Ron Haberman

I was honored to deliver a keynote at this year’s Marconi Society event today at UCLA. I opened with how control (AI), compute, and connectivity technologies each began in the 1980s as monolithic, centralized, and specialized silos, gradually evolving over decades toward distributed, composable fabrics. Today, they are undergoing a deeper transformation toward “distributed convergence,” where these three layers increasingly operate as a "distributed, composable, integrated, and intelligent fabric". Control (AI/ML): symbolic, rule-based AI → neural networks → deep learning → foundational models/LLMs → multi-agent systems Compute: mainframes → PCs → client/server → virtualization → central cloud → distributed cloud (edge) Connectivity: 1G–2G (circuit-switched) → 3G (hybrid) → 4G (packet-switched) → 5G (cloud-native) → toward 6G (AI-native) The next decade will be defined by distributed convergence, where control (AI), compute, and connectivity function as one cohesive system. This is the essence of the AI-native era. Building on this foundation, I reflected on the connectivity progression: We advanced from circuit-switched 2G, to IP-native 4G that unified transport for voice, video, and data, and then to cloud-native 5G, which transformed networks into programmable platforms for distributed workloads. We are now entering the AI-native era, where a new class of devices and applications demands a new class of networks, aligned to the dominant unit of computation: first packets, then workloads, now agents. This set the stage for the architectural continuum I discussed: - Packets move bits - Workloads run services - Agents pursue goals These layers are additive, not replacements. AI-native systems continue to rely on packet-switched substrates and cloud-native orchestration, while adding agentic and semantic capabilities on top. I framed this evolution as a progression from programmability → orchestrability → autonomy → sustainability, guided by new architectural constructs for AI-native networks , including a dual-stack model: today’s syntactic Internet alongside an emerging semantic Internet. My sincere thanks to my colleague, Sheryl Genco, Ph.D. and the Marconi Society for the invitation and for its leadership in bringing the community together to advance the future of connectivity for societal impact.

119

9 Comments

some great points brought up here - 448G is not just upscaled Ethernet - it must bring address the scale at bandwidth density that AI 'scale up and scale out' needs. We need new interconnect technology, careful focus on power and density and modulation and a solid cost base to make this hit the needs and potential of AI silicon.

11

1 Comment

Here is very interesting piece of news I wasn't expecting: Capella Space, an Earth Observation company merging with IonQ, a Quantum Computing company. If you weren't paying attention to satellite Quantum Key Distribution before, you should now. Together, they will provide both the sword and the shield: Quantum processing capable of breaking the asymmetric cryptography in use today and Global QKD to secure future communications & transactions. Not to mention the staggering benefits of Quantum Networking. In short, it's the birth of a powerhouse. Congratulations to both teams, how are you going to get these Photons through the atmosphere reliably ? Can we help ? ;)

10

At NVIDIA GTC 2026, Jensen Huang made something subtle but profound clear: the AI infrastructure race is no longer measured only in FLOPS or GPUs — it’s becoming a token economy. In the AI era, data centers are optimizing around three new core metrics: • Tokens per Watt – how efficiently infrastructure converts power into AI inference throughput • Cost per Token – the economic unit of AI, driven by platform advances from systems like Blackwell and the upcoming Vera Rubin architecture • Tokens per Second per User (TPS/User) – the new user-experience metric for AI responsiveness and real-time interaction This is fundamentally different from the traditional mobile operator economy, where networks evolved around: • Gbps capacity • Cost per GB transported • Latency and throughput as the user experience benchmark In simple terms: Telecom optimized for moving bits. AI infrastructure optimizes for generating intelligence. How telecom can benefit from the AI infrastructure economy paradigm.. Not all the bits on network are equal in value, yet they are charged the same. Will these two worlds converge As AI inference increasingly moves closer to the edge — integrated with telco clouds, RAN edge sites, and distributed compute — mobile operators may need to evolve from a “Gbps economy” to a hybrid “Gbps + Tokens economy.” That shift introduces new strategic questions: • How many tokens/sec can be served at the edge per cell site or metro edge? • What is the cost per AI interaction over a mobile network? • Can operators monetize AI responsiveness the same way they monetized bandwidth? The next phase of telecom infrastructure may not just deliver connectivity — it may deliver real-time intelligence as a service. And if that happens, the network is no longer just transporting data. It becomes part of the AI inference supply chain. #AIInfrastructure #GTC2026 #NVIDIA #Telecom #EdgeAI #AgenticAI

21

2 Comments