All I want for Christmas are these 5 things from telcos

I just love Mariah Carey’s “All I Want for Christmas.” But instead of wishing for something under the tree, I made a different kind of list: five things I wish telcos would actually do in 2026.

The pattern across all five? Speed. Speed to evaluate, speed to build, speed to market. The telcos that win in 2026 won’t be the ones with the best strategy decks. They’ll be the ones that moved first.

Read the full post and pick one to start before the champagne goes flat on New Year’s Eve. Cheers! 🥂

Episode 130

Best of 2025: AI for networks and networks for AI

To round out 2025, we’re bringing back the year’s standout episode: my conversation with Gabriela Styf Sjöman, Managing Director of Research and Network Strategy at BT Group. Gabriela cut through the AI hype with a brilliant, two-question framework to guide technology decisions: does it help us make money or does it help us save money? We dive into BT’s dual AI strategy, why “best network” marketing is dead, and the cultural shift needed for telcos to become subscriber-centric. Don’t miss this, our 2025 Telco in 20 Episode of the Year!

LISTEN NOW: Apple Podcasts, Spotify, YouTube, TelcoDR website

I’m still riding high after winning the gold-medal game with my team Austin Ballers at the Minor League Pickleball (MiLP) DUPR16 nationals for the 50+ division on December 4th! Sadly, I missed The Great Telco Debate, but I think it was worth it. (Sorry TelecomTV!) If you’re looking for some scintillating viewing, check out our final match! 

The competition is a team format, where each team is made up of two men and two women, all over the age of 50, with a maximum combined dynamic universal pickleball rating (DUPR) of 16.3. Each match between the teams consists of four games: women’s doubles, men’s doubles, and two rounds of mixed doubles, rally scoring, to 21. If the teams are tied after the doubles, you advance to a “dreambreaker” round where everyone plays singles. Our secret sauce? We practiced a lot of singles, so we totally crushed the dreambreaker 21-12. When the dust settles, I’ll be ranked in the top 10 in the country. How cool is that?!? 🥇

The Amazon Web Services (AWS) annual jamboree, re:Invent, ran December 2 through 5. I predicted it would be about big advancements in AI, sovereign data centers and chips. This year AWS fused “agentic AI” with next-gen cloud infrastructure—and rolled out some popular discounts. Here’s a quick summary of the highlights.

​In chip news, AWS launched Trainium3, its first 3 nm AI training chip, promising around 4x the performance and significantly better energy efficiency versus Trainium2. AWS also teased a future Trainium4 that will interoperate with Nvidia GPUs, signaling a hybrid strategy. Graviton5 was introduced as AWS’s most powerful general‑purpose CPU to date, targeting broad cloud workloads with better performance‑per‑watt than prior Graviton generations.

​

​In AI, AWS rolled out a number of new models in its Nova family, optimized for different tasks. Bedrock AgentCore gained richer “agentic” capabilities so developers can build multi‑step, policy‑constrained AI agents that maintain memory, respect guardrails, and operate over extended periods. AWS also highlighted “frontier agents” such as Kiro, an autonomous coding agent that adapts to team workflows and can work for long durations with minimal human intervention.

​​

For developers, SageMaker HyperPod added checkpoint‑less and elastic training, targeting faster, more resilient large‑model training with automatic recovery and scaling, among a few other improvements to speed AI development.

​

Database Savings Plans elicited an audience cheer. The new pricing model can help with cost efficiency across database services while retaining deployment flexibility. Duckbill’s Corey Quinn nails it: “Six years of complaining finally pays off.”

In exciting news for modernization, a new AWS Transform family of services can help “crush tech debt with AI-powered code modernization.” I like the sound of that! The tooling can learn an organization’s code patterns to automate refactors for legacy, Windows, and mainframe estates in record time.

For all the details, check out the full list of “top announcements” from AWS.

At Oracle AI World, Larry Ellison revealed that AI is writing “a lot of the code” at Oracle, positioning Big Red at the center of the AI revolution. He’s not alone: Microsoft says 20-30% of its code is AI-generated, Google claims 30%+, and Zuckerberg predicts Meta will hit 50% within a year. Microsoft’s CTO Kevin Scott goes further—95% of all code will be AI-generated within five years. If your IT team is still writing code line-by-line, you’re not just behind. You have an army of paid dinosaurs. 🦖

You don’t need clean data to use AI. Trevor Lee, CEO of Myko AI, speaks truth in a recent LinkedIn post: some AI projects require pristine datasets, but generative AI (GenAI) is different. It’s fluent in human language and can actually create clean data as an output. This misconception is a massive stumbling block for telcos, adding years to AI adoption timelines while teams chase data perfection that was never required. If your data is good enough to run your business today, it’s good enough for GenAI. Stop letting perfection annihilate your progress. The real blocker isn’t dirty data; it’s semantic inconsistency. That’s exactly what BSS Magic solves with its telco-specific ontology layer. 🤔

Totogi just signed a strategic deal with its first European customer—a major tier-one operator on the London Stock Exchange—using BSS Magic to deliver in 8 weeks. The AI-native automation handles 80% of configuration and testing while cutting operational costs in half. Legacy vendors built empires on the lie that BSS transformations require years of customization and armies of consultants—turns out that was never a technical constraint, just a revenue strategy. Pricing changes that used to take months now take days. The era of the multi-year BSS death march is ending, and the operators paying attention are already moving. 🚀

Trump Mobile’s T1 phone is MIA months after its promised launch, with customer service blaming—wait for it—a government shutdown for the delays. 😂 The phone was supposed to be “Made in America,” but that claim quietly vanished from the website. Ship dates have slipped from August to October to November to maybe December, while the company hawks refurbished iPhones and Samsung devices as “brought to life in the USA.” Turns out launching an MVNO is harder than it looks. There’s a reason we don’t have more 1M+ subscriber MVNOs running around: this business eats the unprepared alive.

Adios, Optiva! The Ontario Superior Court gave final approval on December 2nd for Finland’s Qvantel to take Optiva private. Shareholders get C$0.25 per share while noteholders receive 22.4% of Qvantel shares, $25 million in secured notes, warrants for another 3%, plus potential cash from collected receivables. At a quarter per share, this is a fire sale, not a strategic acquisition. Bet the board regrets blocking ESW Capital’s $60/share offer back in 2020—a 122% premium they called “extraordinary” before rejecting it. From $60 to $0.25 in five years. I guess I’m the golden goose after all! 🪿💰