I’m excited to announce Erez Sverdlov has joined Totogi (where I’m founder and CEO) as Chief Revenue Officer. As VP of Cloud and Network Services for Europe at Nokia, he led a major segment of their €3 billion global portfolio. He’s also held leadership roles at Amdocs and Optiva—30 years of being in the room with Tier-1 executives, solving real problems, not pitching slides. So why leave the big-vendor world for a startup? Because Erez has seen what we’re building at Totogi, and he knows AI is going to fundamentally reshape telecom. He wanted to be part of a team that’s actually delivering AI at scale—solutions that drive real business value, not just demos and proofs-of-concept that never make it to production. He’s leaving the safety of the established path because he believes this is where the real action is. Welcome to the team, Erez! I’m so excited to have you with us!
Telecom’s vendor ecosystem is hollowing out—and it’s mostly happening through distress, not strategy. In 2025 alone: Mavenir quit manufacturing radios after a debt restructuring, EchoStar is selling $23 billion in spectrum and decommissioning its network under crushing debt, NEC classified its basestation business as “non-core,” and the RAN market shrank from $45 billion to $35 billion in two years. BSS is following the same pattern: Optiva, struggling for years, sold to Qvantel; Matrixx, also in financial trouble, went to Amdocs. These aren’t strategic acquisitions to take competitors off the table; they’re fire sales. The incumbents aren’t getting stronger; they're just absorbing the wreckage. Here’s what the consolidators are missing: AI is about to reshuffle the entire deck. The vendors who survive the next five years won’t be the ones with the biggest installed base. They’ll be the ones who figured out how to rebuild telecom systems as AI-native from the ground up. That’s exactly what we’re doing at Totogi. While the old guard plays defense and buys distressed assets, we’re building the telco tools that actually belong in an AI-first world.
AWS has the kind of autonomous network telcos dream about: one of the world’s largest fiber networks with 97-98% of events handled autonomously, no network operations centers (NOCs), and humans only involved for setting high-level policies and performing physical repairs. “The scale of our network—it would be terrifying to try to think of how you would have humans in the loop,” says AWS VP of Core Networking Matt Rehder. The secret? AWS built its own network devices and software from day one, avoiding the multi-vendor API nightmare telcos are stuck with. Google Cloud operates similarly and expects full backbone autonomy by the end of 2025. Meanwhile, 70+ telcos have signed TM Forum’s manifesto committing to Level 4 autonomy by 2025-2027—but they’re going to have to solve the problem of fragmented OSS/BSS and heterogeneous legacy infrastructure to get there. If only there were something that could help you solve your semantic inconsistency so you could build new tools and workflows more quickly. 🤔
“Context is king for AI agents,” says BOX CEO Aaron Levie. As AI expertise becomes commoditized—you can rent GPT-4 or Claude for pennies—the companies that win will be the ones who can capture, organize, and operationalize their internal knowledge to give AI the right context. Box is betting big on this with $1 billion in revenue and 64% of the Fortune 500 already on their platform. Here's why telcos should pay attention: you’re sitting on some of the richest customer data on the planet, but it’s trapped in systems where “subscriber” means seventeen different things depending on which database you’re querying. Your AI initiatives will keep failing expensive pilots until you solve the semantic chaos underneath. Context isn’t just king—it’s the whole kingdom. And right now, most telcos don’t have one. Manage your context using Totogi ontology. It gives your AI the guardrails it needs so you can scale AI with confidence.
Google launched its new Universal Commerce Protocol (UCP), an open-source initiative co-developed with Shopify, Target, Walmart, and 20+ other partners including Visa, Mastercard, and Stripe. The insight here isn’t the checkout feature, but that Google figured out AI agents can’t do anything useful in commerce without a universal semantic layer connecting all the fragmented systems underneath. This is another example for telcos to follow—but for your own systems. Build an ontology that unifies your core systems, and suddenly AI can actually scale, letting you build powerful new capabilities on top of what you already have, just like Google’s doing with e-commerce.
Clues for the clueless: Ericsson says that AI demands a new kind of wireless network! Ya think? Ericsson joined the Linux Foundation’s new Agentic AI Foundation (AAIF) as a Gold member—the only wireless RAN vendor among companies like AWS, Google, Microsoft, Anthropic, and OpenAI. The foundation governs open source projects including Anthropic's Model Context Protocol (MCP), which standardizes how AI agents connect to tools and data, and had 97 million monthly SDK downloads in its first year. My read: Ericsson doesn’t know what to do with AI, so they’re joining a club to find out. But here’s the thing. If you’re supposedly the world leader in wireless networking, shouldn’t you be leading on agentic AI for networks, not showing up as a second-tier member after the hyperscalers and AI labs already set the agenda? Talk about being late for school.
Starlink has become a crucial lifeline for Iranian protesters during the country’s most severe internet blackout in history. With an estimated 40,000-50,000 terminals in the country and SpaceX waiving subscription fees, it’s one of the few ways information is getting out. Iran’s response has been aggressive—jamming that degrades connections by 30-80% depending on location, plus door-to-door searches where possession can mean prison. Russia tried to jam Starlink in Ukraine and largely failed. Iran is having mixed results. It’s a stark reminder that satellite connectivity isn’t just about rural broadband anymore. LEO constellations have become geopolitical infrastructure, networks that can bypass authoritarian control entirely.
Are you ready for the Winter Olympics? TIM has partnered with Milano Cortina 2026 as the Official Telecommunications Partner, providing ultra-fast fiber and 5G connectivity across all 15 venues. The scope is massive: upgrading mobile networks at airports and stations, deploying over 10,000 mobile devices and routers, and cabling all Olympic venues with two separate telecommunications networks connecting sites across Milan, Valtellina, Veneto, and Trentino-Alto Adige. It’s the equivalent of running multiple world championships simultaneously across mountain terrain—not exactly easy infrastructure. Will TIM match Orange’s performance at Paris 2024, where it handled 100TB of data per hour during peak events and earned what analysts called a “connectivity gold medal?” The Games start February 6. We’ll be watching…
Anthropic’s new Cowork feature brings Claude Code’s agentic capabilities to non-developers—and they built it in about a week and a half using Claude Code itself. For $100-200/month, any knowledge worker can automate complex multi-step tasks like document creation, expense reports from receipt photos, and file organization. No terminal, no code, just natural language: this is the AI democratization moment telco leaders should watch closely. The productivity transformation isn’t just about your developers anymore. Tools like Cowork are reshaping non-technical roles just as fast as software ones. Before you hire your next analyst or coordinator, maybe give Cowork a shot to do the work instead. You might save yourself the hassle of hiring altogether.
Speaking of hiring, C-suite exec and consultant Dawid Makowski presented his “new team of employees” with a recent post on X: a marketing director, IT specialist, sales manager, operations manager, and senior HR specialist. The photos show each one is a variation of himself—and they’re all named Claude. It’s funny, but it’s also real. This is what I’ve been saying: AI isn’t coming for jobs someday. It’s already here, quietly filling roles that used to require headcount. The smart executives aren’t debating whether AI will transform their workforce. They’re already running leaner teams and pocketing the difference. If you’re still budgeting for the same org chart you had in 2023, you might want to ask yourself how many of those seats could be filled by Claude, Claudette, Claudine, or even Jean Claude!
