with Zendo Energy, a London-based startup building an AI-powered energy management platform, to support a "new generation" of AI-ready data centres powered by renewable energy and energy management technology. The partnership begins at Deep Green’s 400kW site in Urmston, Greater Manchester, where Zendo has secured a renewable energy supply contract for the facility. Designed for high-performance computing and AI workloads, the Urmston site supports rack densities of up to 150kW. Waste heat generated by the servers is captured and reused to heat the swimming pool at Trafford Leisure Centre. According to Deep Green, the heat reuse system is expected to save the leisure centre around £80,000 annually while reducing carbon emissions. The company’s approach centres on deploying modular data centres close to facilities that can use recovered heat, including swimming pools, district heating networks, and public buildings. Renewable energy aiding modular data centre expansion Deep Green says demand for AI infrastructure continues to increase, while grid limitations and planning delays remain challenges for UK data centre developments. The company says its modular deployment model allows new capacity to be brought online more quickly than traditional developments. Through the partnership, Zendo will provide energy monitoring, forecasting, and capacity optimisation using its Energy OS platform. Renewable power for the Urmston facility, in particular, will be supplied by ENGIE. Hazel Lim, Chief Financial Officer at Deep Green, says, “Zendo has been a strong partner in shaping our power procurement strategy for our data centres. "We are excited to draw on their expertise to develop a highly efficient, cost-effective approach that maximises value for our colocation clients by fully capturing the advantages of heat reuse.” Drew Barrett, COO and co-founder of Zendo Energy, adds, “Deep Green has an ambitious vision to accelerate data centre deployments at pace, and we're proud to be the energy technology partner making sure energy is never the bottleneck. “The flexibility we've built into this contract is designed to grow alongside their trajectory, and we see this as a blueprint for what the next generation of data centres should look like: flexible, sustainable, and built for scale.” For more from Deep Green, click here.