Qualcomm struck a deal to acquire AI software company Modular for about $3.9 billion, in all stock-deal aimed at strengthening its software foundation for generative and agentic AI across data centre and edge environments.

The deal will pair Qualcomm’s silicon expertise with Modular’s AI-native software platform, which is designed to run AI models efficiently across CPUs, GPUs, NPUs and custom ASICs without requiring separate rewrites for each type of accelerator.

The chipmaker stated developers and enterprises will only need to build once ahead of deploying across multiple environments at a lower total cost of ownership.

Qualcomm framed the acquisition as a response to a shift in what is constraining AI deployment at scale. As AI use grows, the company argued, efficiency rather than raw capability becomes the limiting factor, since performance-per-watt drives inference costs and those costs determine what can scale.

Closing this gap, according to Qualcomm, requires software that connects system-level optimisation with heterogeneous, disaggregated compute, turning silicon performance into AI services which work reliably across different accelerators and use cases.

The deal is intended to expand Qualcomm’s data centre ambitions, supporting more efficient inference, orchestration and deployment in distributed AI systems while deepening relationships with model creators, developers, hyperscalers, and enterprise customers.

It is also expected to help the company achieve strong day-zero performance on new Qualcomm AI hardware as it rolls out.

Qualcomm President and CEO Cristiano Amon called the acquisition a pivotal moment for the AI industry. He stated agentic AI’s growth across data centres and edge environments is pushing the sector toward disaggregated, multi-vendor architectures which need a more open, modern software foundation.

He said Qualcomm believes the future belongs to developer-friendly platforms which can run across diverse compute environments, giving customers real choice in how and where they deploy AI.

“With Modular, we’re accelerating that shift, combining our scale and energy-efficient data center technologies with an open ecosystem approach to help drive the next chapter of AI,” Amon stated.

Modular co-founder and CEO Chris Lattner stated the company was built on the belief AI needs a more open and efficient software foundation spanning diverse hardware and deployment environments.

The transaction is expected to close in the second half of 2026, subject to customary closing conditions and regulatory approvals. Qualcomm plans to issue 19.2 million shares to Modular equity holders.

The post Qualcomm inks $4B deal for software specialist Modular appeared first on Mobile World Live.

Verizon’s road to Level 4 network autonomy was boosted by turning 33,000 employees into software developers earlier this year using Anthropic’s Claude model.

Verizon CTO Yago Tenorio (pictured) told Mobile World Live (MWL) the company gave about 500 application developers on his team access to Claude Code in December 2025.

“We started using it like there is no tomorrow,” he said. “In a few weeks we made everyone in technology proficient as software developers and then people started coding.

“There is a mantra now that says: ‘If you can imagine it, Claude can code it for you.’ It is amazing the massive amount of enablement, simplification and efficiency we achieved in a few months.”

He explained Verizon’s legacy as a long-established operator means its path to network autonomy is far more complex compared to a greenfield player. Years of M&A, overlapping legacy systems, merged networks and coexisting OSS platforms have created operational complexity which must be abstracted away before AI-driven automation can work at scale.

The AI autonomy revolution is well underway. In 2025, Verizon’s closed-loop automation platforms executed over 70 million network configuration changes autonomously.

“I feel like it’s making the network better. It is reacting in real time and it is freeing up hours of labour that can be then thrown on something that requires a human.”

But Tenorio draws a line between that kind of automation and what’s coming next. Those 70 million loops were largely programmatic, following decision trees and fixed logic which engineers wrote in advance, the network equivalent of a Python script running the same routine repeatedly.

The shift now underway replaces fixed logic with AI agents which can reason through a situation they were never explicitly programmed for because they are trained on the skills and judgment which make Verizon’s network teams effective, rather than coded with rules for every scenario.

Verizon is moving away from traditional hardware constraints to turn its infrastructure into a software-defined, self-organising reasoning engine which optimises itself in real time. The company describes this not as simply adopting tools, but as engineering an AI-native enterprise.

Level 4 network autonomy
In the TM Forum’s six-stage taxonomy for autonomous networks, Level 4 is a major shift from traditional automation to true intelligence. For this level, the network moves away from rigid, rule-based scripting and task-execution, transitioning instead to automating decisions.

CCS Insight research director Ian Fogg told MWL a part of an operator’s network may be at one level of autonomy, but other parts will likely be at a different level depending on how far along each operator’s suppliers and network teams are in modernisation.

“The barriers are that this is simply a very complicated area,” he said. “AI and especially generative AI and agentic AI are increasingly at the heart of autonomous network innovation as operators move to level 3 and beyond. 

“AI is an essential technology to enable the creation of autonomous networks,” he added.

Tenorio is careful not to overclaim on autonomy levels. He said Tesla represents something closer to Level 3, a Waymo represents Level 4, and Level 5 would be a car with no steering wheel at all.

He is not convinced the network needs to go Level 5. In his view, building a network with literally no human element anywhere in it is not necessarily the goal, even if individual domains or use cases can get there.

Across the board, in every domain and every use case, he considers Level 4 a reasonable target, while cautioning operators in the industry tend to generalise their autonomy successes across specific areas.

Rather than chase a number, Tenorio stated Verizon’s focus is on making autonomy matter operationally, whatever level it ends up being called.

He also pushes back on the idea that full autonomy means removing humans from the picture entirely. Even with what he calls a “totally dark” network operations center, people must still be dispatched to base stations and customer premises for hardware replacement or tuning which cannot be done remotely.

He argues agents are trained, not programmed, meaning the human expertise which has shaped a network for years does not disappear, it just gets transferred into the skills the agents learn from.

The loop which runs the network day-to-day can become fully closed, but the loop of deciding how those agents should behave still runs through human judgment.

AI models
To get there, Verizon has doubled down on generative and agentic AI. Tenorio describes Verizon’s approach as model-agnostic. For agents now running parts of its network, it is primarily using Anthropic’s Claude and Google’s Gemini.

Verizon is also optimising across model size and cost: simpler problems get routed to small language or cheaper models, while more complex tasks get the most capable ones available, including Claude Opus 4.7

Verizon had access to Anthropic’s Claude Fable 5 model prior to the US government shutting it down.

Tenorio used Fable 5 on a side project built with Claude Cowork and came away impressed, particularly by what felt like a broader contextual awareness and more lateral reasoning than he was used to. However, he is cautious about claiming a measurable gap versus Opus 4.8, since the same project has continued just as well on Opus 4.8 since Fable 5 access ended.

“I thought it was fantastic,” he said of Fable 5. “I was blown away by it.”

Starting in mid-January, the company expanded Claude access to its entire 33,000-person technology organisation over about six weeks, turning employees across modelling, capacity planning, capital deployment, and operations into capable software developers regardless of their formal coding background.

It now counts between 7,000 and 10,000 weekly active users of the tool. Tenorio frames this as a deliberate first step on the road to autonomy, not a side project.

Before trying to make the network autonomous, Verizon wanted to find out what happens when everyone in the organisation can write software.

Architectural blueprint
Verizon is currently conducting an autonomous network programme with intelligent agents running full-time, although Tenorio resists calling the effort a pilot.

He instead describes it as a deliberate, already-in-production evolution. The same 70 million automated loops from last year are progressively having reasoning and token-based decision-making layered on top of them, with the foundational pieces of true network autonomy now live rather than experimental.

He identifies two areas for genuine autonomous network operations.

The first is a common data layer. Since an autonomous network is only as good as the data feeding it, the telemetry needs to be shared across the organisation in near real time, and it needs to reflect actual customer experience rather than what the network itself reports.

Tenorio stated Verizon began rolling this out at the start of June and it is now in production.

The second is a common repository of skills, tools and agents, the framework by which years of collective engineering expertise and pattern recognition gets encoded into a form AI agents can draw on.

Once both pieces are in place, Tenorio stated deploying domain-specific agents becomes comparatively straightforward.

He said this approach has no off-the-shelf component and is not organised around a specific network domain or generation. It is not built around radio, transport, core, or optical separately, and it is not tied to 5G or 6G.

Instead, Verizon designed it to run across the network as a whole, starting with making every employee a developer rather than buying a vendor’s packaged autonomy product.

A master agent can immediately detect a localised anomaly, take control and instantly spin up a network of specialised sub-agents focused on the exact root cause, pinpointing it down to the precise network domain and specific element.

Work which used to take engineering teams hours to manually diagnose and correct is now identified and resolved in less than two minutes, long before the end user ever experiences a dropped call or a lag in their data connection.

An autonomous system, however, is only as good as the live data feeding it. Traditional network KPIs tend to skew toward favourable conditions and do not always fully capture the true, localised end-user experience.

Verizon’s new AI-driven models aim to close the telemetry gap. By using aggregated, anonymised network performance signals, including floor-by-floor indoor environments, the company builds an accurate, hyper-granular picture of service quality, feeding the common data layer described above.

The rich physical telemetry, fused with existing network data, feeds a continuous optimisation loop which detects and resolves coverage and capacity issues proactively, often before a customer is affected.

Tenorio said the end-result is near real-time visibility into in-building performance environments, enabling automated alerting and network tuning at a level of granularity the company did not previously possess.

“We started by making everyone a software developer,” he said. “I don’t know if other companies are doing it.”

The post Feature: How Verizon is crafting its network for the agentic AI era appeared first on Mobile World Live.

SK Hynix outlined plans to go public in the US to raise KRW45.5 trillion ($29.4 billion), which would be the second-biggest IPO in history after SpaceX’s listing earlier this month.

The South Korean high bandwidth memory (HBM) chip company filed for the proposed IPO with the Securities and Exchange Commission to list American Depositary Receipts on the Nasdaq.

In total it plans to issue 17.79 million new shares to raise more than $29 billion, giving it a company valuation of $1.2 trillion. It expects to start trading on 10 July.

It will use the cash to fund the buildout of chip factories in South Korea and acquire chipmaking equipment including extreme ultraviolet lithography (EUV) scanners. It already has an order in place with ASML worth KRW12 trillion.

A supplier to Nvidia and Google, SK Hynix has been a big winner of the AI boom, with its shares quadrupling in value this year. It competes in the HBM segment with Samsung Electronics and US-based Micron.

Indrani De, global head of investment research at financial company FTSE Russell, told Bloomberg SK Hynix’s outperformance reflects its early move in HBM and strong demand for its chip.

“Supply shortages in HBM where Korean firms dominate with 80% market share are expected to last for years,” she said.

The size of the proposed IPO would see SK Hynix surpass Saudi company Aramco’s IPO in 2019, when it raised $25.6 billion, which is currently the second largest behind SpaceX’s $75 billion.  

The post SK Hynix seeks to raise $29B from US listing appeared first on Mobile World Live.

Industry associations the GSMA and TM Forum agreed a collaboration to aid adoption of practical AI systems across a range of use cases within the telecoms industry.

In a joint statement, GSMA director of AI initiatives Louis Powell and TM Forum EVP products and services Andy Tiller explained the organisations would work together to help create open AI models and frameworks able to meet the industry’s very specific needs.

“Networks underpin the cloud, device and edge capabilities on which AI depends,” the two argued.

“Yet telcos themselves have not fully captured the value of AI within their own operations”. 

“If the industry is to accelerate towards more autonomous, AI-native networks, it must align on shared assets, common standards and practical implementation pathways”.

The organisations address complementary areas towards the common goal. The GSMA is working on relevant models and agentic systems through the Open Telco AI initiative and TM Forum is focusing on frameworks and standards to align the industry and ensure interoperability.

They labelled the collaboration as being “about connecting the AI stack end-to-end, combining shared models, common frameworks and practical deployment pathways, with the trust required for critical network environments”.

Problem
Explaining the rationale for the closer ties between the organisations, Powell and Tiller argued adoption of AI is more complex in telecoms than other industries, but is a board-level priority in many organisations.

Among the difficulties cited is the multi-vendor nature of network architecture, with fragmented, siloed data and “very little tolerance for error”.

They also noted frontier AI models are not designed for operators and have not been trained on industry-specific data, leaving deployments limited to areas outside of the network layer such as customer experience.

Data published by GSMA Intelligence noted 16% of AI deployments target network use cases, despite networks accounting for 34% of opex.

The organisations believe the telecoms industry needs to “align to solve these challenges for itself, building AI approaches that reflect the complexity, economics and operational realities of the industry”.

The post GSMA, TM Forum team on telecom AI appeared first on Mobile World Live.

Orange lined up Usman Javaid (pictured) as its next AI chief, tasking the current chief product and marketing officer of its business unit with delivering on ambitious revenue targets for the technology.

Javaid is scheduled to succeed Steve Jarrett on 1 September, a little over three years after taking on his current post within Orange Business.

Orange is in the throes of implementing a strategy intended to derive more than €600 million in value from AI by 2028. It is also working to expand internal use of the technology and stated implementing agentic iterations would be a key element in Javaid’s remit.

During an Orange Business event in March, Javaid told Mobile World Live past experience in adapting to emerging technologies would serve the company well in the AI era, explaining the focus is on how to employ advances to benefit customers.

Orange credited Jarrett with having laid the foundations of its use of AI and establishing its credentials in the field.

There is a circularity to Orange’s appointment given Javaid began his career with the company in 2005. The executive also racked up experience in various roles held with Vodafone Group and AWS spanning cloud, IoT and network strategy.

The post Orange readies AI chief succession appeared first on Mobile World Live.

LIVE FROM MWC SHANGHAI 2026: AI naturally dominated the opening keynote session of this year’s event, with China Mobile and Filipino operator Globe Telecom offering insight into consumer benefits of the tech and how it is driving growth at both companies.

Carl Cruz, CEO and president of Globe Telecom (pictured), outlined how his business is arguably one of the most advanced operators globally in terms of implementing AI for its own employee work practices as well as customer use cases.

It is a journey which started in 2024 when Globe Telecom established a group to lead its adoption of AI technology and created the position of chief AI officer (CAIO). The role was taken up by Anton Bonifacio and since widened to become Chief Intelligence and Trust Officer (CITO), leading the integrated structure of key technology and enabling groups across AI, data, cybersecurity and digital policy.

“Our goal was simple,” declared Cruz. “Democratise technology so innovation could come from anywhere. We see AI not as a threat but as a tool for power and growth.”

To that end, Cruz said Globe Telecom had witnessed more than 80 citizen-developed use cases across the business using a hub and spoke operating model. He cited other statistics claiming AI is now embedded in 90% of Globe’s business practices, with more than 260 relevant projects running across all business units.

Cruz put the AI focus on four main areas: hyper personalisation; customer experience; autonomous networks; and the creation of new products and services.

He gave the example of how Globe Telecom used AI to build a new billing agent: “It used to take days, even weeks, to reconcile billing. Today it takes seconds and minutes.”

“It’s about deepening customer relevance, unlocking new value and creating new growth engines for Globe.”

Cruz also talked up Globe Telecom’s AI Fiesta initiative, an affordable aggregator platform which bundles leading global AI models including ChatGPT, Claude, Gemini Pro, Perplexity, DeepSeek and Grok under a single, budget-friendly subscription. It is designed to democratise AI access for students, creators and small business owners.

And the operator has launched Globe Cloud, a personal cloud storage and file-backup service. Running Synchronoss technology, the app allows users to free up device space, sync media across devices, and use AI tools for editing and transforming images.

“AI at Globe is not an end to itself, it is a means to serve customers better, unlock new opportunities and growth,” Cruz said. “It’s not about using AI everywhere, it’s about using AI where it matters most; most importantly the communities we serve”.

“We are only scratching the surface of what’s possible.”

China Mobile: AI for All
The world’s largest operator China Mobile opened the keynote with a presentation tagged Mobile AI – AI for All.

Chair Zhongyue Chen highlighted how AI is enabling a hearing-impaired man to carry out everyday tasks.

“AI should not only be serving the elite or remain the domain of the masses,” Chen said.

The company claimed it is open to partnerships to further boost its AI offering. “We are not building a closed garden service for ourselves.”

And its AI for All mantra encompasses a variety of sectors, with Chen highlighting a focus on consumer services, showcasing how the technology is being used to supercharge the TV viewing experience in China of the FIFA World Cup; industries, for example in steel manufacturing where AI is helping reduce exposure to dangerously high temperatures; and infrastructure, where AI is measuring the power of individual models and also the capacity of platforms.

The post Globe, China Mobile drive AI use cases appeared first on Mobile World Live.

SoftBank Group CEO Masayoshi Son reportedly rejected the case put forward by Elon Musk’s SpaceX to build AI data centres in space, arguing the economics favour terrestrial infrastructure.

Bloomberg cited comments made by Son at SoftBank’s annual shareholder meeting, who stated the “the main advantage” of eliminating electricity costs by putting AI data centres in space is outweighed by “higher fees to transport everything into space, maintenance and communication delays”.

Son was apparently responding to a shareholder who asked whether SoftBank would pursue a strategy similar to that proposed by the SpaceX chief, who filed for US regulatory approval in February to deploy up to 1 million solar-powered satellites designed to operate as AI data centres. Musk argues persistent access to solar power will “achieve transformative cost and energy efficiency”, adding “in the long term, space-based AI is obviously the only way to scale”.

While describing the tech mogul as a “remarkable agent of change”, Son reportedly said SoftBank would instead focus on building “formidable” data centre capacity on Earth. “In the battle for AI, the next few years will be far more important than what might happen a decade or so from now,” he noted, adding “he who strikes first wins”.

Despite intensifying competition, Son argued AI has scope for “ten-fold, a hundred-fold” growth, adding there was ample room for OpenAI and rivals including Anthropic and Google to expand amid accelerating AI demand.

The remarks put Son’s strategy at odds with Musk’s long-term vision for AI. SpaceX launched a record-breaking $75 billion IPO earlier this month, with proceeds funnelled towards funding its space-based AI initiative alongside the expansion of its Starlink satellite network.

Meanwhile, SoftBank has committed around $65 billion to OpenAI despite reports the investments were straining its balance sheets, and pledged to invest hundreds of billions of dollars in terrestrial AI infrastructure and data centres globally.

The post SoftBank chief dismisses Musk orbital data centre gambit appeared first on Mobile World Live.

The Five Eyes intelligence agency called for urgent action to respond to the evolving AI landscape, warning while the technology would help improve defences over time, it also accelerates the speed, scale and sophistication of attacks.

Five Eyes is a global intelligence-sharing alliance between the UK, Canada, US, Australia and New Zealand: it stated frontier models are expected to exceed current industry expectations, transforming offensive and defensive cyber capabilities.

The timeline for this “is not years, it is months”, meaning action must be swift “to remain ahead”.

“In this environment, cyber resilience is integral to advancing business continuity, market confidence and long-term value,” it stated.

As part of a number of recommendations, the alliance urged leaders to understand and assess, risk, readiness and accountability; prioritise foundational cybersecurity practices and controls; empower cyber leaders with authority and resources; and actively engage as threats evolve.

“Success will come from getting the basics right, acting quickly and integrating cybersecurity into core business strategy. Those that do not will face growing operational and strategic disadvantage.”

It emphasised cyber risk cannot be treated purely as a technical issue, but rather a core business risk, arguing ensuring systems are secure by design should become standard industry practice, resilience should depend on a single solution or technology and there must be an understanding fresh vulnerabilities would emerge as AI systems evolve.

Five Eyes also called for better processes to swiftly patch faulty software, address legacy systems, and limit unnecessary system access and external connectivity.

On the flip side, the alliance stated AI should be used to bolster defences, explaining organisations integrating AI tools into security operations can detect vulnerabilities faster, improve quality, monitor unusual behaviour and respond to threats swiftly.

The post Five Eyes sounds alarm on AI cyber threat appeared first on Mobile World Live.

SpaceX reportedly signed a significant computing power deal with Reflection AI, making the open-source start-up the latest outside company to tap Elon Musk’s Colossus infrastructure.

CNBC reported Reflection will get access to Nvidia’s GB300 chips, the company’s top-tier hardware for training and running advanced AI models, housed at SpaceX’s Colossus 2 facility in Memphis, Tennessee.

Reflection agreed to pay SpaceX $150 million per month starting 1 July through 2029, with total payments reaching roughly $6.3 billion if the agreement runs its full term.

Either party can terminate the contract with 90 days’ notice after the first three months. Bloomberg reported Reflection will pay SpaceX’s AI division, SpaceXAI, under the arrangement.

Colossus was originally built in part to power Grok, Musk’s AI chatbot as a rival to ChatGPT, but SpaceX has increasingly opened the infrastructure to outside AI companies seeking compute capacity.

The deal is smaller than two of SpaceX’s other compute agreements. According to TechCrunch, SpaceX’s existing contracts with Anthropic and Google cost those companies $1.25 billion and $920 million per month, respectively, with both agreements also running through July 2029.

CNBC reported SpaceX has also struck a computing power-related deal with Cursor prior to buying the company for $60 billion earlier this month.

In October 2025, Nvidia led a $2 billion funding round for the open-source start-up, directly contributing $800 million of the total to value Reflection AI at $8 billion.

Reflection AI was founded in 2024 by Misha Laskin and Ioannis Antonoglou, both alumni of Google’s DeepMind lab.

The post SpaceX inks computing deal with Reflection AI appeared first on Mobile World Live.

Groq raised $650 million in funding as the company revamps itself from a chip startup into a data centre operator focused on AI inference.

The round was led by growth-focused investment company Disruptive and hedge fund Infinitum, both of which already hold board seats at Groq.

The funding follows a non-exclusive licensing agreement Groq struck with Nvidia in December 2025. At Nvidia’s GTC event earlier this year, the chipmaker unveiled its next-generation LPX platform, which incorporates Groq’s inference technology.

US-based Groq stated it plans to use the new cash to expand its network of 13 data centres in North America, Europe, the Middle East and Asia-Pacific, and quadruple total capacity to 200 megawatts by the end of 2027.

The new capital will also fund the fit-out of its existing footprint with Groq’s latest inference technology, including Nvidia’s new LPX system.

In September 2025 Groq raised $750 million, which valued it at $6.9 billion. Its backers include KDDI, Samsung Electronics, Cisco Systems and BlackRock.

Groq chair Alex Davis stated the company spent years building the technology, infrastructure and operational expertise needed for AI’s next phase, adding it is positioned to become a foundational layer of the AI economy.

Groq also announced several executive additions alongside the funding news. Alan Rice, who previously worked on xAI’s Colossus data centre project and Meta data centres, is now COO.

Sinclair Schuller will start as CTO in July, with Rakesh Malhotra joining as chief product officer the same month. Schuller and Malhotra previously worked together at enterprise cloud platform Apprenda and later co-founded a separate software engineering firm.

The post Groq lands $650M to aid shift from chips to AI Inference appeared first on Mobile World Live.

Nvidia unveiled what it described as the industry’s first full-stack safety platform for physical AI designed to accelerate the deployment of humanoid robots in industrial environments.

The company said its expanded Halos platform combines AI compute, system software, sensors and inspection tools into a single safety architecture aimed at helping robotics companies build and certify autonomous systems operating alongside human workers.

The chipmaker explained as robots become more autonomous and increasingly rely on AI foundation models, distributed sensors and accelerated computing, “scaling these systems requires a full-stack safety architecture” before they can be trusted to work alongside humans in factories, warehouses and logistics operations.

The stack includes Nvidia’s IGX Thor industrial computing platform, Holoscan Sensor Bridge for sensor connectivity and Halos OS software stack.

Nvidia also opened its Halos AI Systems Inspection Lab to robotics partners, aiming to help them complete pre-certification testing and validate safety systems before approaching regulators.

Digit
Humanoid robotics company Agility Robotics is the first partner to adopt elements of the platform into the proprietary safety system for its Digit humanoid robot, which is already being deployed for industrial applications with customers including Amazon, Schaeffler and Toyota Motor Manufacturing Canada.

The robotics player will also use Nvidia’s Inspection Lab to validate Digit’s safety-related software, AI components and cybersecurity protections against strict standards.

Agility Robotics CEO Peggy Johnson said: “For humanoids to deliver value at scale, safety has to be built into the robot and validated across the entire system.”

She added the partnership with Nvidia “extends our leadership in responsible automation”, describing it as “a non-negotiable requirement for bringing humanoids safely into industrial workflows”.

Deepu Talla, VP of robotics and edge AI at Nvidia said: “Physical AI is transforming how factories, warehouses and logistics operations work. He added developers will be able to “harness Nvidia’s proven autonomous vehicle safety foundation to develop safer robots faster” and deploy systems “alongside workers with greater confidence”.

The post Nvidia rolls out robot safety platform appeared first on Mobile World Live.

Amazon is now water positive in India, meaning the company now returns more water to communities than it uses across its direct operations, including corporate offices, data centres and fulfilment centres.

The milestone arrives ahead of the company’s original 2027 target and marks a step towards its broader global water stewardship strategy.

The achievement is noteworthy in a country facing acute water stress. India holds 18% of the world’s population but only 4% of its freshwater resources.

The company’s approach rests on three pillars: reducing water use, reusing water through treatment and harvesting systems, and replenishing water in stressed communities.

By comparing water returned through these initiatives against water withdrawn for operations, the company says it returned 120% of the water it used in 2025, with progress verified through internal and third-party audits.

“Water is fundamental to the communities where we operate, and we recognise both the scale of India’s water challenge and the opportunity to make a meaningful difference,” stated Abhinav Singh, Amazon’s VP of operations for India, APAC, Middle East and Turkiye.

On the reduction side, Amazon’s data centres in India do not use water for cooling, while water use across fulfilment centres and offices is largely tied to restrooms, kitchens and drinking water, managed through low-flow fixtures and smart meters that detect leaks.

On reuse, on-site sewage treatment plants recycled an estimated 298 million litres in 2025 for flushing and irrigation, while rainwater harvesting pits collected roughly 178 million litres which went into local groundwater.

For replenishment, Amazon has committed more than $6.5 million to water stewardship projects across Maharashtra, Karnataka, Telangana, Delhi NCR, Andhra Pradesh and Tamil Nadu, which are expected to replenish more than 4 billion litres annually.

The work ties into Amazon Web Services’ 2022 global pledge to become water positive by 2030 across its data centre operations.

The company stated it is now 75% of the way to its goal, returning three litres for every four used in 2025.

Amazon continues to expand its footprint in India, with plans to invest more than $35 billion in the country by 2030 to boost AI capabilities and digital transformations.

Microsoft and Google have also announced significant data centre investments in India over the past year.

The post Amazon hits water-positive target in India ahead of schedule appeared first on Mobile World Live.

Mavenir and Red Hat announced what they are calling an Integrated AI Platform, a joint offering designed to give mobile operators the infrastructure to sell services and capture revenue currently flowing to hyperscalers and frontier AI companies.

After making the decision to pivot away from open RAN last year, the vendor has turned its focus to AI, which is evidenced in its partnership with Red Hat.

The Integrated AI Platform blends Mavenir’s AI software with Red Hat’s enterprise Kubernetes and AI capabilities using the latter’s OpenShift and the former’s Digital Enablement (MDE) platform for billing and charging.

The result is a platform which enables operators to monetise AI through token-based consumption plans, with full control over pricing, SLAs and the large language models (LLMs) which power every interaction.

The two companies believe token-based AI consumption, the unit model powering every LLM interaction, mirrors the mobile data billing paradigm operators have long used.

Mobile operators spent decades mastering consumption-based data billing, and Mavenir CEO Pardeep Kohli believes the same model can now be applied to AI while also capturing AI revenue currently flowing to the likes of OpenAI, Anthropic and the major hyperscalers.

 “At the end of the day, most of the money is going to go to OpenAI and all these other players,” Kohli told Mobile World Live. “What we are hoping is that this time, there’s an opportunity for operators not only to use AI to cut their own cost and reduce opex and do more automation but also be a provider of AI services to enterprises or even to consumers.”

How it works
The AI platform operates across three deployment models: operator-branded AI services for their own subscribers; as the AI infrastructure layer powering operator AI grid deployments; and as a managed AI platform through which operators can offer the capabilities to enterprise customers on a consumption basis.

The architecture is designed as hybrid, using on-premises and open-source models to serve most of the traffic, while Mavenir’s model router provides policy-governed, metered access to external frontier models for use cases which require advanced reasoning or multimodal capability.

Kohli explained every generation of mobile technology, from 3G to 5G, saw the bulk of value flow to application developers and hyperscalers while operators became commoditised connectivity pipes.

He believes AI represents a rare chance to break that cycle, provided operators move fast enough to control the infrastructure layer. Kohli pointed to live deployments, including Turkcell running on-premises language models and Telefonica working with Mavenir on a range of use cases announced at MWC, as evidence the strategy is already in motion rather than theoretical.

Use cases
One example he provided MWL is real-time voice translation. An operator customer of Mavenir is deploying an agent which converts live calls between languages.

At a small scale, routing the workload through a hyperscaler is economically viable. At the scale of a major mobile network, it becomes prohibitively expensive unless the inference runs on infrastructure the operator controls directly.

The AI platform, as Kohli describes it, is the mechanism which makes the transition possible without rebuilding the operator’s entire operational model.

Kohli also said currently when a mobile operator customer sends a query to Gemini or OpenAI, it bypasses the operator entirely and then goes straight to the hyperscaler’s data centre with the operator earning nothing beyond the data connection.

His AI grid model changes the scenario by making the operator an intelligent routing layer. A simple query gets handled by a local on-premises model at low cost; a complex reasoning task gets routed to a frontier model; a non-urgent request could be sent to a cheaper, more distant data centre.

With the AI grid, the operator becomes a traffic controller in the AI value chain rather than, as Kohli put it, “a dump pipe”.

Sovereign angle
Regulated operators in markets with strict data residency requirements have ample reasons to prefer on-premises inference over sending subscriber data to hyperscaler data centres in unspecified jurisdictions.

The physical proximity argument of operators has merit as Kohli noted operators with urban data centres genuinely are better placed than hyperscalers whose capacity often sits in rural locations.

“We are going use case by use case, but the actual framework and platform is common, and it can be used for many applications,” he said.

Analyst outlook
AvidThink founder and principal Roy Chua raises several issues with the token-as-currency model. One of his points is a token is not a megabyte. A megabyte is a fixed, fungible unit of information.

He explained a token from a frontier reasoning model and a token from a fine-tuned small language model running on premises are not equivalent. One may resolve a complex problem immediately while the other “takes 10 times the tokens getting to a worse answer”.

Chua said the intelligence embedded in a token varies wildly between models, and tokenisation methods differ between model families. Until there is an established, auditable standard for what a token represents in terms of capability, he argues, “the token can’t function as the billing primitive”.

There is also the question of where developer loyalty sits today. Enterprises and software developers have built their AI workflows around the frontier labs and around aggregation layers which already solve model routing across providers.

“For developers and enterprises, trying to insert the operator into a value chain where most operators currently have no relationships and add little differentiated value is an uphill climb,” he said.

He stated a stronger move is not selling tokens as a currency, but selling sovereign, compliant, latency-advantaged AI services wrapped in outcomes, particularly for regulated buyers who need in-country control.

“Let me give credit where it’s due,” Chua said. “Pardeep Kohli and the Mavenir team have correctly identified the macro tailwind. The geopolitics around AI access are real, data sovereignty pressures are real, and the desire for operators to climb the value stack is older than the cloud itself.”

He continued: “However, token-as-a-service, isn’t a sellable product yet for most telcos, with very few exceptions. Someone like SoftBank could do it in the future.”

DTW Ignite
Kohli said Mavenir will have a booth for the first time at TM Forum’s DTW Ignite conference next week. Historically, he noted, DTW was the domain of operators’ IT and billing-side teams while Mobile World Congress served the network side.

Mavenir’s presence at DTW now reflects his view those two worlds are converging as operators build common platforms spanning both functions.

 “We have a booth now to explain how we can actually have a common platform, which can not only do billing and charging, customer onboarding and customer facing work, but also can manage your network,” he said. “And it learns about packet code, and it learns about voice calling, it learns about SMS, MMS and not just your bill.

“I believe this is a golden opportunity,” he added.

The post Analysis: Mavenir pitches AI tokens as a telco cash cow appeared first on Mobile World Live.

Mobile World Live (MWL) brings you our top three picks of the week as Ericsson CEO Borje Ekholm stepped down, Anthropic faced a US clampdown on its latest AI models and the UK outlined plans to ban under-16 users from social media.

Ericsson CEO Ekholm set for exit

What happened: Ericsson lined-up a changing of the guard, announcing Borje Ekholm will step down as CEO on 30 September after more than nine years at the helm and lining up current networks boss Per Narvinger to take over.

Why it matters: Ericsson chair Jan Carlson said the incoming CEO brings “deep technical knowledge of our industry as well as extensive commercial experience”, while crediting Ekholm with “extraordinary leadership and strategic advancements”. The chief joined as CEO in 2017 and inherited a business under pressure, steering it through payments scandals, which resulted in a 2019 US fine of more than $1 billion and a $206.7 million. Ekholm will remain at the company as an executive adviser until 15 June 2027.

Narvinger now takes over as 5G spending slows, 6G looms and AI reshapes network demand., He told the Financial Times (FT) the industry is at a “pivotal time”, adding users will “hopefully value your connectivity more” as AI dependence rises. David Hammarwall, currently head of Customer Unit T-Mobile in Market Area Americas, will be Narvinger’s successor as head of the Networks division.

Ericsson CEO Ekholm set for exit

US orders Anthropic to suspend Fable 5, Mythos 5 models

What happened: US President Donald Trump’s administration ordered Anthropic to immediately suspend access to its two most advanced AI models, Fable 5 and Mythos 5, over national security concerns.

Why it matters: The order applied to “any foreign national, whether inside or outside the United States”, prompting Anthropic to disable the models for all customers globally while it worked out how to comply. Anthropic said it believes officials were responding to a “potential narrow, non-universal jailbreak” affecting cybersecurity safeguards, but argued the issue involved a small number of known minor vulnerabilities which could be replicated on other publicly available models. The company pushed back on the recall, noting Fable 5 had been tested with the US government, the UK AI Safety Institute and private groups before launch. The Wall Street Journal (WSJ) reported Amazon CEO Andy Jassy was among tech executives who raised security concerns about the models with officials. Anthropic called the dispute a “misunderstanding” and is seeking to restore access, adding “if this standard was applied across the industry, we believe it would essentially halt all new model deployments for all frontier model providers”. Gina Neff, professor of responsible AI at Queen Mary University London, told the BBC the speed of AI development means and government and enterprises “need to be able to build up capabilities to keep our companies safe from cyberattacks”. She explained the restriction could limit safe AI testing and international cooperation, adding “we’re in uncharted territory at this point”. The latest development marks a sharp escalation in US scrutiny of Anthropic, which is already suing the Trump administration after the Department of War (DoW) blacklisted it as a “supply chain risk” earlier this year.

US orders Anthropic to suspend Fable 5, Mythos 5 models

UK plots under-16 social media ban

What happened: The UK government laid out plans to ban under-16 users from social media and impose wider child-safety restrictions on gaming, live-streaming and AI platforms, with the landmark online safety measures slated for early 2027.

Why it matters: The proposed ban would apply to platforms including Snapchat, TikTok, Instagram, YouTube, Facebook and X, with messaging apps such as WhatsApp and Signal excluded from the scope. Prime Minister Keir Starmer said the move will “give kids their childhood back”, arguing “tech giants had their chance and failed”. However, Assembly Research analyst Megan Jenkins said the government appeared to choose “what feels like a punitive measure for children” over making platforms “safer by design”. The UK’s plan follows Australia, which became the first country in the world to ban social media for under-16s last year. However, the government added it will “learn the lessons from Australia’s experience” by introducing tougher age verification tools. Technology secretary Liz Kendall told the BBC the government plans to disclose further details in July on VPNs, a potential method for bypassing restrictions, as well as curfews, chatbots and other restrictions. Alex Brown, head of TMT at law firm Simmons and Simmons said the move signals “a clear hardening of approach to online safety”, but warned companies must now “grapple with the difficult task of age verification”.

UK plots under-16 social media ban

The post The Friday File: Ericsson; Anthropic; UK appeared first on Mobile World Live.

Hyundai Motor Group was reported by media in South Korea to be set to buy SoftBank Group’s remaining 9.7% stake in robotics company Boston Dynamics for $325 million, assuming full ownership of the business.

Newspaper Maeil Business reported SoftBank had an option in its contract when it sold the bulk of the business to Hyundai to divest the remainder under certain terms.

Having apparently informed Hyundai of its intention to sell, the automotive company decided it would buy the rest subject to approval at a series of board meetings.

SoftBank sold 80% of its holding in former Alphabet business Boston Dynamics in 2021. Its interest in the company has subsequently reduced to the share Maeil Business reported it held today (19 June).

Boston Dynamics’ robots are a regular sight at technology events and other showcases, including robotic dog Spot and humanoid unit Atlas. Its products are aimed at industrial sectors including manufacturing, logistics and construction.

The newspaper noted Hyundai would likely focus on commercialisation of Atlas and undertaking a long-rumoured IPO of the robotics company after completing the buy.

The post SoftBank tipped to cash out of Boston Dynamics appeared first on Mobile World Live.

China’s commerce ministry is reportedly targeting broader use of AI across the country’s consumer economy, unveiling 17 policy measures intended to push the technology beyond industrial applications.

State broadcaster CCTV reported the ministry will use the measures to embed AI into households and businesses, viewing the technology as a lever for economic stimulus.

As part of the plan, it is apparently looking at a fundamental shift in how consumer electronics are designed and sold.

The way devices operate is also a target, moving from functional tools towards AI-integrated systems.

Advancing humanoid robotics in the home is high on the agenda, and the sector has already gained government and private investment.

Chinese players including Unitree and Ubtech were cited as front-runners in the market. They are vying to establish themselves commercially in industrial and consumer spaces.

Structural issues
On the services side, China wants to expand AI’s presence in retail services to public and personal areas.

CCTV cited comments from Lin Jian, deputy director of the international trade cooperation institute at the ministry of commerce, who said the measures are designed to solve structural issues within the country’s services sector.

These include addressing challenges around high labour costs and poor standardisation in service consumption, added Lin.

Beijing is stepping up its AI plans, with Bloomberg reporting earlier this month the government is planning to invest $295 billion over the next five years to build out data centres in the country.

The post China unveils plan to embed AI across society appeared first on Mobile World Live.

Orange and the French Alternative Energies and Atomic Energy Commission (CEA) launched a joint laboratory to explore semantic communications, a technology designed to cut data traffic and energy use in future AI-driven networks.

The AI-Native Communications lab will run for five years and develop semantic communications technologies and applications, which prioritise whether the receiver understands the meaning or content of a transmitted message, even when the received data is not exactly identical to what was sent. This marks a shift away from conventional data transmission, where networks aim to reproduce bits perfectly and treat any alteration as an error.

The partners said the approach could reduce the volume of data exchanged across networks, lowering energy consumption and infrastructure requirements. The technology could also support communication between AI agents, which are expected to play a greater role in future network architectures.

The laboratory will also combine research from Orange and CEA to study network evolution, create shared semantic representations between AI models and contribute to future standards work.

Announced at the annual VivaTech event in Paris by Orange CEO Christel Heydemann and CEA general administrator Anne-Isabelle Etienvre, the pair said the initiative forms part of a broader push to strengthen French and European sovereignty over future intelligent network technologies.

Lyse Brillouet, EVP of research at Orange, said the joint laboratory reflects the operator’s commitment to building “the networks of the future: smarter, more sustainable, and natively designed for artificial intelligence”. She added the move was “a key step in reconciling the performance of our infrastructure with the need to reduce our future carbon footprint”.

Meanwhile, Etienvre stated semantic communications marked “a major breakthrough in how future networks will process information”, adding the collaboration aimed to strengthen Europe’s ability to “design and influence technological standards”.

The post Orange, CEA lay groundwork for AI-native networks appeared first on Mobile World Live.

Anthropic CEO Dario Amodei and Google DeepMind’s Demis Hassabis reportedly lobbied for a US-led coalition to shape global AI rules, as Anthropic remained locked in negotiations with the Trump administration over export controls on its newest models.

CNBC reported that during a closed-door lunch at the G7 Summit attended by President Donald Trump and other world leaders, Amodei called for international cooperation to cover structured access to frontier models, a chip and critical components trade framework which excludes China and coordinated responses to AI risks in cybersecurity, bioterrorism and intelligence.

OpenAI CEO Sam Altman also called for a formal international forum to establish globally accepted testing standards, provide impartial analysis of AI capabilities and risks, and serve as a venue for cooperation among nations, according to an OpenAI briefing.

The outlet reported Canadian Prime Minister Mark Carney agreed the US could lead such a coalition, according to people familiar with the discussions.

In addition to Trump, Treasury Secretary Scott Bessent, Commerce Secretary Howard Lutnik and Secretary of State Marco Rubio were also in attendance.

Anthropic ban
The backdrop of the AI discussions was the export controls imposed by the Trump administration on Anthropic’s Fable 5 and Mythos 5 models, which went into effect on 12 June after the government cited national security concerns.

News sites have reported Anthropic is currently in negotiations with the Trump administration over the directive.

Financial Times (FT) reported the US decision alarmed European governments and Silicon Valley alike, while raising concerns even allied nations could be cut off from cutting-edge AI.

French President Emmanuel Macron stated the Anthropic dispute clarified what was at stake, warning if the US could “from one day to the next turn off the switch” it would ultimately damage the multitrillion-dollar US companies leading the global AI race.

Macron called for stronger AI regulation and a platform for democratic cooperation to establish common standards. Indian Prime Minister Narendra Modi stated democratic nations needed access to frontier models to protect critical infrastructure, according to people familiar with the discussions.

The post Tech bosses urge G7 to back US-led AI group appeared first on Mobile World Live.

Consulting company Deloitte is set to launch an AI Studio at its London campus in partnership with Google Cloud, a co-innovation hub designed to support the UK’s transition towards agentic AI.

Opening in late July, the studio will serve clients across financial services, the public sector, retail, healthcare and life sciences, and technology, media and telecommunications.

Deloitte is pitching the facility as a place where production-ready agentic solutions are built and delivered. Innovation labs hosted at the studio will allow clients to prototype and validate agentic AI systems in as little as four weeks, the company stated, compressing time-to-market for complex deployments.

For its retail clients, the aim is to move beyond generative AI chat interfaces toward sophisticated systems capable of modernising the customer journey end-to-end.

In healthcare and technology, media and telecommunications, the studio will focus on how agentic AI can reshape traditional business models and daily workflows rather than simply augmenting them.

Alongside the studio launch, Deloitte announced an upskilling plan to train 1,000 members of its UK AI and data workforce on Google Cloud’s Gemini Enterprise. The certification programme is designed to ensure Deloitte’s engineers are equipped to implement Google’s most advanced agentic architecture, giving UK clients access to what Deloitte described as one of the largest pools of certified AI talent in the region.

“Our clients are looking for more than just productivity gains — they want AI that can take action and drive real-world business outcomes,” stated Hayley McKelvey, chief AI officer for Deloitte UK, speaking at the Google Cloud London Summit. “This is about moving from curiosity to a new era of autonomous business operations.”

The London studio is in addition to Deloitte’s existing AI Experience Centres in Egypt and India.

The post Deloitte teams with Google Cloud to open London AI Studio appeared first on Mobile World Live.

Alibaba Cloud launched a cloud region in France featuring two availability zones in Paris as part of a plan to launch a suite of enterprise-grade agentic AI services in European markets later this year.

The move marks the Chinese giant’s third European hub following infrastructure builds in Germany and the UK.

The fully operational region brings enterprise-grade cloud services to European customers spanning computing, storage, containerisation, networking, security, databases, and developer tools.

The launch extends Alibaba Cloud’s global network to 105 availability zones across 32 regions.

The France deployment builds on European countries seeking to host their own data centres and cloud regions instead of relying on US companies such as Google, Microsoft and Amazon Web Services.

Built with data sovereignty in mind, Alibaba Cloud stated the Paris infrastructure is designed to comply with European regulatory frameworks covering data privacy, cybersecurity and resilience.

Alibaba Cloud noted the new region in France comes as European demand accelerates across sectors including retail, software development, AI-powered creative platforms and sports tech.

The company’s suite of agentic AI services includes AgentRun, a one-stop development platform for AI agents; STAROps, an intelligent operations platform; and ACS Agent Sandbox, which provides hardware-level security isolation while reducing operational costs.

The company is also bringing Agent Security Centre, AI Security Guardrails 2.0, and Agentic SOC to the region. Together, the portfolio is designed to cover the full agent lifecycle from development and debugging through deployment and security.

“This expansion, alongside the introduction of our agentic AI services to Europe, aligns with our broader strategy to bring our full-stack AI+Cloud ecosystem to global customers as we enter the agentic era,” stated Alibaba Cloud CTO and president of international business Dr. Feifei Li.

Last year, Alibaba Group CEO Eddie Wu committed to building “full stack technology capabilities” across AI chip design, infrastructure, cloud services and model development.

The post Alibaba Cloud expands agentic AI push to France appeared first on Mobile World Live.