Jamie Dimon’s 2026 shareholder letter devoted extensive space to AI, and telco CEOs should read every word. JPMorgan is spending nearly $20 billion on technology, with a massive chunk going to AI. Dimon says AI adoption will be “far faster than prior technological transformations” and will affect every function, application, and process at the bank, eliminating some jobs. His advice: deploy small, empowered teams that move like Navy SEALs, not massive transformation programs. (Sound familiar?) Dimon also says there’s no bubble, arguing the capital flowing into AI reflects genuine transformation, not speculation. When the CEO of a $3.9 trillion bank says this isn’t a drill, believe him.
Anthropic announced a new, unreleased frontier model named Claude Mythos that has discovered thousands of high-severity vulnerabilities across every major operating system and web browser. To give makers of the world’s most critical software a jump on hackers, Anthropic launched Project Glasswing with Amazon Web Services (AWS), Apple, Cisco, Google, Microsoft, and Nvidia. For anyone running critical legacy code, this is a massive breakthrough. Every telco on the planet is about to get a tool that can audit and harden code at a scale and depth that was physically impossible before. If you’re a CTO haunted by decades of security debt, get ready to get to work. Amazing!
Was 2025 the year of AI agents? Industry consensus from SDxCentral and multiple analyst reports think 2025 was experimental, 2026 is still early, and agentic AI projects have been harder to deliver than expected. Why? Because of the semantic inconsistency in telcos’ underlying architecture: different BSS systems have different definitions of the same customer, causing hallucinations at critical decision points. Luckily, the Totogi Ontology was built to solve this problem, providing an overarching semantic layer that gives AI agents a single, consistent definition of every concept across your telco. Give it a try. (Full disclosure: I’m Totogi’s CEO.)
More telecom professionals (57.5%) believe AI offers major new revenue opportunities according to TelecomTV's latest DSP Leaders survey—up from 52% last year. That’s good news. Sadly, the rest of the respondents (42.5%) either doubt it’ll ever happen or have no idea how to get it done. That’s the story throughout the survey: belief in AI’s potential is growing, but there are still too many doubters. I was super sad to see that 57.5% of respondents think telecom networks are still too complex for agentic AI. I hate to say it, but these people are dead wrong. What will it take to convince them? Probably just one telco doing it, and everyone will follow.
On April 1, a missile struck Batelco’s headquarters in Bahrain, which hosts AWS infrastructure. In March, drone strikes took two of three AWS availability zones offline in the UAE. Cloud infrastructure is now a target in armed conflict. If you weren’t using the public cloud before, you should start now for disaster recovery (DR). It’s the only infrastructure that makes true geo-diverse recovery possible at speed and scale. When Zain Sudan lost both of its on-prem data centers, its incumbent vendor quoted six to nine months to rebuild. Totogi restored service to 20 million subscribers in 18 days on AWS in a separate geography. It’s time to reevaluate your DR strategy, as the threat matrix has dramatically changed.
Microsoft partnered with Armada to deploy Azure Local on ruggedized, modular data centers for industrial edge AI—think oil rigs, mines, factories, anywhere terrestrial connectivity is limited. The unit arrives self-contained, runs private 5G traffic locally, keeps all data sovereign and air-gapped, and plugs into the same Azure management plane as headquarters. No custom server rooms, no specialized cooling, no construction projects required. Haven’t operators been promising this kind of service for years? Enterprise customers won’t wait for operators to get it together. Speed matters. It’s time to up your game.
Telefónica just sold Movistar Mexico—20 million subscribers, 25 years of operations—for $450 million to a consortium led by Oxio, a cloud-native Telecom-as-a-Service (TaaS) platform. Telefónica has been operating as a virtual operator on AT&T’s infrastructure in Mexico since 2019—and still couldn’t make the economics work. Oxio is betting that its TaaS stack can extract margin where a $36 billion European telco group couldn’t. I believe it. If you strip out the legacy BSS/OSS overhead and run 20 million subscribers on a cloud-native, automated platform, the same revenue base becomes profitable. (I blogged about it in The modern telco.) Oxio has its work cut out running a massive MVNO AND a software company, but the bet itself is worth watching. 👀
Four astronauts aboard NASA’s Artemis II Orion just made a video call from 248,655 miles away. The Orion spacecraft is testing a laser communications system built by MIT Lincoln Lab that delivers 260 Mbps burst downlink from lunar distance—faster than the average fixed broadband connection in dozens of countries. NASA’s old S-band radio could return about 7 GB in an hour of contact; the new optical link returns 36 GB in the same window. If we can beam HD video from a quarter million miles away, we can certainly figure out how to run a telco on modern software, like software from Totogi!
