Peru’s leading MVNE Guinea Mobile swapped its legacy on-premise BSS for Totogi’s cloud-native, multi-tenant platform to power seven digital brands including Cuy Móvil. The move cuts new MVNO launch times from months to weeks and deploys Totogi’s AI to actually drive subscriber growth and reduce churn—capabilities single-tenant, on-premise systems simply can’t deliver. This is what winning looks like: while competitors stay trapped in vendor prisons that require custom code for every brand launch, Guinea Mobile can now spin up MVNOs at internet speed and use AI to grow them. When your old vendor gets acquired by another struggling player, that’s usually a sign you made the right call to switch to Totogi!
Now you can buy stuff directly from ChatGPT through Instant Checkout—starting with Etsy and soon expanding to a million Shopify merchants, with OpenAI open-sourcing the Agentic Commerce Protocol (so anyone can integrate). The question for telcos: how fast can you enable your products to be sold via large language model (LLM) chats? Or are you still stuck trying to get your BSS applications to talk to each other after 18 months of integration hell? This is where legacy vendors have you trapped. While the world moves to conversational commerce, you’re burning cycles on basic system integration. Totogi’s BSS Magic can wire your systems together and get you selling in ChatGPT in weeks instead of spending another year trying to integrate your own stack.
OpenAI launched AgentKit at DevDay, a visual drag-and-drop toolkit that turns months of complex agent orchestration into hours. Fast is good. But speed without context is just expensive mistakes at scale. Tools like AgentKit accelerate AI workflows, but they still need deep domain knowledge to understand telco-specific data models, business processes, and vendor APIs—exactly what Totogi’s BSS Magic ontology provides. The winning formula isn’t just drag-and-drop workflows—it’s marrying tools like AgentKit with a telco-specific ontological layer that acts as the “universal translator” for disparate BSS and OSS systems. While you debate whether to build or buy AI agents, smart telcos are combining both: enterprise-grade tools like AgentKit for speed, plus domain expertise like BSS Magic for context. That’s how you ship AI that actually works.
Anthropic released Claude Sonnet 4.5, a new AI model designed for state-of-the-art coding performance that builds production-ready applications, not just prototypes. Early trials showed Claude autonomously coding for up to 30 hours, and handling complex tasks like setting up databases, buying domains, and conducting security audits. Claude Sonnet 4.5 scored 77.2% on the SWE-bench Verified coding test, beating OpenAI’s ChatGPT-5 Codex at 74.5%—a record that stood for less than two months. Other industries are jumping on these tools while telcos are waiting for their vendors to save them. (Newsflash: they’re struggling just like you.) The imperative here isn’t just about adopting technology, but also about how fast you can develop talent and adapt to the pace of AI innovation.
Vibe coding has only been a thing for a few months, but it’s already being replaced by “agentic swarm coding.” One guy used a swarm of Claude AI agents on a transatlantic flight to build a production-ready software platform in six hours—work that would normally take 18 developer days. These swarms, composed of specialized agents (planners, coders, reviewers), produce high-quality, secure, and fully documented applications, not just prototypes. At Totogi, we use swarm coding in BSS Magic to compress what used to take years into days—this is how we’re integrating legacy telco systems faster than the vendors who built them and creating full enterprise applications to boot. Swarm coding is fast, sure, but it also lets small, agile teams outship entire engineering organizations stuck in waterfall processes. Want to learn how to do this in your organization? Give me a call.
NVIDIA and OpenAI signed a $100 billion agreement (er, “non-binding letter of intent”) for OpenAI to deploy up to 10 gigawatts of GPU infrastructure, with Nvidia investing up to $100 billion in what’s essentially vendor financing. Some telcos see deals like this and think “we should do GPUaaS too!” without realizing they’re wading into a Martingale strategy against the world’s most valuable company and hyperscalers with infinite cash reserves. Your GPUaaS play will get crushed before it even launches. The REAL opportunity may be in monetizing enterprise LLM data needs: massive pipes, bulletproof redundancy, and rock-solid connectivity for businesses running on AI instead of humans. When every enterprise decision flows through an LLM, network reliability becomes existential, not operational. That’s a telco play worth making.
BT is developing AI that detects “agonal” breathing during emergency calls to help responders prioritize cardiac events, plus uses background noise for environmental context and categorizing non-emergency calls. Finally, a telco AI application that actually saves lives instead of just optimizing network efficiency! But the team is hitting integration challenges getting its systems to talk to each other—a pattern we see everywhere in telco. While BT’s innovation team has the vision to build life-saving applications, it’s burning cycles on basic system integration instead of shipping code. This is the tax legacy infrastructure imposes: great ideas stuck in integration hell while the clock ticks. Totogi’s BSS Magic exists specifically to solve this problem. It’s a layer on top of your existing systems, so there’s no rip-and-replace required. Just connect your systems, code your application, and go! Deploy innovation instead of filing change requests.
T-Mobile announced Srini Gopalan will become CEO on November 1, succeeding Mike Sievert, who becomes “Vice Chairman.” Currently the COO, Gopalan previously led Deutsche Telekom’s Germany business, doubling growth rates and scaling fiber. He’s spearheading T-Mobile’s push to become the most data-driven, AI-enabled, digital-first operator. I’ll be watching to see if Gopalan gets that modern telcos need to be tech companies and if he’s bold enough to drive the massive transformation required. Being AI-first means changing everything, top to bottom: how you hire, how you build software, how you compensate teams, how you make decisions. Does Gopalan have the courage to blow up what’s working today to build what’s required for tomorrow? In the age of AI, that’s the true leadership test.
Ahead of the UN General Assembly late last month, the US Secret Service uncovered a massive SIM farm—over 300 SIM servers and 100,000 SIM cards—in a hidden network across New York, New Jersey, and Connecticut, connected to Chinese actors. Concentrated within 35 miles of UN Headquarters, the devices had the potential to disrupt cell towers and eavesdrop on world leaders. But here’s the strategic question for telco leaders: 100,000 active SIM cards co-located in a 35-mile radius should be trivial to detect with AI and machine learning (ML) pattern recognition. Geolocation data, activation patterns, usage anomalies—this is exactly what ML was built to find. If the Secret Service can discover these networks through routine threat assessments, your fraud detection systems should be catching them in real-time. This is an opportunity for MNOs. We should be leveraging AI to identify these concentration patterns before they become national security incidents.
Qvantel is acquiring Optiva in what’s being called a “strategic transaction,” but let’s be honest: it’s a lifeline for Optiva. The combined organization of over 1,000 employees (!) promises a “full-stack, AI-enabled platform” serving 70+ operators in 40+ countries. The deal structure is brutal for minority shareholders like me (We get $0.25 per share versus the $60 per share ESW Capital offer that was shot down 2021.), while noteholders receive Qvantel equity, new debt, and warrants. (I collated the action in my second-ever newsletter's Moves in the Cloud.) If you’re an Optiva customer, get ready for massive change: new processes, new account teams, new pricing models, and the inevitable integration chaos that comes with vendor M&A. History suggests combining two struggling companies with legacy tech debt rarely produces the AI-powered innovation they’re promising. Customers, this is your signal to evaluate alternatives like Totogi before your business is dependent on their successful integration.
