News

Advancing the agentic world

In this article, Chinese multinational technology company Huawei tells us about this year's Huawei Connect 2026: Huawei Connect 2026 kicked off recently in Shanghai. The first keynote of the date was by David Wang, the Deputy Chair of the Board and Rotating Chair at Huawei. In his speech (Advancing the Agentic World, Building a Solid Silicon Foundation), David highlighted the work the company is doing alongside industry stakeholders to build powerful AI infrastructure, lay a solid computing foundation, and address the challenges and opportunities in the intelligent world to come. AI is sweeping the world faster than any previous technological revolution. Today, foundation model parameters are rapidly approaching 10 trillion, and are projected to exceed 100 trillion by 2030. AI agents can now work on the same task continuously, for hours on end. By 2030, they will be able to handle tasks that span months. In China alone, the average number of inference tokens consumed every day has surged to around 500 trillion, and is expected to reach quintillions (10¹⁸) by 2030. On-device AI is also advancing rapidly. On-device models for smartphones have expanded from three billion parameters in 2024 to 30 billion today, and will push toward hundreds of billions in the near future. These trends will set a much higher bar for the scale, performance, and reliability of underlying technical systems. Only by building powerful AI infrastructure can the industry lay a solid foundation for the future intelligent world. An intelligent world is approaching – and faster than ever. To lead the charge into this new world, Huawei is laser-focused on building out AI infrastructure – the silicon foundation for the future to come. In particular, Huawei's AI strategy is centred on computing power, with a focus on monetising hardware. The company is also sharpening its competitive edge through systems and architectural innovation. Centring these efforts on SuperPoDs and SuperClusters, the company aims to build a solid computing foundation and offer a new option for the world. Huawei is a major contributor to open computing ecosystems, and will continue to support native training for mainstream foundation models on its systems, as well as supporting a vast range of models and applications. For customers, Huawei provides flexible on-premises and cloud compute solutions for its customers to accelerate intelligent transformation across industries. With diverse forms of compute, including solutions for micro-, low-tier, mid-range, and massive computing power – Huawei is driving the expansion of on-device and in-vehicle AI, making intelligence truly ubiquitous. Additionally, Huawei is dedicated to building next-generation communications networks to bring readily available compute and intelligence to every person, home, and organisation. SuperPoDs gain broad consensus, with adoption growing in industries, academia, and research institutes To date, over 1,000 Atlas 900 A3 SuperPoDs have been deployed, and Atlas 950 SuperPoD is seeing large-scale commercial use. While adoption continues to grow, SuperPoDs have gained broad acceptance across industry, academia, and research institutions as a key direction for AI infrastructure. Currently, a SuperPoD is explicitly defined as a computing system in which multiple computing nodes are tightly coupled through high-speed interconnect protocols, featuring unified memory addressing across physical nodes — functioning like a single logical computer. SuperPoDs are the go-to choice for AI infrastructure buildout. Right now, 100k-NPU computing clusters have become the baseline for training SOTA models. However, traditional server architectures result in intra-cluster communications that account for over 40% of total training time, severely constraining Model FLOPs Utilisation (MFU). Simulation results from Huawei's Markov Lab show that a 100k-NPU cluster built with 4k-NPU SuperPoDs can deliver a 2.75x increase in MFU compared to a 100k-NPU cluster composed of 8-NPU servers. 11-chip UnifiedBus-powered portfolio for SuperPoDs and SuperClusters; the Atlas 960E SuperPoD – the industry's first to use NPO The Ascend series of chips is the most critical component in Huawei's 11-chip UnifiedBus-powered portfolio for SuperPoDs and SuperClusters. Development on Ascend 960 has exceeded the company's expectations, with performance doubling as planned. Ascend 960DT will be available in Q1 2027, three quarters ahead of the company's original roadmap. And the Ascend 960PR will be ready in Q3 2027, one quarter ahead of schedule. "We're evolving our Ascend chip series on a one-generation-a-year cycle," says David in his keynote. "In 2028 and 2029, we will roll out the Ascend 970 and 980 chips, respectively. Thanks to the Tau (τ) Scaling Law, not only will their compute specifications continue to double, but you can also expect to see huge improvements across the board in terms of memory bandwidth, memory capacity, interconnect bandwidth, and more." In addition to Ascend chips, Huawei has also developed a complete portfolio of chips for AI infrastructure, based on UnifiedBus, delivering key capabilities that cover computing, interconnect, storage, and management. "SuperPoDs are designed to coordinate multiple NPUs through interconnect," continues David. "We have developed a next-generation optical interconnect product based on near-packaged optics (NPO): the High-density Optical-interconnect-Node Engine (Hi-ONE). Built on Huawei's proprietary technologies, Hi-ONE has a multi-physics design for balancing optical, mechanical, electrical, electromagnetic, and thermal performance, realising a transmission capacity of 7.2 Tbit/s per single engine. "This is the industry's first NPO product ready for mass production, delivering the largest transmission capacity. It is also the industry's first NPO product with a built-in light source." This product combines high bandwidth and high reliability with low latency and low power consumption. This, coupled with UnifiedBus, will make it far easier to scale up SuperPoD interconnect systems. Recently, Huawei submitted an implementation agreement (IA) on NPO to the Optical Internetworking Forum (OIF), a standards organisation. The response from numerous industry partners has been widely positive. Huawei will continue its efforts to further refine the NPO industry ecosystem. Using Ascend 960 chips and Hi-ONE, Huawei has developed the industry's first NPO-based SuperPoDs: the Atlas 960E SuperPoDs. A single Atlas 960E SuperPoD can scale up to 4,096 NPUs, delivering 8EFLOPS of FP8 compute performance, with up to 1PB of HBM capacity. With 5,500 Hi-ONE units, this SuperPoD doesn't need the 48,000 800G optical modules that would traditionally be required to connect all the NPUs. This cuts power consumption by over 550kW, while doubling the system's fault-free operating time, achieving 99.8% system availability. Combining the upgraded TaiShan 950 SuperPoD and context memory storage to power an ultrascale cluster with one million NPUs As SOTA models scale to 10 trillion parameters, training and inference can no longer rely on a single AI server or AI SuperPoD – they require a more complex computing system. This system includes AI SuperPoDs, general-purpose SuperPoDs, and an interconnect system that features peer-to-peer interconnect and zero protocol conversion. For inference, including a petabyte-scale KV cache cluster is also a must. To meet these demands, Huawei has fully upgraded its TaiShan 950 SuperPoD. Powered by UnifiedBus all-optical networking, this new SuperPoD supports up to 4,096 nodes with a unified memory pool of up to 256TB. This setup significantly improves agent performance. For sandbox-intensive workloads, startup speeds for 100,000 sandboxes are 30 times faster than traditional servers, and sandbox density can be improved by an additional 25%. For vector search across 10 billion x 1,000-dimensional vectors, this SuperPoD delivers twice the search efficiency of traditional servers. Huawei has also launched OceanStor M900 – a UnifiedBus-powered context memory storage cluster that delivers multi-tier KV caching for agent-heavy and longer-context workloads. Designed for agentic inference, this cluster supports one-hop direct access and provides a petabyte-scale KV cache for the L3.5 layer. OceanStor M900 also uses hybrid media and an optimised retention algorithm, extending SSD read/write lifespan by 16-fold. This ensures a higher KV cache hit rate alongside long-term stability and reliability from the ground up. Combining its strengths in computing and communications, Huawei has built a brand-new agentic SuperCluster to accelerate training and inference for 10-trillion-parameter models. This SuperCluster uses UnifiedBus to consolidate multiple interconnect protocols into a single unified protocol, significantly reducing protocol conversion overhead. This delivers peer-to-peer interconnect between subsystems like Ascend SuperPoDs, Kunpeng SuperPoDs, and KV cache clusters. The SuperCluster also comes with a multi-tier, high-bandwidth, and large-capacity storage system that enables direct single-hop access for all KV cache tiers. With a two-tier, four-plane Clos architecture, the SuperCluster can interconnect up to 512,000 NPUs. When combined with a multi-rail topology, this cluster can support up to one million NPUs. One of Huawei's core strategies: Going open source and open system to build out computing ecosystems The Kunpeng ecosystem is driving digital and intelligent innovation across a wide range of industries. To date, the Kunpeng ecosystem has attracted over 4.16 million developers and more than 7,200 ecosystem partners from around the globe. The community currently supports over 560 open-source projects worldwide. openEuler has seen more than 20 million installations, securing the largest share in China's server OS market. The Ascend ecosystem has reached a new inflection point. The Compute Architecture for Neural Networks (CANN) is the foundation of the Ascend ecosystem. Today, CANN has moved to sustained, community-driven open-source development, which has brought the platform from usable to user-friendly. External CANN developers now comprise 61% of all CANN developers, outnumbering internal developers for the first time. With over 5,200 monthly active developers, the CANN community has become the most vibrant open-source community in China. What's more, over 40 models have been natively pre-trained on Ascend and CANN, making it the only proven domestic stack capable of model pre-training. Ascend now supports over 90 leading third-party open-source projects, including PyTorch, Triton, vLLM, and veRL. With strong support from the Linux Foundation, Ascend is the first official Chinese compute platform on PyTorch's website. This gives developers around the world ready access to new innovations in the Ascend ecosystem. Diverse forms of compute for ubiquitous on-device and in-vehicle AI AI is expanding faster into all kinds of devices. To deliver an unparalleled AI experience across all scenarios, Huawei will continue to strengthen capabilities in four key areas: First, Huawei will combine Kirin and Ascend chips to drive self-reliance and autonomy in on-device compute. Second, Huawei will bring together Pangu models and third-party models to make on-device intelligence better and easier to use. Third, HarmonyOS, as an Agent OS for ubiquitous intelligence, will be completely redefined from the ground up – spanning system architecture, how it operates, and interaction logic – to enable human-agent collaboration. Fourth, Huawei will keep cultivating a diverse AI ecosystem, which is the foundation for its system agent Celia to thrive. Huawei plans to build four on-device computing platforms: for AI phones, AI PCs, vehicles, and homes. Through cross-device and device-cloud compute synergy, Huawei will be able to provide distributed swarm intelligence, delivering integrated and continuous intelligent services across personal mobile, office, vehicle, and home spaces, ultimately bringing intelligence to every person and every space. Building next-generation communications that prioritise readily available compute, because without networks, all compute is siloed Next-generation communications networks are crucial for fully unleashing the value of AI compute. We are driving the upgrade to networks that, in addition to connecting people, will prioritise delivering readily available compute. These networks will be underpinned by 5G-A/6G, 10-gigabit optical networks, and multi-tier, low-latency bearer networks, delivering intelligent connectivity across data centres, the edge, and devices. Concluding his keynote, David expresses that AI "may well be the final technological revolution in human history," noting that its impact is deeper and broader, and coming faster than anyone could have ever imagined. "No single company," he says, "can build an intelligent world alone." He stresses Huawei's ongoing commitments moving forward: • Huawei will remain committed to building a solid silicon foundation to make computing power readily accessible to all.• The company will continue to open source its software, helping developers unleash their full potential.• It will continue to embrace a wide range of models and applications, unlocking value in every form. "And we will continue to work together to drive shared success, growing together with our customers and partners around the world," David concludes. "Let's work together to build a fully connected, intelligent world." Themed Advancing the Agentic World, Huawei Connect 2026 delved into AI across three dimensions: strategy, technology, and ecosystems. You can read about an in-depth look at our latest strategic initiatives, and we'll also be unveiling our all-new digital and intelligent infrastructure products, scenario-specific solutions for industries, and development tools. The event ran from 17 to 19 September at the Shanghai World Expo Exhibition & Convention Center and Shanghai Expo Center. For more information, please visit Huawei Connect 2026 online by clicking here. For more from Huawei, click here.

Datacenter United signs multi-megawatt capacity agreement

Datacenter United, an independent Belgian data centre operator, has signed a binding agreement to provide multi-megawatt data centre capacity for AI workloads at DC Machelen in Belgium. The environment will use liquid cooling to support the power densities and heat loads associated with high-density AI systems. The customer's identity and commercial terms of the agreement will remain confidential. AI systems concentrate increasing amounts of computing power within a limited footprint, raising both the power required per rack and the amount of heat that must be removed. Power supply, cooling, connectivity, redundancy, and operations must, therefore, be engineered as an integrated environment. Liquid cooling removes heat closer to its source and can enable high-density configurations that are increasingly difficult to support with air cooling alone. The system at DC Machelen will be engineered around the customer's confirmed architecture and operational requirements. The agreement reflects demand for specialised AI infrastructure in Belgium, with the project moving from planning to deployment. Liquid cooling for high-density AI Friso Haringsma, CEO of Datacenter United, comments, “Demand for specialised AI infrastructure in Belgium is moving from ambition to actual deployment. "This binding agreement translates a specific customer requirement into a concrete, multi-megawatt project in Machelen. It demonstrates that a Belgian operator can support high-density AI workloads at this scale.” The liquid-cooling design is intended to support higher computing densities whilst improving cooling and energy efficiency. Datacenter United says the project also demonstrates the importance of access to power, connectivity, suitable locations, and predictable permitting for AI infrastructure development in Belgium. DC Machelen is designed for organisations with high power and capacity requirements. The site combines power, space, carrier-neutral connectivity, and the flexibility required for large-scale, high-density infrastructure. Technical implementation will proceed according to the agreed system configuration and deployment plan. Further information on the commissioning schedule is set to be provided at a later stage. For more from Datacenter United, click here.

iPronics, BSC advance optical networking for AI

iPronics, a developer of silicon photonics-based optical circuit switching, has announced a two-year technology collaboration with the Barcelona Supercomputing Center – Centro Nacional de Supercomputación (BSC-CNS) to advance programmable optical networking for GPU-based AI and high-performance computing infrastructure. The collaboration will bring iPronics's programmable optical switching platform together with BSC's research-class GPU and HPC infrastructure. The teams will investigate how dynamically reconfigurable optical connectivity can respond to the communication requirements of AI workloads. iPronics will provide its programmable optical switching platform, low-level software, and APIs, whilst BSC will develop software above the switch-management layer. The work will focus on networking approaches that align optical connectivity with AI applications. “AI infrastructure is reaching a point where scaling compute also requires rethinking the network,” says Christian Dupont, CEO of iPronics. “Networking needs to become more dynamic and more closely aligned with the behaviour of AI workloads. "Our strategic collaboration with BSC is an important milestone in bringing programmable optical networking deeper into the AI infrastructure stack.” Programmable optics for AI workloads The platform will be integrated into BSC's GPU and HPC infrastructure, allowing the teams to investigate networking approaches for AI training and inference, including large language models (LLMs) and mixture-of-experts (MoE) workloads. Antonio Peña, Leading Researcher and Group Leader at BSC, notes, “As AI and HPC systems scale, the interconnect is becoming an increasingly important part of overall system architecture. “Our collaboration with iPronics gives us the opportunity to explore how programmable optical networking and the software stack can work together to support more dynamic and efficient computing infrastructures.” The collaboration is intended to generate system-level insights that can inform the architecture of future AI infrastructure. The announcement follows iPronics's recent $125 million (£93 million) Series B funding round, co-led by Maverick Silicon and Light Street Capital, with participation from NVIDIA. The funding brings the company's total raised to $177 million (£132 million). iPronics has also opened a US office in Santa Clara, California, which it says will support its product strategy, customer deployments, and AI infrastructure partnerships. For more from iPronics, click here.

DataVita invites Greens, APRS to Chapelhall data centre

DataVita, a UK data centre and cloud services provider, has published open letters inviting Ross Greer MSP, Co-leader of the Scottish Greens, and Kat Jones, Director of Action to Protect Rural Scotland (APRS), to visit its data centre campus at Chapelhall in North Lanarkshire, Scotland. The invitation to Ross also extends to his co-leader, Gillian Mackay MSP, with both letters welcoming colleagues, including planning and technical advisers. DataVita says visits can take place at a time convenient to the recipients. The Scottish Greens are expected to use an opposition debate on data centres to renew calls for a moratorium on large-scale data centre applications across Scotland. The proposed visits would provide an opportunity to see DataVita's existing operations, discuss construction under way at Chapelhall, and ask questions about its published proposals for a fourth data centre, DV4. DataVita, which is headquartered in Lanarkshire, has operated data centres in Scotland for more than 10 years. It built, owns, and operates its DV1 facility in Chapelhall, is constructing DV3 alongside it, and operates a second facility in Glasgow. The company says its facilities use renewable energy and closed-loop cooling, with plans for new facilities to contribute to Scotland's electricity grid. DataVita recently secured around £300 million in funding from a syndicate of lenders, including the UK Government's National Wealth Fund, to expand DV1 and build DV3. The company says this will support the development of Scotland's first AI data centres within a designated AI Growth Zone. UK Government figures estimate that, if fully adopted, AI could be worth up to £47 billion a year to the UK economy over the next decade. DataVita says the wider data centre campus could also support more than 3,400 new jobs and more than £8 billion in private investment in North Lanarkshire. More immediately, the expansion of DV1 and development of DV3 are expected to create around 600 construction jobs, as well as around 100 permanent high-skilled roles once completed. Data centre plans and public discussion The Lanarkshire AI Growth Zone features a Community Investment Fund that will provide up to £543 million in direct funding to North Lanarkshire over 15 years. This includes skills and training programmes in AI, coding, and other areas of technology for local people, alongside support for local food banks and charities, delivered in partnership with the community. Danny Quinn, Managing Director of DataVita, comments, “Nearly all of us use data centres every single day whether we realise it or not - from being able to use online banking and working from home, to scientists delivering groundbreaking medical research. They are the critical infrastructure that makes many of the things we do in modern life possible, and that is only going to accelerate as technology becomes even more entwined with daily life. “For more than a decade, we’ve been supporting Scotland’s public services, businesses, and even households to take advantage of the many benefits from online services. And, although AI is very much a hot topic at the moment, there is no doubt it is here to stay, and the question for Scotland is how can we best use it to support jobs, skills, the public services we all rely on, and the way we live more generally. “DataVita fully supports an open, public debate about the right kind of controls to have in place for AI and, where necessary, government action to protect our economy, society, and environment. But, the importance of the decisions we make about AI for Scotland's future economy cannot be overstated. We need to move beyond a debate in which one side has to be entirely right and the other entirely wrong. “Where opinions differ, open and honest dialogue is the best way forward. A visit gives us the opportunity to show how our facilities operate and have a detailed conversation about our plans. We can explain the choices behind their design, hear the questions people have, and discuss them directly. “Data centres differ in what they do and how they are designed and operated. We would welcome the chance to explain how those differences apply to our facilities and proposals, and we are making these invitations public because the developments are of public interest. “People should be able to see that we are willing to explain our work, listen to different views, and have a constructive discussion. We hope the invitations will be accepted.” The letters are published in full on the DataVita website. For more from DataVita, click here.

techUK calls for 10-year digital infrastructure strategy

Trade association techUK has called on the UK Government to publish a 10-year 'Digital Infrastructure Strategy' by 2027, warning that economic growth, productivity, and national resilience depend on stronger digital infrastructure. The organisation’s latest report argues that connectivity, compute, and data infrastructure are now as fundamental to the economy as transport, energy, and housing. It says the UK’s digital technologies already contribute an estimated £286 billion in annual Gross Value Added (GVA), equivalent to around 13% of national output. The UK is home to Europe’s largest data centre market, alongside research capabilities in quantum, photonic, and neuromorphic computing, whilst full-fibre broadband and 5G Standalone deployment continue to expand. However, techUK warns that high industrial electricity prices, lengthy grid connection times, planning constraints, and comparatively weak network performance are limiting investment and threatening the UK’s competitive position. Five objectives for 2035 The report sets out five objectives for 2035: universal full-fibre and 5G Standalone coverage, with commercial 6G deployment underway; a globally competitive position in next-generation compute; electricity prices below the IEA median; governance and planning frameworks that can keep pace with technological change; and widespread digital adoption across the UK. To support these objectives, techUK recommends a coordinated programme of government and industry action. This includes establishing a cross-government digital infrastructure strategy, appointing a Minister of State with responsibility for its delivery, and creating a dedicated digital infrastructure unit within the Regulatory Innovation Office. The organisation also calls for an interim update to the UK Compute Roadmap by 2027, covering neuromorphic, photonic, and quantum computing, alongside measures to develop commercial supply chains. On energy, techUK recommends a roadmap to bring non-domestic electricity prices below the IEA median by 2035. Its proposals include moving electricity levies to general taxation by 2029 and accelerating grid connection reforms. Sophie Greaves, Associate Director for Digital Infrastructure at techUK, argues that the UK needs to treat digital infrastructure as a national economic priority. She says, “Digital infrastructure is no longer simply an enabler of economic activity; it is the foundation on which future economic growth, innovation, and national resilience will be built. The countries that invest now in connectivity, compute, and data infrastructure will be best placed to capture the economic opportunities of the next decade. “By following the roadmap we have set out for a resilient digital infrastructure by 2035, the UK can secure its position as a leader in compute, bringing significant benefits to the economy.” techUK warns that failing to address these infrastructure challenges could result in slower growth, weaker productivity, reduced investment, and greater dependence on infrastructure and technologies developed elsewhere. For more from techUK, click here.

Schneider Electric expands prefabricated power portfolio

Global energy technology company Schneider Electric has launched a range of standardised Prefabricated Power Modules and Power Skids for hyperscale, neocloud, and colocation data centres supporting AI and high-performance computing workloads. Powered by the company’s Galaxy VXL UPS technology, the prefabricated power modules provide up to 2.5MW of capacity per module. Standard configurations are available at 2MW, 2.25MW, and 2.5MW, allowing multiple units to be deployed as repeatable power blocks across larger data centre campuses. The systems are based on Schneider Electric’s reference designs with NVIDIA, including integrated power management and liquid cooling controls for the GB300 NVL72 platform. They are intended to support AI training clusters and other GPU-intensive workloads. Standardised systems target faster deployment Schneider Electric says traditional customised prefabricated power systems can require lead times of 12–14 months. Its standardised approach is intended to simplify specification, design, and manufacturing, with production possible in as little as six months. The modules combine UPS, switchgear, cooling systems, lithium-ion batteries, busway, and power distribution within a factory-tested enclosure. The power skids provide similar electrical infrastructure in an open format for indoor deployment. Tarunjeet Sarao, SVP Data Center Systems at Schneider Electric, says, “Today’s data centre operators are under immense pressure to deploy AI-ready infrastructure faster, more efficiently, and at greater scale. “By standardising our prefabricated power systems and incorporating our leading Galaxy VXL UPS technology, we are enabling hyperscale, neocloud, and colocation providers to deploy scalable, high-density power infrastructure in months rather than years.” The modules and skids are engineered, manufactured, and tested at Schneider Electric’s facility in Sant Boi de Llobregat, Barcelona. The products are available immediately in Europe, with North American availability planned later this year, followed by other international markets and APAC. For more from Schneider Electric, click here.

Aggreko achieves grid code compliance across Europe

Aggreko, a British multinational temporary power generation company, has achieved grid code compliance for its 1.5MW generator fleet across 11 European countries, allowing the company to support data centres and other large power projects facing grid capacity constraints. The certification covers the UK, Ireland, Germany, France, Italy, Spain, the Netherlands, Belgium, Poland, Romania, and Czechia. It allows Aggreko’s generators to integrate with public electricity networks for applications including additional generation capacity, peak demand management, exporting power to the grid, and grid balancing. Aggreko says the certification can also help accelerate large-scale power projects, with gas-powered generation of more than 50MW potentially deployed within months while longer-term grid infrastructure is developed. Generation supports grid capacity The company says decentralised generation can help address capacity shortfalls, support renewable energy integration, and reduce pressure on electricity networks. Aggreko has previously supplied 43MW of power to Romania’s national grid over four years, using a combined heat and power system while providing grid balancing services. Connor Docherty, Head of Product for Gas at Aggreko, suggests, “Decentralised energy will have a critical role to play during this critical period, which is where grid code compliance comes into play. “With the ability to integrate with grid infrastructure, we can support grid operators and high energy users alike through decentralised energy generation. This will help to balance the grid, make up capacity shortfalls to ensure that large power projects can go ahead, assist increased renewable integration, and avoid blackouts.” Aggreko says that its fleet also incorporates remote monitoring and control systems for equipment diagnostics and plant-level analytics. For more from Aggreko, click here.

Expert warns of extreme weather disruption to data centres

A recent warning from the UN about a "supersized" El Niño has made it clear that extreme weather events will become more common, coming as the UK braces for a stormy autumn of heavy rain and strong winds. Murray Gates, CEO at Westgate Global and a protection risk expert for data centre builds, warns that El Niño will have a significant and costly impact on data centre construction across the UK for the rest of 2026 (and into 2027) if the industry doesn’t evolve its approach to weather risk management. He argues, “The UK’s predictably ‘unpredictable’ weather pattern isn’t anything new. However, with autumn and early winter likely to be wetter and more stormy, it will become virtually impossible to deliver data centre builds on time and within budget without the right planning and infrastructure. “Due to an uncharacteristically warm summer, many data centre build projects will already be playing catch-up due to the delays incurred by extreme heat. The prospect of now dealing with heavy rainfall and high winds over the autumn months will be a cause for concern on projects taking a reactive approach to weather risk management.” Adapting as growth accelerates Two years on from the Government announcing data centres as critical national infrastructure (CNI), the expansion of data centres across the UK has increased at pace. Murray continues, “Failing to prepare for these harsher weather conditions from the outset will have a detrimental impact on the UK’s ability to meet its AI ambitions and keep pace with global competition. “Unfortunately, engineered weathering is too often seen as what you reach for ‘if the weather turns’, rather than a planned, budgeted part of the build programme. This must change if data centre builds want to stay on track. “From the outset, data centre operators, contractors, and quantity surveyors must consider how to mitigate against the challenge of building in live - and increasingly unpredictable - environments. "When it comes to weather protection during the build process, procurement decisions are often boiled down to cost. Should the weather suddenly change, this leaves projects susceptible to delays that could have been avoided had the right infrastructure been installed from the start.”

Australia could cut AI emissions through data centres

Australia could reduce global AI emissions by hosting more data centre infrastructure and using renewable energy to power computing, according to a study by economics research and advisory firm Mandala Partners (commissioned by Australian industry body Data Centres Australia). The report argues that Australia could capture AI demand that might otherwise be hosted in Southeast Asia, where more carbon-intensive energy generation is used. Hosting a 1.9GW AI compute export opportunity in Australia could produce an estimated 3.2 million tonnes of CO2 emissions, compared with 11.9 million tonnes if the same capacity were hosted in Indonesia. Mandala says Australia is well positioned to become a regional hub for lower-emission AI compute, effectively exporting renewable energy through computing services rather than physical electricity infrastructure. Data centres could deliver economic benefits Capturing the AI export market could generate up to $4.1 billion (£3 billion) in additional annual economic activity, support up to 17,120 jobs, and drive up to $52 billion (£38.5 billion) in new investment. Mandala Partner Tom McMahon says, “We know Australia is capable of producing abundant clean energy, but the challenge is how to export that energy to the world. "Exporting clean energy via compute takes advantage of existing subsea cables that are responsible for carrying data and internet traffic in and out of Australia.” Belinda Dennett, Chief Executive Officer of Data Centres Australia, also notes that investment in digital infrastructure could strengthen Australia's role in the energy transition while supporting economic growth and sovereign capability. The report forecasts that Australia's data centre industry could generate $5.6 billion (£4.1 billion) in annual economic activity and support 23,040 ongoing jobs by 2030, with construction activity contributing a further $19.8 billion (£14.6 billion) and 18,930 jobs. The full report is available to view via Data Centres Australia's website.

T5 Services creates EverOn data centre business

T5 Services has announced changes to its construction and operations businesses, with T5 Construction becoming EverOn Data Center Services (EverOn DCS), while T5 Operations is set to join Salute. Effective from 1 September, EverOn DCS will operate as an independent data centre construction company focused on mission-critical digital infrastructure. At the same time, T5 Operations has entered into a definitive agreement to join Salute, bringing the two organisations' operations businesses together. EverOn DCS has grown from $87 million (£64 million) in revenue in 2021 to $1.7 billion (£1.2 billion) in 2025, with 83% of its 2025 revenue coming from third-party construction projects. The business was originally established within T5 Data Centers and now works with hyperscale, developer, enterprise, and AI customers across North America. According to T5 Services, establishing EverOn DCS as an independent company will allow it to focus on its third-party construction activities as demand for data centre capacity increases. The company has completed more than 260 construction projects, delivering more than 12 million square feet (1.1 million square metres) of data centre space across the United States. Its activities include ground-up campus construction, data hall fit-outs, retrofits and upgrades, design-build delivery, LLE equipment procurement, commissioning oversight, QA/QC, owner advisement, and preconstruction services. EverOn also has experience with direct liquid cooling, high-density AI infrastructure, and construction in live, energised mission-critical environments. T5 Operations to join Salute Salute has announced an agreement to acquire T5 Operations, which will bring the operations teams together and aims to give customers access to a broader range of technical resources and geographic coverage, according to the companies. EverOn and Salute will continue to work together as strategic partners, with the companies stating that collaboration between construction and operations will support operational readiness and long-term facility performance. Pete Marin, CEO and Chairman of T5 Data Centers, comments, "This is a strategic evolution that strengthens both businesses. Our construction business has grown significantly beyond its original mandate, earning the trust of third-party developers and hyperscalers, and establishing EverOn gives [the] organisation the independence and focus to continue scaling. "At the same time, bringing our T5 Operations team together with Salute creates an even stronger platform dedicated to helping customers maximise uptime and operational excellence. We’re incredibly proud of what both teams have built and excited about what comes next.”