News

Industry analysts urge data trust over AI hype

As organisations continue to increasingly embrace AI to unlock new operations, industry analysts - at the Gartner Data & Analytics Summit in Sydney - urged a critical reminder that without trustworthy data, even the most advanced AI systems can lead businesses astray. Amid rising interest in generative AI and autonomous agents, business leaders are being reminded that flashy AI capabilities are meaningless if built on unreliable data. According to information technology research and advisory company Gartner's 2024 survey, data quality and availability remain the biggest barriers to effective AI implementation. If the foundation is flawed, so is the intelligence built on top of it. While achieving perfect data governance is an admirable goal, it's often impractical in fast-moving business environments. Instead, analysts recommend implementing "trust models" that assess the reliability of data based on its origin, lineage, and level of curation. These models enable more nuanced, risk-aware decision-making and can prevent the misuse of data without stalling innovation. Richard Bovey, Chief for Data at AND Digital, comments, "Trust in data isn't just a technical challenge, it's deeply cultural and organisational. While advanced tools and trust models can help address the reliability of data, true confidence in data quality comes from clear ownership, clear practices, and company-wide commitment to transparency. "Too often, organisations are rushing into AI initiatives without fixing the basics. According to our research, 56% of businesses are implementing AI despite knowing their data may not be accurate in order to prevent from falling behind their competitors. "Businesses must consider taking a data and AI approach to their technical operations to build trust, cross-functional collaboration, and ongoing education. Only then can AI initiatives truly succeed." At the summit, autonomy was a central theme. AI systems may act independently in low-risk or time-sensitive situations, but full autonomy still raises concerns as, while users accept AI advice, they're still adjusting to autonomous AI decision-making. Stuart Harvey, CEO of Datactics, argues, "One of the biggest misconceptions we see is the belief that AI performance is purely a function of the model itself, when in reality, it all starts with data. Without well-governed, high-quality data, even the most sophisticated AI systems will produce inconsistent or misleading results. "Organisations often underestimate the foundational role of data management, but these aren't back-office tasks, they're strategic enablers of trustworthy AI and those businesses that rush into AI without addressing fragmented or unverified data sources put themselves at significant risk. Strong data foundations aren't just nice to have in today's technical landscape, they're essential for reliable, ethical, and scalable AI adoption." Gartner predicts that by 2027, 20% of business processes will be fully managed by autonomous analytics and these "perceptive" systems will move beyond dashboards, offering proactive, embedded insights. The company also believes that by 2030, AI agents will replace 30% of SaaS interfaces, turning apps into intelligent data platforms. To thrive, data leaders should thus prioritise trust, influence, and organisational impact, or risk being sidelined. For more from Gartner, click here.

Spirent and Juniper Networks collaborate for first public UET test

Spirent Communications, a provider of test and assurance solutions for next-generation devices and networks, and Juniper Networks, a multinational corporation developing and marketing networking products, have successfully completed the first live public demonstration of Ultra Ethernet Transport (UET), marking a milestone for the evolving Ultra Ethernet Consortium (UEC) ecosystem. The demonstration took place at Interop25 Tokyo, where Spirent, Juniper, and Spirent’s long-standing partner in Japan, TOYO Corporation, jointly generated and forwarded UET traffic in a live network environment, utilising the Spirent B3 800G Appliance and the Juniper QFX 5240-64OD Switch. The test topology also included RoCEv2 traffic, demonstrating coexistence and interoperability, and earned the companies an Interop ShowNet Special Prize. The Juniper switch with 800G interfaces successfully recognised and forwarded all traffic types, validating its readiness for UET-based deployments. As hyperscalers continue to prioritise open, interoperable solutions for AI/ML back-end networks, Spirent believes the significance of the UEC initiative and its approach to delivering AI and HPC performance is becoming ever more apparent. With the consortium approaching its first official specification release, Spirent and Juniper argue that they are playing a pivotal role in driving progress through real-world validation, helping to strengthen and refine the evolving ecosystem. “We are honoured to have been recognised alongside Spirent with this prestigious award at Interop Tokyo,” says Praful Lalchandani, VP of Products, Data Center Platforms and AI Solutions, Juniper Networks. “This award underscores the strength of our partnership and the innovation we’re driving together in high-performance, AI-driven data centre networking. Juniper’s data centre solution, a cornerstone of Juniper’s AI-Native Networking Platform, continues to lead with operational simplicity, performance, and open architecture. “Validating performance and interoperability with emerging transport standards like UET is essential for ensuring next-generation infrastructure can meet the evolving demands of AI-driven networking.” “Dramatic growth in AI applications and workloads continues to create challenges for the industry, and this public demonstration illustrates our dedication to helping industry partners to meet these challenges,” comments Masatoshi Nishihara, General Manager, Spirent Japan. “Spirent is committed to supporting emerging standards. By facilitating early adoption and validation of standards with comprehensive testing solutions, Spirent empowers organisations to confidently prepare for the transition to next-generation networking while maintaining performance and reliability in increasingly demanding environments.” For more from Spirent, click here.

3000km of DWDM deployed in 72 hours in Eastern Europe

GNM, a backbone ISP operating its own, international peer-to-peer distributed Internet Exchange Network, GNM-IX, based on an 18,000+ kilometre Dense Wavelength Division Multiplexing (DWDM) network, has completed the deployment of nearly 3000km of long-haul DWDM infrastructure across Romania, Bulgaria, and Serbia, with installation and commissioning finalised in just 72 hours. The project was delivered entirely by GNM’s in-house engineering team, without the involvement of external contractors. All stages - from planning and logistics to setup and final network configuration - were handled internally, aiming to ensure full control, integration with existing infrastructure, and adherence to tight deployment timelines. "This kind of delivery isn’t a special case for us," says Alex Surkov, Head of Development at GNM. "We’ve built our processes to operate at this pace - with full internal coordination, no outsourcing, and flexible routing options already in place. That’s what allows us to move quickly and deliver stable, production-ready transport in days, not weeks." The new segment, according to the company, strengthens GNM’s DWDM footprint, which spans Europe, the Middle East, and Asia, and connects to key interconnection hubs including Frankfurt, Amsterdam, Helsinki, and Singapore. It says services are designed to meet the needs of carriers, content providers, and data centre operators, with a focus on scalability, redundancy, and delivery speed.

PoliCloud raises €7.5 million

PoliCloud, a provider and developer of high performance computing (HPC) cloud infrastructure, has announced its €7.5 million (£6.42 million) seed fundraise. The funding was led by Global Ventures, a VC firm in MENA, with participation from MI8 Limited, a Hong Kong multi-family office; OneRagtime, a Paris-based venture capital firm; Inria, France’s National Institute for Research in Digital Science and Technology; and other private investors. The proceeds will be used to hire the operating team and grow the business globally with a focus on public entities in Europe. PoliCloud says it is responding to demand following global cloud growth (~20% annually). Accelerated demand for AI requires affordable and scalable computing power, and the market is ripe for a Europe-led solution to lessen dependence on US cloud providers, who currently dominate the $800 billion (£583.9 billion) market. David Gurlé, Founder of PoliCloud, claims, “PoliCloud is meeting a critical market demand for sovereign cloud infrastructure that is not only secure and abundant, but also eco-responsible. Our unique edge computing capabilities deliver significant benefits to both public and private sector users. “The time is right for a new, European solution that reduces reliance on US cloud providers and offers affordable, scalable computing power, especially as AI adoption accelerates. We are grateful to Global Ventures and all our investors for their support as we enter this exciting phase of expansion.” Current cloud expansion suffers from high usage costs and dependence on hyperscalers - such as Google or Amazon - whose models use massive, centralised data facilities with high implementation costs and challenging environmental conditions. In this respect, PoliCloud claims it has the following competitive advantages: ● 'Unlimited and flexible computing power,' provided by federating with the grid. By y/e 2025, it says it will have >1,000+ GPUs and by y/e 2026 >20,000+ GPUs;● Computing resources are delivered to where they are needed and 'empower local communities;'● Small footprint and energy needs;● Rapid time to market, with flexibility and adaptability;● Capex and Opex offset by sharing unused capacity; and● 'More resilient, higher performance, and more scalable by design.' PoliCloud’s operating model combines its hardware and infrastructure with Hivenet’s distributed storage and computing software. PoliCloud designs, builds, and operates its own computers and micro-data centres. It was launched in February 2025 at the World Artificial Intelligence Cannes Festival (WAICF) with support from the five cities of the Alpes-Maritimes. Simon Sharp, Senior Partner of Global Ventures, comments, “Global Ventures is delighted to lead PoliCloud’s seed fund raise and work again with David and his talented management team, following their track record of successful delivery in Hivenet. [...] Their distributed data centres have multiple competitive advantages: delivering next-gen, sovereign computing resources where they are needed; with more resilience; faster performance; greater security; while being cheaper to build and maintain. The exponential growth in AI demand and the need for reliable, scalable computing power means the company’s future is a very bright one.” Stephanie Hospital, Founder & CEO of OneRagtime, adds, “As an early investor and believer in David and Hivenet, [...] OneRagtime is excited to invest in PoliCloud. The company is uniquely positioned to provide decentralised, unlimited computing power affordably, securely, and in an eco-responsible way – for which substantial demand exists.” Bruno Sportisse, CEO of Inria, says, "Inria Participations is delighted to become an investor in PoliCloud as it is a logical extension of Inria's existing strategic partnership with Hivenet. Inria and PoliCloud share the same philosophy of a decentralised path to the cloud and for secure, distributed computing, but where resources can also be shared according to need. Achieving this goal is of strategic importance for France and its digital sovereignty." Guillaume Dhamelincourt, Managing Director of Mi8, concludes, “The opportunity to invest in PoliCloud was compelling for Mi8 as the world embraces AI and rapidly adjusts its demand for computing power. The multiple use cases for PoliClouds, such as SMEs - but also public enterprises who want to stay mindful of their IT strategy's impact - is an attractive market environment and we look forward to PoliCloud’s future growth with great confidence.”

Nasuni achieves AWS Energy & Utilities Competency status

Nasuni, a unified file data platform company, has announced that it has achieved Amazon Web Services (AWS) Energy & Utilities Competency status. This designation recognises that Nasuni has demonstrated expertise in helping customers leverage AWS cloud technology to transform complex systems and accelerate the transition to a sustainable energy and utilities future. To receive the designation, AWS Partners undergo a rigorous technical validation process, including a customer reference audit. The AWS Energy & Utilities Competency provides energy and utilities customers the ability to more easily select skilled partners to help accelerate their digital transformations. "Our strategic collaboration with AWS is redefining how energy companies harness seismic data,” comments Michael Sotnick, SVP of Business & Corporate Development at Nasuni. “Together, we’re removing traditional infrastructure barriers and unlocking faster, smarter subsurface decisions. By integrating Nasuni’s global unified file data platform with the power of AWS solutions including Amazon Simple Storage Service (S3), Amazon Bedrock, and Amazon Q, we’re helping upstream operators accelerate time to first oil, boost capital efficiency, and prepare for the next era of data-driven exploration." AWS says it is enabling scalable, flexible, and cost-effective solutions from startups to global enterprises. To support the integration and deployment of these solutions, AWS established the AWS Competency Program to help customers identify AWS Partners with industry experience and expertise. By bringing together Nasuni’s cloud-native file data platform with Amazon S3 and other AWS services, the company claims energy customers could eliminate data silos, reduce interpretation cycle times, and unlock the value of seismic data for AI-driven exploration. For more from Nasuni, click here.

Bitrise first mobile DevOps platform to launch data centre in EU

Bitrise, a mobile DevOps platform, today announced plans to launch a data centre in the Netherlands in a response to increased demand for data residency in the European Union (EU). The new data centre will be the first in the EU operated by a DevOps platform, aiming to provide businesses with a fully-hosted and managed solution to meet the stringent data security and compliance requirements of the region. Bitrise will invest $3 million (£2.2 million) in the project, supporting the anticipated 22% year-on-year growth in European data centre capacity in 2025 as the continent focuses on operational resilience and data sovereignty. “In an era of geopolitical volatility and increasing regulatory complexity, mobile innovation in Europe demands sovereignty, speed, and security,” announces Barnabás Birmacher, CEO and Co-founder of Bitrise. “By launching the first EU-hosted DevOps platform, Bitrise is giving customers total control over their data, ensuring compliance and empowering them to scale development faster and more securely.” This investment in an Amsterdam-based data centre marks a step forwards in enhancing support for EU customers. By replicating the data centre model used in the US, Bitrise intends to deploy the same high-performance Apple M4 and Linux-based infrastructure in Europe, allowing businesses to choose their data residency and meet risk and compliance requirements. This expansion is a direct response to the growing demand from EU-based companies and global brands operating in the region. By strengthening data security and sovereignty, customers should have access to the tools they need to scale development securely and stay competitive in a rapidly changing market. “With data sovereignty becoming a critical priority for European businesses, Bitrise’s move to launch an EU-based data centre couldn’t be more timely," comments Reza Malekzadeh, General Partner at Partech and Bitrise board member. "Bitrise is setting a new standard for DevOps in Europe by giving companies the ability to meet stringent regulatory requirements without compromising on speed or innovation.” Recent regulatory changes and international data transfer challenges have created a complex environment for companies operating across borders. The data centre market in Europe is estimated to grow by $291.7 billion (£214.3 billion) from 2024 to 2028, driven by demand for local data processing and storage solutions. European companies in security-sensitive and regulated industries often rely on cloud providers in the US or spend millions to build their own local infrastructure. This has created a major gap in the market for compliant, hosted solutions. “We recognise the critical need for sovereign hosting solutions for mobile CI/CD infrastructure in the EU,” Barnabás says. “This move not only strengthens our presence in Europe, but underscores our commitment to solving the complex challenges our partners face, allowing them to innovate and scale without compromise.” Bitrise’s Amsterdam data centre will, according to the company, emulate Bitrise’s existing infrastructure model, providing:• Access to the fastest Apple Silicon and Linux machines for iOS and Android.• Advanced physical and network security measures.• Full compliance with EU data protection standards.• High-speed connectivity to ensure rapid build times. The data centre will support all Bitrise products and services, aiming to allow customers to build, test, and automate their applications without source code ever leaving the EU. In contrast to the majority of DevOps providers with US-only hosting capabilities, Bitrise’s expansion, the company claims, creates a DevOps platform that caters to data residency and digital operational resilience requirements. “By filling this gap in the market, we’re addressing a critical need for businesses throughout the EU,” Barnabás continues. “Our ability to quickly deploy and scale infrastructure based on our successful US model allows us to move fast and establish a strong presence in this underserved market.”

Chemists create molecular magnet, boosting data storage by 100x

Scientists at The University of Manchester have designed a molecule that can remember magnetic information at the highest temperature ever recorded for this kind of material. In a boon for the future of data storage technologies, the researchers have made a new single-molecule magnet that retains its magnetic memory up to 100 Kelvin (-173 °C) – around the temperature of the moon at night. The finding, published in the journal Nature, is a significant advancement on the previous record of 80 Kelvin (-193 °C). While still a long way from working in a standard freezer, or at room temperature, data storage at 100 Kelvin could be feasible in huge data centres, such as those used by Google. If perfected, these single-molecule magnets could pack vast amounts of information into incredibly small spaces – possibly more than three terabytes of data per square centimetre. That’s around half a million TikTok videos squeezed into a hard drive that’s the size of a postage stamp. The research was led by The University of Manchester, with computational modelling led by the Australian National University (ANU). David Mills, Professor of Inorganic Chemistry at The University of Manchester, comments, “This research showcases the power of chemists to deliberately design and build molecules with targeted properties. The results are an exciting prospect for the use of single-molecule magnets in data storage media that is 100 times more dense than the absolute limit of current technologies. “Although the new magnet still needs cooling far below room temperature, it is now well above the temperature of liquid nitrogen (77 Kelvin), which is a readily available coolant. So, while we won’t be seeing this type of data storage in our mobile phones for a while, it does make storing information in huge data centres more feasible.” Magnetic materials have long played an important role in data storage technologies. Currently, hard drives store data by magnetising tiny regions made up of many atoms all working together to retain memory. Single-molecule magnets can store information individually and don’t need help from any neighbouring atoms to retain their memory, offering the potential for incredibly high data density. But, until now, the challenge has always been the incredibly cold temperatures needed in order for them to function. The key to the new magnets’ success is the unique structure, with the element dysprosium located between two nitrogen atoms. These three atoms are arranged almost in a straight line – a configuration predicted to boost magnetic performance, but now realised for the first time. Usually, when dysprosium is bonded to only two nitrogen atoms it tends to form molecules with more bent or irregular shapes. In the new molecule, the researchers added a chemical group called an alkene that acts like a molecular pin, binding to dysprosium to hold the structure in place. The team at the Australian National University developed a new theoretical model to simulate the molecule’s magnetic behaviour to allow them to explain why this particular molecular magnet performs so well compared to previous designs. Now, the researchers will use these results as a blueprint to guide the design of even better molecular magnets.

LSC completes new dark fibre route in Kansas City

Light Source Communications (LSC), an owner-operator of networks serving enterprises throughout the US, has announced it has completed work on a new dark fibre network in Kansas City, Missouri, USA, aiming to deliver new opportunities to the region’s tech-rich ecosystem at a time when artificial intelligence (AI) and other high-performance computing (HPC) technologies are driving demand for greater connectivity. The 35-mile metro ring already has a major hyperscaler as the anchor tenant, with more on the way, as well as connections to four data centres so far. The Kansas City route is the first of LSC’s four new network builds to be completed this year, with projects in Las Vegas, Phoenix, and Tulsa also on track to be finished in 2025. “We’re thrilled to raise a toast to this exciting milestone,” says Debra Freitas, CEO of LSC. “Expanding into the Kansas City market is a key step in our strategic growth across high-demand US regions. As AI and HPC continue to drive unprecedented connectivity needs, we remain committed to delivering high-capacity, low-latency solutions to organisations of all sizes. As a carrier-neutral, customer-agnostic provider, LSC is proud to support the evolving demands of today’s digital economy.” Kansas City is the third-fastest-growing tech market in the US and has emerged as a hub for data centre projects. The area’s infrastructure, existing tech sector, and trained workforce make it a prime location for LSC’s dark fibre network. The route will be entirely underground with a high fibre count and conduit system. The Kansas City project follows a similar pattern to LSC’s other dark fibre builds underway. In Las Vegas, the company is building a 60-mile route that intends to bring hyper-connectivity to one of the country’s fastest-growing data centre markets. The Phoenix network will encompass 300+ miles and 15 rings. In Tulsa, LSC is adding 80 miles of new fibre to its existing 50-mile network. In addition to all of the networks being underground, all are anchored by a hyperscale tenant.

IBM, RIKEN unveil first Quantum System Two outside of the US

IBM, an American multinational technology corporation, and RIKEN, a national research laboratory in Japan, have unveiled the first IBM Quantum System Two ever to be deployed outside of the United States and beyond an IBM quantum data centre. The availability of this system also marks a milestone as the first quantum computer to be co-located with RIKEN's supercomputer, Fugaku — one of the most powerful classical systems on Earth. This effort is supported by the New Energy and Industrial Technology Development Organisation (NEDO), an organisation under the jurisdiction of Japan's Ministry of Economy, Trade, and Industry (METI)'s 'Development of Integrated Utilisation Technology for Quantum and Supercomputers' as part of the 'Project for Research and Development of Enhanced Infrastructures for Post 5G Information and Communications Systems.' IBM Quantum System Two at RIKEN is powered by IBM's 156-qubit IBM Quantum Heron, one of the company's quantum processors. IBM Heron's quality as measured by the two-qubit error rate, across a 100-qubit layered circuit, is 3x10-3 — which, the company claims, is 10 times better than the previous generation 127-qubit IBM Quantum Eagle. IBM Heron's speed, as measured by the CLOPS (circuit layer operations per second) metric, is 250,000, which would reflect another 10 times improvement in the past year over IBM Eagle. At a scale of 156 qubits, with these quality and speed metrics, Heron is the most performant quantum processor in the world. This latest Heron is capable of running quantum circuits that are beyond brute-force simulations on classical computers, and its connection to Fugaku will enable RIKEN teams to use quantum-centric supercomputing approaches to push forward research on advanced algorithms, such as fundamental chemistry problems. The new IBM Quantum System Two is co-located with Fugaku within the RIKEN Center for Computational Science (R-CCS), one of Japan's high-performance computing (HPC) centres. The computers are linked through a high-speed network at the fundamental instruction level to form a proving ground for quantum-centric supercomputing. This low-level integration aims to allow RIKEN and IBM engineers to develop parallelised workloads, low-latency classical-quantum communication protocols, and advanced compilation passes and libraries. Because quantum and classical systems will ultimately offer different computational strengths, this hopes to allow each paradigm to perform the parts of an algorithm for which it is best suited. This new development expands IBM's global fleet of quantum computers and was officially launched during a ribbon-cutting ceremony today (24 June 2025) in Kobe, Japan. The event featured opening remarks from RIKEN President Makoto Gonokami; Jay Gambetta, IBM Fellow and Vice President of IBM Quantum; Akio Yamaguchi, General Manager of IBM Japan; as well as local parliament members and representatives from the Kobe Prefecture and City, METI, NEDO, and MEXT. "The future of computing is quantum-centric and with our partners at RIKEN we are taking a big step forward to make this vision a reality," claims Jay Gambetta, VP, IBM Quantum. "The new IBM Quantum System Two, powered by our latest Heron processor and connected to Fugaku, will allow scientists and engineers to push the limits of what is possible." "By combining Fugaku and the IBM Quantum System Two, RIKEN aims to lead Japan into a new era of high-performance computing," says Mitsuhisa Sato, Division Director of the Quantum-HPC Hybrid Platform Division, RIKEN Center for Computational Science. "Our mission is to develop and demonstrate practical quantum-HPC hybrid workflows that can be explored by both the scientific community and industry. The connection of these two systems enables us to take critical steps toward realising this vision." The installation of IBM Quantum System Two at RIKEN is poised to expand previous efforts by RIKEN and IBM researchers as they seek to discover algorithms that offer quantum advantage: the point at which a quantum computer can solve a problem faster, cheaper, or more accurately than any known classical method. This includes work recently featured on the cover of Science Advances, based on sample-based quantum diagonalisation (SQD) techniques to accurately model the electronic structure of iron sulphides, a compound present widely in nature and organic systems. The ability to realistically model such a complex system is essential for many problems in chemistry, and was historically believed to require fault-tolerant quantum computers. SQD workflows are among the first demonstrations of how the near-term quantum computers of today can provide scientific value when integrated with powerful classical infrastructure. For more from IBM, click here.

Arelion connects EcoDataCenter to Nordic AI superhighway

Arelion, a Swedish telecommunications company formerly known as Telia Carrier, today announced it is upgrading its existing Point-of-Presence (PoP) at Swedish sustainable data centre operator EcoDataCenter’s Falun, Sweden, data centre campus with the latest-generation open optical line systems. Arelion’s network enhancement enables connectivity to its AI superhighway in the Nordics, leveraging scalable 400G and 800G coherent optics. This upgrade continues Arelion’s investment in its Scandinavian network by enhancing capacity and diversity to support its wholesale and enterprise customers’ AI deployments. EcoDataCenter builds data centres designed for demanding AI workloads and is used by global AI companies such as DeepL and CoreWeave. The company is also currently experiencing growth across the region. “Arelion’s investment is crucial in bolstering Scandinavia’s latest wave of technological innovation amid the region’s AI market growth,” argues Peter Michelson, CEO of EcoDataCenter. “By providing high-capacity connectivity to sustainable data centre infrastructure, we will collaborate to empower customers with Tier-1 services that support the AI ecosystem, enabling digital transformation across vital industries.” Sweden’s AI sector is experiencing growing investment, with analysts projecting the market will reach $6.35 billion by 2031, growing at a Compound Annual Growth Rate (CAGR) of 26.24%. The country’s data centre market is also growing to support these applications through digital infrastructure, with experts estimating it will reach $2.73 billion by 2031, growing at a CAGR of 10.39%. This growth is most likely driven by the higher availability of sustainable power in Sweden, with the country’s existing data centre capacity totalling over 130 MW, which is critical for supporting energy-demanding AI workloads. Arelion is offering its enhanced capabilities with delivery starting early Q3 2025. “This strategic deployment continues our Scandinavian network investments, allowing us to provide the vital backbone connectivity needed to support AI and cloud applications in the region’s booming technology markets,” comments Patrik Andreasson, Head of Sales Nordic & CEE at Arelion. “Our partnership combines high-capacity services with energy-efficient infrastructure, accelerating our customers’ AI deployments to spur further innovation across the Nordics.” For more from EcoDataCenter, click here.