Data Centre Infrastructure News & Trends

LINX, JPIX launch London-to-Tokyo network route

Internet exchange operators the London Internet Exchange (LINX) and Japan Internet Xing (JPIX) have launched direct network routes between London and Tokyo, giving European networks access to JPIX’s interconnection platform and local peering opportunities in Japan. The partnership began in 2018, initially allowing networks in Tokyo to access content in London through LINX. Under the new service, LINX members can now connect directly to JPIX Tokyo through the LINX LON1 peering platform. JPIX says it will manage the connection and ordering process through a carrier partner, removing the need for customers to arrange separate local connectivity. Mike Hellers, Product Development Manager at LINX, comments, “It’s great to see this partnership flourish. We have a handful of networks now peering in London from Tokyo with quite high bandwidths. We are pleased to now be able to work with JPIX to provide the reverse route from London to Tokyo.” European networks gain Tokyo peering access The new route is intended to support demand for Japanese digital content from European networks, including content associated with anime, manga, and gaming. An introductory service for LINX members provides 1Gbps connectivity to JPIX for £400 per month, including transport bandwidth between LINX LON1 and JPIX Tokyo. Other bandwidth options up to 5Gbps are also available. Tetsuya Hamada, Senior Executive Expert at JPIX, says, “This new route is a wider project we are working on to increase the awareness of local peering in Japan for networks in Europe.” Tokyo is a major interconnection location, with more than 20 international subsea cable systems connecting Japan with North America, Southeast Asia, and Australia. The service expansion forms part of LINX IX Connect, a programme covering its collaborative work and network routes with other internet exchange points (IXPs), including JPIX, Namex in Rome, and NYIIX in New York. For more from LINX, click here.

FlexSysAI launches platform to manage DC energy demand

Australian energy technology company FlexSysAI has launched a platform designed to make data centre power demand more flexible by shifting GPU-based AI workloads between locations and times without affecting service. The platform connects electricity market and grid conditions with AI workload balancing, allowing compute to be shifted to locations where electricity is cheaper and more abundant. Non-critical workloads can also be reduced when the grid is under strain and electricity prices are high. FlexSysAI says the approach could help data centre operators connect to the grid sooner, reduce electricity costs, access renewable power when it is available, and receive payments for supporting grid operations. The company was founded by energy specialists, technologists, and entrepreneurs including Victor Feoktistov (pictured above), Angelo Perera, and Sean Senvirtne. The team has developed its demand response technology over several years and secured a retail licence providing customers with direct access to energy markets. Australia provides test market for flexible computing FlexSysAI is initially targeting Australia, where the electricity grid, growing data centre demand, and high penetration of renewable generation provide a testing environment for the platform. The company says it is exploring adoption with multiple data centres in Australia. It is backed by EnergyLab and Sean Senvirtne, and is part of NVIDIA’s Inception programme. The launch comes as governments consider approaches to managing increasing data centre electricity demand. The Australian Energy Market Commission recently recommended a national framework that includes operational flexibility as one potential approach to managing demand. FlexSysAI co-founder Victor Feoktistov says, “Data centres need a strategy for the power crunch that is holding back the sector. Physical grid constraints are beginning to bite and time to power is already a major constraint for operators. "These pressures are likely to worsen over the next two to three years, while regulatory pressure is moving even faster as more jurisdictions look to require flexibility from large energy users. “Operators stay in control, see a live price for flexibility, and choose when to shift workloads while connecting to the grid sooner and getting ahead of mandatory requirements that continue to emerge across key markets.” Angelo Perera, Chief Technology Officer at FlexSysAI, adds, “We are starting in Australia’s National Electricity Market as it has one of the world’s most challenging grids, and its structure and volatility strongly reward smarter, more flexible demand. "This makes Australia an ideal market to prove the technology before deploying it globally as we provide a long-term solution to the accelerating electricity demand problem.” FlexSysAI has formally stated it intends to expand into international markets as data centre electricity demand increases.

AI infrastructure: Planning for long-term growth

This article from Antonio Castano, Global Market Development Director at AFL, highlights how a combination of technologies, architecture, and deployment models is driving the next generation of physical infrastructure. In today’s landscape, several infrastructure approaches are becoming increasingly relevant: • Neoclouds provide specialised compute capacity • Brownfield deployments accelerate AI implementation within existing facilities • DCI extends AI infrastructure across multiple facilities • CPO brings optical interfaces closer to compute and networking components • VSFF supports higher-density optical connectivity Each approach addresses a different requirement. However, the consistent, underlying planning challenge remains that AI deployments can change significantly between hardware generations, whilst optical cabling, pathways, and connectivity infrastructure typically remain in service for much longer periods. This means physical infrastructure needs sufficient capacity and flexibility to accommodate technologies that may not yet be widely deployed. Designing around a single AI architecture can limit future options, particularly where increasing fibre density, distributed connectivity, or new optical interfaces require changes to the physical layer. The long-term objective is to establish an infrastructure foundation that can support successive generations of AI systems without requiring extensive physical redesign. AI is not scaling through a single infrastructure model. As neoclouds, brownfield deployments, DCI, CPO, and VSFF connectivity reshape the data centre landscape, the physical layer must be designed as a flexible foundation that can adapt across multiple generations of AI architectures. Approaches to the infrastructure of tomorrow Neoclouds provide dedicated access to accelerated computing, allowing organisations to scale AI capacity without building equivalent facilities. This model can shorten deployment timelines whilst increasing demand for high-density power, cooling, networking, and optical connectivity. For infrastructure planners, the key consideration is ensuring supporting physical infrastructure can accommodate rapid changes in compute requirements. Brownfield deployments can accelerate AI capacity by reusing existing power, cooling, pathways, and facility space. However, infrastructure designed for conventional workloads may not accommodate the fibre density required by modern AI systems. For example, an NVIDIA NVL72 rack can require up to 1,152 fibre connections. Retrofitting requires careful planning for capacity and future upgrades. Data centre interconnect (DCI) allows AI environments to operate across multiple buildings, campuses, or locations. This approach can provide greater flexibility when capacity, power, or resilience requirements exceed what one facility can support. However, longer connections introduce additional considerations around latency, optical performance, power, and network architecture that must be addressed during infrastructure planning. Co-packaged optics (CPO) places optical interfaces closer to compute and networking components, reducing electrical transmission distances within systems. The architecture can support higher bandwidth while changing how fibre connectivity is presented around equipment. Physical infrastructure, therefore, needs sufficient flexibility in cable routing, fibre management, and connectivity capacity to accommodate evolving optical architectures. Very small form factor (VSFF) connectivity enables more optical connections within limited rack and panel space. Higher connection density becomes increasingly important as AI systems require greater numbers of fibres for high-speed networking. The benefit depends on adequate pathway capacity, patching space, and cable management, making VSFF part of a wider physical infrastructure strategy. Long-term AI growth: Building flexible foundations AI infrastructure will continue to combine different deployment models, facilities, optical technologies, and connectivity architectures. Because physical infrastructure remains in service longer than compute and networking hardware, capacity and flexibility are critical. A modular optical foundation allows operators to accommodate future AI requirements whilst protecting existing infrastructure investments. With AI infrastructure demands scaling fast, bringing new considerations that challenge traditional data centre design, AFL’s ‘AI Infrastructure’ whitepaper series, including Architecting AI at Scale and Building AI Training Clusters at 16K Accelerators, examines these requirements in greater detail. For a deeper dive, read AFL’s blog, What Does Sustained AI Growth Mean for Data Center Fiber Infrastructure? For more from AFL, click here.

euNetworks sets new sustainability loan targets

euNetworks, a European bandwidth infrastructure company, has introduced two environmental performance targets through its Sustainability-Linked Loan (SLL), linking sustainability measures to the design and development of new network infrastructure. The revised framework introduces Network Development Impact by Design Plans for major network projects, alongside a target for continuous improvement in the company’s GRESB infrastructure benchmark score. euNetworks first established its €760 million (£650 million) SLL in 2021 to support the expansion of its fibre network across Europe. The facility was then refinanced and expanded to €1.26 billion (£1 billion) in 2024. The Impact by Design Plans will now apply to major projects requiring significant new network construction. These projects account for a large proportion of euNetworks’ annual capital investment and approximately two thirds of its current greenhouse gas emissions. The plans will assess lower-carbon materials, construction techniques, and supplier options during the design stage, before project specifications are finalised. The approach is intended to incorporate environmental considerations into commercial and engineering decisions alongside cost, delivery times, and customer requirements. New targets added to €1.26bn loan The GRESB target will measure continuous improvement against an infrastructure-focused benchmark covering governance, environmental management, and operational performance. Marisa Trisolino, CEO of euNetworks, says, “Our new SLL targets mark an important step in euNetworks’ commitment to growing our business sustainably, focusing our efforts on the areas where we can deliver the greatest impact. “The introduction of our NetDev Impact by Design Plans represents a significant evolution in how we approach major network development projects, embedding sustainability considerations from the very beginning of the design and planning process.” The targets complement euNetworks’ existing sustainability commitments, including its validated Science Based Targets, net zero by 2040 commitment, supplier engagement programme, and carbon measurement tools. For more from euNetworks, click here.

Echelon, Trinovium to develop liquid cooling technology

Echelon Data Centres, a developer and operator of hyperscale data centres, has entered into a collaboration with clinical diagnostics company Trinovium, a subsidiary of Trinity Biotech, to develop liquid cooling technologies for AI and high-density computing environments. The collaboration will focus on cooling requirements created by increasing computing and power densities in data centres, including the management and monitoring of coolant used in liquid cooling systems. As operating temperatures and demands increase, factors including corrosion, particulate contamination, fluid degradation, and microbial growth can affect the reliability of cooling systems. The collaboration combines Echelon's experience in designing, developing, and operating hyperscale data centre infrastructure with Trinovium's expertise in high-purity fluid manufacturing and analytical technologies. Under the agreement, the companies will develop and refine liquid cooling technologies based on the requirements of Echelon's hyperscale and AI infrastructure portfolio, which currently includes more than 700MW of capacity in development and more than 1.4GW of secured capacity across Ireland, the UK, Italy, and other markets. Trinovium was established to apply Trinity Biotech's experience in healthcare-grade fluid manufacturing to AI infrastructure. Its initial direct-to-chip cooling formulation is based on high-purity aqueous chemistry, corrosion inhibition, and coolant system protection, alongside consistency and traceability. Monitoring liquid cooling systems Alongside its cooling fluid, Trinovium is developing a fluid health and system intelligence platform using analytical technologies from Trinity Biotech, including connected electrochemical sensing and mass spectrometry. The platform is intended to monitor corrosion and scaling, particulate contamination, and microbial growth and biofilm formation, providing information about the condition of liquid cooling systems. Development is expected to begin shortly, with initial work covering direct-to-chip cooling fluids, thermal management systems, and modular liquid cooling technologies for AI and high-performance computing. Niall Molloy, CEO of Echelon Data Centres, comments, “AI is changing the infrastructure requirements of data centres. As more computing power is concentrated into smaller spaces, managing the heat it generates becomes an increasingly important engineering challenge. “Liquid cooling will be an important part of meeting that challenge, but it is not simply about moving heat more efficiently; the quality, stability, and monitoring of the fluids within those systems will also be important to their long-term reliability.” John Gillard, CEO of Trinity Biotech, adds, “The rapid growth of AI is creating unprecedented demand for advanced cooling technologies capable of supporting increasingly dense and power-hungry computing environments. “By combining Trinovium’s capabilities in high-performance fluid design, precision manufacturing, and advanced analytics with Echelon’s hyperscale data centre expertise, we believe we can accelerate the development of innovative liquid cooling solutions designed specifically for the next generation of AI infrastructure.” For more from Echelon, click here.

Schneider Electric launches Easy UPS 3S Pro

Global energy technology company Schneider Electric has launched the Easy UPS 3S Pro, a three-phase uninterruptible power supply (UPS) designed for critical applications in small and medium-sized data centres, as well as for telecommunications, commercial buildings, healthcare facilities, manufacturing, and transport. Available in 10–40kVA capacities, the UPS supports internal and external battery configurations and is now available across IEC markets, including Europe, the Middle East, and Africa (EMEA). The system provides efficiency of more than 96% in double-conversion mode and up to 99% in ECO mode. It has a compact form factor and integrated breakers intended to simplify installation and commissioning. Thierry Chamayou, Vice President, Cloud & Service Providers, EMEA at Schneider Electric, says, “Today’s businesses require resilient and efficient power protection solutions that are simple to deploy and manage. “Easy UPS 3S Pro has been designed to make business continuity easy by combining reliable power protection, simplified installation, and integrated monitoring capabilities in a highly cost-effective package for customers and channel partners alike.” UPS designed for simplified maintenance The Easy UPS 3S Pro includes an Easy Loop test function, allowing UPS performance to be verified without a load bank. Front, side, and rear access is provided for maintenance. Other features include an enhanced human-machine interface, a wide operating temperature range, embedded dust filters, and conformal coating. The UPS can be operated in parallel for capacity expansion or N+1 redundancy, whilst shared battery configurations can be used across multiple units. An embedded Network Management Card enables remote monitoring and management through Schneider Electric's EcoStruxure IT software, while BACnet support allows integration with building management systems. The Secure Network Management Card is certified to IEC 62443-4-2, providing cybersecurity controls for connected infrastructure. For more from Schneider Electric, click here.

Lightpath expands managed bandwidth into Atlanta

Lightpath, a New York-based fibre network and connectivity provider, has entered the Greater Atlanta market, offering managed bandwidth services over its fibre network to data centre and enterprise customers. The company operates 12,100 route miles (19,473 kilometres) of fibre networks across 10 US metropolitan markets and is now providing Wavelength and IP Transit services in Atlanta at speeds from 10Gbps to 800Gbps per wavelength. Aggregate system capacity can scale into multi-terabit configurations, supporting applications including AI workloads, cloud interconnection, and large-scale data transfers. Atlanta has become a growing data centre market, with enterprises and hyperscalers requiring high-capacity optical transport for connectivity between data centres, cloud platforms, and other infrastructure. Lightpath operates a long-haul network hub in Atlanta connected to New York, Philadelphia, Ashburn, Chicago, Dallas, and Miami. Its Atlanta network provides connectivity to 30 data centre campuses using underground fibre infrastructure. Atlanta connects to Lightpath's national network Chris Morley, CEO of Lightpath, says, “Atlanta is one of the fastest-growing data centre markets in the country, and demand for high-capacity managed bandwidth continues to accelerate. “Entering this market connects this growing ecosystem to our national backbone, giving customers a direct path to the intensive capacity that AI and cloud workloads require.” Tim Haverkate, Chief Commercial Officer at Lightpath, adds, “Our customers need optical transport that scales with their workloads. By pairing our long-haul backbone with high-capacity managed Wavelength and IP Transit services, we give Atlanta-area customers a single, trusted path to the bandwidth their businesses demand.” For more from Lightpath, click here.

Rubicon surpasses 1GW battery storage milestone

Rubicon Professional Services (RPS), a US design and construction firm for critical facilities, says it has surpassed 1GW of installed battery energy storage capacity across projects in the United States, as demand for data centre power infrastructure grows alongside AI. The company says the capacity is distributed across projects supporting data centres, telecommunications networks, healthcare facilities, manufacturers, educational institutions, housing developments, retail operations, and utilities. Battery energy storage can provide additional power resilience, help manage electricity costs, and reduce demand on the electrical grid during periods of peak consumption. For data centres and other critical facilities, storage can also support continuity of power during disruptions. RPS says the 1GW milestone represents battery storage capacity comparable with the output of a large utility-scale power plant. William Pirrone, Founding Principal at RPS, comments, "Surpassing one gigawatt of installed battery energy storage capacity represents more than a company milestone; it reflects years of helping customers solve increasingly complex power challenges. "Today, and into the foreseeable future, the rapid growth of AI compute will place unprecedented demand on our nation's electrical grid. Organisations need experienced partners who can deliver resilient, scalable power alternatives, efficiently and on schedule." Battery storage for rising power demand The company says its battery energy storage projects are being deployed across a range of sectors as electricity demand increases. RPS has worked on energy infrastructure projects across the US since 2019, including projects supporting data centres and other critical facilities. The company has 20 years of experience delivering electrical infrastructure projects and says around 95% of its business comes from repeat customers. It also says the growth of AI computing is increasing demand for electrical infrastructure and is contributing to greater interest in battery energy storage as part of power planning for data centres. The 1GW milestone covers installed battery energy storage capacity across RPS projects nationwide.

Enlightra advances AI interconnects with Yokogawa

Enlightra, a Swiss photonics company, is using optical spectrum analysers from Yokogawa, a Japanese manufacturer of industrial measurement, control, and automation systems, to develop multi-channel laser sources for optical interconnects in next-generation AI data centres. The company is developing compact laser sources based on optical frequency comb technology. Its microcomb platform generates multiple precisely spaced optical wavelengths from a single laser, providing an alternative to using separate lasers for each optical communication channel. The systems are designed to generate between eight and 32 comb lines, with channel spacing ranging from 100 to 800GHz and power per line exceeding 7dBm across O-band datacom and C-band telecom wavelengths. Enlightra needs to measure the number of channels, line spacing, power per line, channel-to-channel power variation, and optical noise floor during development. The company characterises thousands of photonic chips during development, making measurement speed and repeatability important. It uses Yokogawa optical spectrum analysers to capture high-resolution measurements across a broad wavelength range. The analysers provide a resolution of 0.02nm and a close-in dynamic range of more than 70dB, allowing engineers to distinguish individual comb lines and measure the noise floor between channels. The measurements are used to assess line spacing, power levels, spectral flatness, and optical signal-to-noise ratio across O-band and C-band systems. Optical measurements for development Fast sweep performance allows Enlightra to maintain measurement throughput when characterising large numbers of photonic chips. The instruments also provide broad wavelength coverage for development and validation work. Measurement quality is also important when Enlightra demonstrates its technology to potential customers at exhibitions and other live demonstrations. The company says portable field instruments previously used for demonstrations did not always reproduce the resolution and low-noise performance available in its laboratory. Limited resolution could affect the displayed shape and peak power of individual comb lines, while making the noise floor more difficult to distinguish. Using Yokogawa optical spectrum analysers allows Enlightra to reproduce the measurement quality of its laboratory equipment during demonstrations. Hanae Zegmout, Lead Photonic Designer at Enlightra, says, “In the frequency comb space, Yokogawa is the gold standard. Choosing its spectrum analysers ensures that our data is immediately trusted and recognised. “Our experience with the two Yokogawa systems in our lab shows that these analysers are robust, hold their calibration well, and offer a seamless user experience.” The instruments are used across Enlightra's research and development work, as well as demonstrations of its comb-laser technology for optical interconnects. To read the full case study, you can click here. Yokogawa is also due to exhibit its optical test and measurement equipment alongside frequency-comb laser technology at ECOC 2026 in Málaga, Spain, from 21–23 September 2026.

Russelectric details its transfer switches for data centres

US power control manufacturer Russelectric, a Siemens business, has outlined the features of its automatic transfer switches for data centres and other critical facilities. The transfer switches are available in 480VAC configurations with three-cycle and 30-cycle UL-tested closing and withstand ratings under UL 1008. The 30-cycle rating can support selective coordination with overcurrent protection devices, including systems with extended trip delays. This can be used to meet National Electrical Code (NEC) requirements for emergency and legally required standby systems. The switches use preloaded springs and an over-centre mechanism driven by an electric operator to open and close the power contacts. During an open-transition transfer, the contact mechanism remains locked until the over-centre position is reached. The preloaded springs then open the closed contacts and close the open contacts rapidly, with a momentary break between the two operations. The quick-break function provides rapid arc interruption at maximum voltage and current, reducing contact erosion. The design is intended to withstand the mechanical and thermal demands associated with fault durations of up to 30 cycles. Transfer switch configurations Control components and wiring can be replaced without removing the transfer switch from its enclosure. The front-connected wiring uses flame-retardant, 600V SIS-rated switchboard wiring, with connections identified using sleeve-type markers that are visible from the front of the cabinet. Open-transition versions are available for basic transfer applications and loads with high inductive characteristics. They use a quick-break, quick-make transfer mechanism, positive mechanical interlocking, and electrical operators. The switches are available in three- and four-pole configurations, with front-accessible wiring. The 30-cycle withstand rating allows the switches to accommodate a range of short-time overcurrent protection settings, which can simplify coordination studies in high-density electrical systems. Closed-transition versions provide closed-transition retransfer to the preferred power source for testing without interrupting power. If the preferred source is lost, the switches provide high-speed quick-break, quick-make open-transition transfer. These versions are available in two-, three-, and four-pole configurations, with front-accessible wiring. The 30-cycle rating is designed to maintain the transfer switch's operating capability following exposure to fault conditions. The transfer switches can coordinate with a range of circuit breakers, including devices with extended trip delays. This provides flexibility when electrical systems are modified or expanded and can help reduce the potential for disruption to servers, cooling systems, and other critical IT infrastructure. For more from Russelectric, click here.