• Ooredoo leans on Nvidia for sovereign AI launch
• The FCC’s chairman sets out his stall
• Quantum specialist IonQ eyes key role in Texas

In today’s industry news roundup: Ooredoo launches sovereign AI services in Qatar, just a year after striking a deal with Nvidia; the FCC’s chairman Brendan Carr outlines the key points of his Build America Agenda; quantum technology vendor IonQ embeds itself in the Texas Quantum Initiative; and much more!

Qatar-based telco Ooredoo has launched “advanced sovereign AI cloud services” from its local datacentre facilities that house the latest Nvidia Hopper GPUs. “By enabling advanced artificial intelligence applications without relying on international providers, and by delivering high-performance computing infrastructure locally, Ooredoo is accelerating AI adoption across key sectors, including energy, finance, logistics, healthcare and smart city development. For customers, this means faster access to computing power, the ability to handle massive datasets securely and the convenience of deploying AI solutions under national data policies, all with the low latency and reliability of in-country infrastructure,” noted Ooredoo, which originally struck an AI infrastructure partnership with Nvidia in June 2024. Ooredoo Qatar CEO Sheikh Ali Bin Jabor Al-Thani, stated: “We are proud to bring this world-class AI infrastructure to Qatar, equipping our customers with the tools they need to turn ambition into real-world solutions. Our collaboration with Nvidia paves the way for a new generation of innovation, empowering everyone from startups to government entities to benefit from the full potential of artificial intelligence, securely and efficiently. This, in turn, will fuel economic growth, upgrade citizen experiences and cement Qatar’s position as a digital leader in the region.” In addition, the launch “directly supports the ambitions of the Qatar Digital Agenda 2030 and the Qatar National AI Strategy, which call for robust digital infrastructure, local hosting of critical technology, and the responsible development of AI to benefit society and the economy,” the telco added.

Federal Communications Commission (FCC) chairman Brendan Carr has delivered the first major policy speech of his chairmanship in Sioux Falls, South Dakota, during which he outlined his “Build America Agenda” for the US regulator. “It is time to unleash new growth and opportunity.  It is time to build. Through the FCC’s Build America Agenda, we will do exactly that,” he stated. As part of his agenda, Carr intends to “free up more spectrum that connects our communities and restore America’s wireless leadership,” as well as “unleash high-speed infrastructure builds in communities all across the country.  We made a lot of good progress on this front during President Trump’s first term but there is more work to be done to cut red tape that is driving up costs and holding back internet builds.” Carr also intends to “ensure the US maintains secure and resilient networks. It will extend America’s leadership over China in the race for critical technologies like 5G, 6G and AI so that American companies continue to be the gold standard around the world.”

First it was Texas Hold’em, now it’s Texas Quantum. The stock of IonQ, Inc, the US quantum computing hardware and software company that develops general-purpose trapped ion quantum computers and the software needed to generate, optimise and execute quantum circuits, is on the up and up. The company is working with the likes of SK Telecom and Turkcell on their quantum-safe networking developments and systems, and was in the headlines three weeks ago when it paid $1.075bn to acquire Oxford Ionics, headquartered in the English university city. IonQ is a big believer in quantum computing and a major supporter in the ambition of the ‘lone star state’ to become “a major hub for quantum innovation”. Last week, the Texas State Legislature passed the Texas Quantum Initiative that establishes a strategic framework “to leverage the state’s academic strength, technology sector and entrepreneurial drive.” The aim is to provide guidance, investment incentives and research opportunities that advance quantum computing, networking and sensing technologies. Earlier in the year, the CEO of IonQ, Niccolo de Masi, made a very favourable impression when he testified before the Texas Senate Business and Commerce Committee during which he emphasised the critical role of quantum computing in ensuring cybersecurity in a very uncertain future. “Quantum technologies offer unparalleled potential to protect digital infrastructure, optimise systems and uncover solutions that were once beyond reach,” stated de Masi in this announcement. “We believe Texas is poised to play a defining role in the future of this technology.” IonQ’s current generation of quantum computers are the IonQ Forte and IonQ Forte Enterprise. The Forte is the company’s latest evolutionary step towards the ultimate goal of software-configurable quantum machines where qubit and gate configuration can be tailored to the needs of individual clients. It is IonQ’s largest single core quantum processor and highest performing system to date. Meanwhile, the Forte Enterprise is another step towards commercial quantum computing-based system miniaturisation, modularity and rack-mountable form factor. IonQ’s goal is to build quantum computers with 2 million physical qubits by 2030. The Texas Quantum Initiative establishes a structured approach to capitalise on the state’s strengths in academia, technology and entrepreneurship, and will establish a dedicated funding mechanism to support quantum research and development. Funds will be allocated to “promising startups, established companies and university research labs, providing the financial resources needed to advance quantum technologies and bring them to market. The initiative’s long-term vision is to create a self-sustaining ecosystem where innovation flourishes and Texas remains at the forefront of the quantum revolution.” Yeehaw!

Ericsson has inaugurated its new regional HQ in Riyadh, Saudi Arabia, a move it says will “enable us to further deepen our close partnerships with customers and partners, accelerate the adoption of advanced technologies and build a more connected future,” noted the vendor’s president of Ericsson Europe, Middle East and Africa, Patrick Johansson. Customers, including STC Group and Mobily, as well as government ministers, attended the opening of the HQ. 

Latvia’s largest telco, Latvijas Mobilais Telefons (LMT), has selected Ericsson to modernise its subscriber management systems. The vendor will upgrade LMT’s cloud infrastructure and replace its legacy communication services and user data management systems with Ericsson’s network functions virtualisation infrastructure (NFVI) solution, “giving it the ability to deploy both virtualised and cloud-native network functions, and opening up a wide range of transformation avenues,” noted the vendor in this announcement. “The new agreement also upgrades LMT’s IP multimedia subsystem (IMS) and consolidates LMT’s subscriber data systems within its core network into a future-ready architecture, establishing a solid foundation for scalable, high-quality service innovation,” added Ericsson. 

Google has just bought 200 megawatts (MW) of energy to be generated by a nuclear fusion technology that doesn’t exist yet! Fusion technology has long been regarded as the theoretical ultimate and perfect source of infinitely renewable, safe, clean power, but providing it as a practical daily reality remains elusive. However, Google, and other big tech firms, are building more and more datacentres to power their vaunting AI ambitions even as the US electricity grid comes under increasing stress and in danger of collapse. It is evident that change will have to come to the power supply sector and many companies, organisations, university laboratories and government departments around the world are vying to be the first with a real fusion solution. As CNN reports, Google has prised open its purse sufficiently to pay Commonwealth Fusion Systems (CFS) of Massachusetts, US, for the planned 200MW of power in the hope that its investment will pay off. The price paid has not been revealed, but CFS is now contracted to generate 200 MW of electricity from its Arc power plant. Current plans are for the Arc fusion plant to be built (but not necessarily operational) by “the early 2030s”, so Google will have to wait at least a decade to see if it has backed a winner. The US government estimates that demand for electricity will double by 2035. Currently – see what I did there? – CFS is constructing a “demonstration reactor” (called Sparc) on the outskirts of Boston. It will be completed late next year. Arc, CFS’s proposed commercial nuclear fusion plant, will be built further south, near Richmond, Virginia. Google is also an investor in CFS, having participated in its $1.8bn round of funding in 2021, and the big tech firm is to increase its stake with a further unspecified investment, CFS noted in this press release. To date, CFS has raised more investment cash than any other US fusion technology company. Much of the new investment will be spent on the development of a demonstration ‘tokamak’, a doughnut-shaped machine that applies massive magnets and molten plasma to force two atoms to merge. Fusion generates 10 million times more energy than fossil fuels without any resulting pollution. Fusion reactors can be powered by seawater and, unlike the nuclear fission plants of today, there is no radioactive danger or waste to be kept sealed under mountains for ten of thousands of years. Google isn’t the first hyperscaler to take a punt on nuclear fusion. Back in 2023, Microsoft signed a deal to take all the power output from the Washington State-headquartered Helion Energy’s first commercial reactor when it comes online, hopefully by 2028. According to the company website, Helion is developing “a magneto-inertial fusion technology to produce helium-3 and fusion power via aneutronic fusion, which could produce low-cost clean electric energy using a fuel that can be derived exclusively from water.”

Here’s a novel idea: Using datacentres as “shock absorbers” when power supplies elsewhere are under stress. Governments around the globe are increasingly concerned that national electricity grids could eventually collapse if the massive increases in the power necessary to cool the huge AI datacentres that are now being built are not moderated via the use of alternative energy sources, including onshore and offshore wind turbines, solar, wave and tidal technologies, hydro, river and estuary barrages and nuclear fusion reactors. Newly launched Emerald AI, headquartered in Faridabad, Haryana, India, has just been given the stamp of approval by Nvidia (the world-leading designer and seller of GPUs), together with $24.5m in seed funding from a consortium that includes NVentures, Nvidia’s venture capital arm. Describing itself, briefly and rather modestly, as “a technology company offering energy optimisation solutions”, Emerald AI also has the backing of a number of high-profile individual investors, including Jeff Dean, the chief scientist at Google, John Kerry, sometime US secretary of state, and Malcolm Turnbull, the former prime minister of Australia. The investment will be used further to develop the Emerald Conductor platform which, the company claims, could enable datacentres to obtain a grid connection significantly more quickly by managing energy consumption through AI. In this scenario, Emerald AI would operate as a “mediator’ between the electricity grid and datacentres and orchestrate AI workloads in real time to enable the datacentres to dynamically adjust energy consumption and help maintain grid stability whilst assuring “acceptable AI compute performance”. Dr Varun Sivaram, the CEO of Emerald AI, said the company has developed software that “transforms datacentres from grid liabilities into flexible assets.” He added: “We’re at a critical inflection point as exponential growth of AI computing pressures our electrical infrastructure. To unshackle AI technology progress from power constraints, Emerald AI transforms data centres from grid liabilities into flexible assets, enabling grid operators to swiftly interconnect AI, bolster reliability and energy security, and more efficiently harness spare capacity on today’s grids.” In a blog post, Nvidia explains that the Emerald Conductor’s ability to modulate power usage can be exploited by power-generating utilities to avoid rolling blackouts, protect communities from rising utility rates and support the integration of clean-energy systems into the electricity grid. Ayse Coskun, the chief scientist at Emerald AI (and, incidentally, a professor at Boston University), commented: “Renewable energy, which is intermittent and variable, is easier to add to a grid if that grid has lots of shock absorbers that can shift with changes in power supply. Datacentres can become some of those shock absorbers.” The first real-world demonstration of Emerald AI’s software-based platform transforming a production AI cluster into a grid-responsive asset took place in Phoenix, Arizona. It was conducted by Emerald AI in partnership with Oracle cloud infrastructure, Nvidia and the regional utility Salt River Project (SRP) under the aegis of the EPRI DCFlex initiative. EPRI, is an independent, non-profit US organisation that focuses on the generation, delivery and use of electricity to help address challenges in the energy industry, including reliability, efficiency, affordability, health, safety and the environment. The test offered the first validation of a system providing sustained, accurate power reductions using only workload orchestration – all without requiring any hardware retrofits or energy storage systems. It conclusively showed that an AI compute cluster using GPUs in a commercial datacentre could achieve a reduction of power usage of 25% over a three-hour peak period during times of high grid demand. Emerald says the results were achieved without compromising the performance of AI workloads. The company added that the intelligent management of energy consumption via software “shows that datacentres can bypass grid interconnection queues that can stretch 7 to 10 years or more.” The next step is for Emerald AI and its partners to conduct a much larger-scale demonstration that will be based in Phoenix but will extend across the US. 

– The staff, TelecomTV