Latest news with #IvoIvanov
Forbes
3 days ago
- Business
- Forbes
From Humanoids To Driverless Cars: Why The Next Rung On The AI Ladder Will Require A New Kind Of Network
Ivo Ivanov is the CEO of DE-CIX. It's been a busy year for AI. So much so that the biggest tech and trade events on the calendar were dominated by it, from the Consumer Electronics Show (CES) in Las Vegas to Hannover Messe in Germany. But what sets 2025 apart from previous years is how AI is now being framed—humanoid robots capable of conversational nuance and problem-solving, autonomous vehicles that can react in milliseconds to changing road conditions and new devices that are engineered from the ground up to deliver AI as a core function rather than just a novelty. Blink and you'll miss it, but AI is now stepping out of the cloud and into the world. This illustrates the turning point we've reached in 2025. Automation is yesterday's news. These aren't systems designed to crunch data or follow routines; rather, they're machines built to engage, react and make decisions in real time, and they're already feeling the constraints of centralized AI infrastructure. For over a decade, we've fed ever-larger models with ever-larger datasets, assuming more compute would always mean better outcomes. But inference—the act of using AI, not just training it—demands more than processing power. It demands presence and proximity. And that demand is colliding with a system designed for throughput, not immediacy. The future of AI isn't gated by model size or memory capacity. It's gated by latency—the new bottleneck, and the new benchmark, for AI in motion. Why Latency Is The New Currency For years, the AI conversation has revolved around training. Where to run it, how fast it can be done and how much GPU horsepower it will need. It's led to a boom in hyperscale data centers, sprawling training clusters and chip innovation that's breaking new ground almost monthly. But this emphasis on training has cast a long shadow over the other half of the equation: inference. Unlike training, which can be done in the background and often at a distance, inference is live. It's AI in action—detecting, responding, adapting—and it can't afford to wait. For a voice assistant to feel truly responsive, a self-driving car to process visual inputs and respond safely, a care robot to interact with its patients, the threshold of latency is in the single-digit millisecond range—and should move closer to zero over time. So, while an individual demonstration of the technology might offer a nice head rush, the reality of rolling it out (at scale) is far more sobering. As inference workloads move into physical environments, latency becomes a liability, a variable that can't be abstracted away with more compute. And the closer we try to align AI with human-level cognition and perception, the clearer the challenge becomes: milliseconds matter. The 'Triangle Of Inference'—And Its Fragility Behind every responsive AI system is a simple but unforgiving structure. Picture a triangle. On one corner sits the AI model itself, whether it's a language model parsing speech or a vision model interpreting road signs. On the other hand, there is the connected device: the robot, the vehicle, the drone or even the industrial sensor, triggering a chain of decisions. And anchoring the third corner is the transmission technology (fiber, 5G, satellite) that links them both together. In theory, it's an elegant system. But in practice, real-time AI only works when all three corners of the triangle are tightly synchronized. If the model is too far from the device, or the transmission link introduces jitter, or the compute node is overloaded, the system stalls. In this scenario, AI without immediacy is like a smartphone with a dead battery: It's intelligent, but for all its bells and whistles, it can't do what it's designed to do. This isn't a software problem that can be engineered out. You can have the best-trained model, the smartest agent and the most efficient algorithm, but if it takes too long to reach the device that needs it, the whole triangle collapses. Congested mobile networks, poorly placed infrastructure, edge devices hovering at the fringes of coverage—these are daily roadblocks to real-world deployment. What we're confronting is not a lack of AI capability; it's a systemic disconnect between compute, connectivity and context—and to solve it, we need to reevaluate how we design, deploy and optimize infrastructure. Location, Location, Location The Internet was never built for responsiveness or performance—it was simply built for reach. And for most of its history, furthering that purpose has made perfect sense. Web pages could take a few seconds to load. Streaming services could buffer. Even cloud apps could afford a few hundred milliseconds of delay. But AI has changed the stakes. Inference doesn't just need connectivity or 'reach'—it needs closeness. A response delivered a second too late might as well not arrive at all. This is where geography matters. Much of today's Internet traffic moves through what might be called digital 'flyover cities'—places where data passes through on long-haul fiber but has no way to break out locally. The infrastructure exists, but the interconnection points don't. That creates avoidable detours and latency penalties, even for traffic that originates just down the road. To keep pace with AI's demand for proximity, we need a new layer of infrastructure: small interconnection hubs distributed within 50 to 100 miles of users. These aren't full data centers or edge compute clusters—they're lean, neutral exchange points designed to keep data local when it makes sense and send it on its way smartly when it doesn't. Because if AI is going to live in our homes, cars, factories and cities, connectivity can't just exist in the usual tech hotspots. It needs to be wherever we are. Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
Forbes
08-07-2025
- Business
- Forbes
What Three Of The Biggest 2025 Tech Events Revealed About Connectivity
Ivo Ivanov is the CEO of DE-CIX. If 2024 explored the 'what' of generative AI, 2025 is exploring the 'how' of it. At the world's biggest tech gatherings—CES, MWC and Hannover Messe—the story wasn't just about new devices, dazzling demos or smarter machines. It was about what those innovations quietly demanded behind the scenes: a level of connectivity infrastructure that many businesses, governments and even industries aren't yet prepared to deliver. As the industrial-grade AI we're working toward becomes increasingly closer, the need for near-instantaneous data transfer across geographies is also becoming more acute. Life at the network edge, where these applications need to be, is making one thing crystal clear: Connectivity is no longer just an accessory to innovation. It's either the bottleneck or the breakthrough. Among the many aspects of connectivity, latency has emerged as the most visible constraint. It's also the least tolerable. Whether it's an agentic AI system designed to 'think' instantly about a customer query, a robot that relies on millisecond coordination across a production line, or a self-driving vehicle with sensors that are rendered useless without the network feedback to act on their output, a lack of 'responsiveness' is now a deal-breaker. Low latency is no longer a luxury for high-performance applications—it's the new baseline for the plethora of AI applications coming our way. This has been the breakout theme of some of the biggest tech events on the calendar in 2025; as we build smarter systems, we also need to reimagine the architecture that connects them. Not as a patchwork of nodes and endpoints, but as a living, scalable ecosystem designed for intelligent workloads and data transfer. What follows are three snapshots from CES, MWC and Hannover Messe that illustrate this turning point. Consumer Electronics Show (CES) 2025: Visible Expectations, Invisible Strain The year kicked off in Las Vegas, where CES 2025 went all out to demonstrate the remarkable impact AI is about to have on everyday life—from smart homes and cars to wearables and entertainment systems. Voice assistants now hold contextual conversations, while extended reality (XR) platforms deliver adaptive, AI-driven gaming experiences that blur the line between the real world and the virtual world. These applications rely on more than raw compute power—they demand fast, stable, low-latency connectivity between users, edge infrastructure and AI models running in far-off data centers, where even slight delays can dispel immersion or break the carefully modelled illusion of intelligence. The event also featured AI-enabled healthcare wearables, hearing aids and autonomous mobility solutions, all of which depend on dynamic, location-aware data. Yet most residential and public networks weren't designed for this level of complexity. The smarter the service, the more sensitive it becomes to latency and jitter. CES made one thing clear: The infrastructure behind our digital lives needs to evolve fast, or risk holding back the very innovations it's meant to support. Mobile World Congress (MWC) 2025: The Cloud Gets Crowded In February, Barcelona set the stage for another reality check—this time about where AI actually runs. At MWC, booths weren't just showing off new handsets or 5G upgrades, they were unveiling LLM-powered humanoid robots, 'sight-beyond-sight' vehicle-to-cloud software and the world's first official AI smartphone. But more than that, industry experts were talking about the ecosystem behind the tech: telcos, hyperscalers and AI developers jostling to define the infrastructure that will power AI at scale. What emerged was a clear trend toward decentralization. AI is no longer confined to hyperscale data centers. It's moving outward—to regional hubs, colocation sites and edge data centers and to a certain extent, into the devices themselves. The shift of focus from centralized training to distributed inference is now fully underway, and it's putting immense pressure on existing cloud and network architectures. As these ecosystems converge, so too do their limitations. Proprietary cloud environments and vendor-specific platforms were built for scale, not necessarily for interoperability. But real-time AI—whether for autonomous vehicles, smart factories or agentic assistants—requires seamless connectivity across providers, clouds and geographies. Without common standards and robust interconnection, distributed AI workloads will hit friction fast. Hannover Messe 2025: Industrial AI Has A Logistics Problem By April, the conversation had shifted from homes and handsets to factories and supply chains. At Hannover Messe, one of the world's largest industrial trade fairs, AI's potential was on full display—cognitive robotics, digital twins, autonomous systems and intelligent collaboration across factory floors and logistics hubs. But with every new demonstration came a familiar question: How fast can systems talk to one another? The most revealing takeaway from Hannover wasn't necessarily the sophistication of the AI models, but the infrastructural fragility beneath them. Industrial sites often span regions with wildly different levels of network maturity. Machines outfitted with advanced inference capabilities can only operate effectively if the data they depend on—sometimes hundreds of miles away—is delivered almost instantaneously. Add to this the rise of collaborative robotics and AI-managed energy systems, and the need for deterministic, ultra-reliable connectivity becomes mission-critical. Hannover Messe made the case that the intelligence of industry isn't limited by innovation, it's limited by distance, bandwidth and latency. Connecting The Dots Across three of the most influential tech events of the year, the message was consistent—the ability to move data quickly, securely and intelligently across the network of networks will determine which ideas scale and which ideas stall. CES showed us a near-future where consumer experiences become dependent on real-time, AI-powered interactions. MWC revealed how the infrastructure behind those experiences must interoperate in order to reach our AI goals. And Hannover Messe reminded us that when it comes to industrial automation, the stakes are higher, the environments are harsher and the tolerance for latency is virtually nonexistent. Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
Mid East Info
17-06-2025
- Business
- Mid East Info
UAE-IX Powered by DE-CIX Introduces 400 Gigabit Ethernet Connectivity, Strengthening Position as the Largest Internet Exchange in the Middle East
DE-CIX and datamena have upgraded the UAE-IX to offer 400 Gigabit Ethernet, enhancing its status as the largest and most advanced Internet Exchange in the Middle East and supporting the region's digital growth. Dubai, UAE, May , 2025: datamena, carrier neutral data centre and connectivity platform from du, the leading telecom and digital services provider, and DE-CIX, the global leader in Internet Exchange (IX) operations, announced a significant upgrade to the UAE-IX powered by DE-CIX, making it the first IX in the region to offer 400 Gigabit Ethernet (400 GE) access. This enhancement solidifies the UAE-IX's position as the largest and most advanced Internet Exchange in the Middle East, enabling unmatched connectivity and interconnection capabilities for enterprises, carriers, and content providers across the GCC and beyond. In the last year alone, connected customer capacity on the UAE-IX grew by two terabits, representing a 30% increase. With over six terabits of aggregate connected capacity and close to 110 participating networks—including Internet service providers, carriers, cloud platforms, content providers, and global enterprises—the exchange has evolved into a critical digital infrastructure hub for the region. Karim Benkirane, Chief Commercial Officer at du, said: 'We are proud to partner with DE-CIX in leading digital growth in the Middle East with the upgrade of the UAE-IX powered by DE-CIX to 400 GE access. It is our vision to foster a seamlessly interconnected landscape where businesses and consumers alike can benefit from unparalleled Internet exchange capabilities, heightened performance, and robust security. This milestone aligns with our commitment to maintaining the UAE-IX as a pioneer in interconnection and marks a transformative leap for regional digital ecosystems.' The UAE-IX stands out not only as a leading peering exchange but also as a provider of advanced enterprise-grade interconnection services, including cloud exchanges, cloud routing, and application connectivity such as the Microsoft Azure Peering Service (MAPS). Ivo Ivanov, CEO of DE-CIX said: 'The UAE-IX today stands as a global Internet hub, bringing together the network operators, content, applications, and cloud services to serve the entire GCC region with resilient and low latency connectivity. This upgrade further reinforces the importance of the UAE-IX, now ready to serve the rising demand for everything digital. The excellent collaboration with our partner du has enabled the UAE-IX powered by DE-CIX to shine as the most important aggregation point for network interconnection in the Middle East. I look forward to a bright future working together for the next decade of digital development.' DE-CIX, renowned for its neutral interconnection ecosystem and extensive global footprint, has been active in the Middle East for over a decade, with IX operations spanning Iraq, Jordan, Qatar, Turkey, and the UAE. The UAE-IX powered by DE-CIX in Dubai is operated under the DE-CIX as a Service (DaaS) model, offering a turnkey solution for partners—including carriers and data center operators—through comprehensive services such as setup, maintenance, provisioning, marketing, and sales support. This upgrade is set to empower regional businesses, foster technological advancement, and support the rapidly growing demand for high-performance, low-latency digital connectivity, paving the way for continued innovation and growth in the Middle East. About datamena: datamena is the UAE's fastest growing carrier neutral data centre and leading digital hub serving the Middle East and Africa (MEA) region. It is owned and operated by du, from Emirates Integrated Telecommunications Company (EITC). Carriers, hyperscalers, cloud and content providers and enterprises choose datamena because it provides a growing ecosystem of partners that enables them to meet, connect and do business to accelerate growth in Asia, Africa, Europe and the Middle East. datamena provides customer access to critical digital infrastructure and partners within a simple and seamless ecosystem to allow them to benefit from global cloud and content hubs by providing a seamless and high-performance data centre, connectivity and interconnectivity solutions dedicated to enabling local and global digital transformation and we're only just getting started. The simple, secure and open ecosystem has been awarded 'Best Internet Exchange Innovation' at the 15th edition of the prestigious Global Carrier Awards 2019, which took place in London, United Kingdom. About DE-CIX: DE-CIX, pronounced DEE-KICKS dˈi:-kˈɪks , is the world's leading operator of Internet Exchanges (IXs). Founded in 1995, the company is celebrating its 30 th anniversary in 2025. DE-CIX offers its interconnection services in 60 locations in Europe, Africa, North and South America, the Middle East, and Asia. Today, accessible from data centers in over 600 cities world-wide, DE-CIX interconnects thousands of network operators (carriers), Internet service providers (ISPs), content providers and enterprise networks from more than 100 countries, and offers peering, cloud, and other interconnection services. DE-CIX Frankfurt is one of the largest Internet Exchanges in the world, with a data volume of over 45 Exabytes per year (as of 2024) and close to 1100 connected networks. Close to 250 colleagues from over 35 different nations form the foundation of the DE-CIX success story in Germany and around the world. Since the beginning of the commercial Internet, DE-CIX has had a decisive influence – in a range of leading global bodies, such as the Internet Engineering Task Force (IETF) – on co-defining guiding principles for the Internet of the present and the future. As the operator of critical IT infrastructure, DE-CIX bears a great responsibility for the seamless, fast, and secure data exchange between people, enterprises, and organizations at its locations around the globe.
Zawya
06-02-2025
- Business
- Zawya
Ooredoo & DE-CIX bring world-class Internet exchange to Qatar with Doha IX
Qatar - Ooredoo, in partnership with DE-CIX, a global operator of carrier-neutral Internet Exchanges (IX), has officially announced Doha IX powered by DE-CIX, Qatar's first standalone commercial Internet Exchange (IX). Leveraging DE-CIX's extensive global expertise, developed across nearly 60 locations worldwide, this initiative strengthens Qatar's position as a regional digital hub by enhancing connectivity, reducing costs, and delivering exceptional customer experiences. Doha IX will offer a secure, carrier-neutral platform that facilitates low-latency traffic exchange, improves network performance, and supports remote peering services. Businesses in Qatar and across the region will benefit from cost-effective, direct access to global and regional content providers, streamlining connectivity through a single port while significantly reducing traditional IP transit costs. Doha IX is built on DE-CIX's cutting-edge interconnection technology and Ooredoo's state-of-the-art data centre infrastructure, supported by both partners' strong relationships with global content providers and networks. These critical assets ensure seamless and efficient traffic exchange, reducing latency and optimising network performance, as well as creating a robust IX ecosystem in Qatar. 'We are proud to introduce Doha IX, which represents a significant step in upgrading Qatar's digital infrastructure. Doha IX delivers faster, more reliable connectivity while reducing operational costs for businesses and driving innovation across various industries, aligning with the goals of Qatar's National Vision 2030 and advancing our digital transformation initiatives,' said Thani Ali al-Malki, chief business officer at Ooredoo Qatar. Ivo Ivanov, CEO of DE-CIX, said: 'With Doha IX powered by DE-CIX, we are bringing DE-CIX's global expertise to Qatar, enabling businesses and networks to benefit from superior interconnection services. Doha IX is the ideal place for international networks interested in reaching this important Middle Eastern market. 'The new IX, established through the partnership between DE-CIX and Ooredoo, will unleash the potential of the country's digital economy by providing better performance and user experience of content and applications, and affordable and high-quality Internet access for enterprises and individuals.' He added: 'This partnership marks an important milestone in strengthening regional connectivity and creating an advanced digital ecosystem that supports economic growth and innovation in the GCC for the amazing digital decades ahead of us.' DE-CIX is an established name in the Middle East, with a proven track record of growing healthy IXs and vibrant interconnection ecosystems. Doha IX, which will be built and operated under the DE-CIX as a Service (DaaS) model, is the sixth IX operated by DE-CIX in the region. Through this collaboration, Ooredoo and DE-CIX are setting the foundation for advanced interconnection in the region. Together, they support Qatar's digital transformation goals and position the country as a leader in the global digital economy, aligning with Qatar National Vision 2030. © Gulf Times Newspaper 2022 Provided by SyndiGate Media Inc. (
