Ouch. My previous employer, Optiva, aka Redknee, has issued a Going Concern notice to investors as its $100M+ of debt is due in July and it doesn’t have a way to pay it back. Buried on page 22 of its most recent Management Discussion and Analysis document it states, “The Company’s ability to continue its operations is dependent upon its ability to refinance this debt or implement other financial alternatives including other sources of financing through debt or equity, however there is no assurance that this will be successful. These factors indicate the existence of a material uncertainty that may cast significant doubt on the Company’s ability to continue as a going concern.” I’m sure the team has been scrambling to ask ANYONE to help refinance the debt, current investors, banks, VC, PE, hedge funds, heck… probably even uncles and aunties… but to no avail. Remember, at one point the Board of Directors had a CAD $60/share offer on its desk that it REFUSED. Will Optiva secure financing before the due date? Will the current investors let it roll over the debt? Will it go bankrupt? I guess we’ll find out soon enough… 🍿🍿🍿
Well I guess the pilot worked, because Telefónica Germany has opted to run even more of its 5G network in the public cloud. It’s now targeting 20% of traffic on the AWS platform, proving that network workloads CAN run on hyperscaler infrastructure. When faced with scaling challenges, AWS stepped up by customizing Outposts for its needs. Meanwhile, Ericsson keeps pushing its full-stack approach that keeps operators locked in and paying premium prices. You guys know who my money’s on… You guys know who my money’s on… ⛅️ ⛅️ ⛅️
SoftBank and Tech Mahindra are making waves with their new "large telco models" (LTMs)—specialized LLMs trained on network data using NVIDIA's AI platform. While this is a strong move (SoftBank's model cuts base-station configuration time from days to minutes with 90% accuracy!), I'm not 100% sold. Generic models miss what makes YOUR telco unique. What telcos need is an ontology-driven approach that models your specific network, assets, and processes. The perfect solution? Use both! Use an ontology to capture your telco's processes, then power it with an LTM (and other models) to scale. That way, you get AI that understands YOUR unique business challenges, not just generic telco patterns. That’s how you get value out of AI. Want an ontological approach to AI? Try out Totogi’s BSS Magic!
Liberty Global CEO Mike Fries sees a bright future for AI in telecommunications, but says two big challenges are preventing the industry from adopting it today: hiring the right talent, and being willing to take the risk. I couldn’t agree more! But I have a plan to address both, Mike. Have you checked out Totogi’s BSS Magic and PlanAI? These two solutions bring huge, AI-driven improvements to your organization and bottom line. What are you waiting for? And as for AI talent, I’m building it myself by being a part of GauntletAI. The first cohort graduates this week, and will be joining Totogi in April! Want to see them in action? Book a meeting with Totogi and we’ll show you how we can build mobile apps in two hours.
Love is in the air when it comes to the satellite market. Bharti Airtel has signed an exploratory deal with SpaceX's Starlink to bring satellite internet to India's 560+ million unconnected citizens. Meanwhile, Comcast Business has become the first major network provider to integrate Starlink into its Managed Connectivity portfolio, joining other telcos like AT&T, Verizon, and T-Mobile in leveraging LEO satellites for expanded coverage. And New Zealand's 2degrees, which has an existing Starlink partnership, is inking a deal with AST SpaceMobile to bring space-based cellular broadband to New Zealand consumers. (The Starlink service is still live as connectivity for businesses.) As I said in my recent blog, The Wolf of MWC, partnership with satellite providers can definitely deliver benefits, but operators need to be aware that partners may become competitors in the future. Best to enter into these agreements with eyes wide open and a clear Plan B.
SK Telecom is keeping its options open. It just enhanced its AI personal assistant, adot, by integrating Google's Gemini 2.0 Flash, bringing its supported AI models to a whopping 12. The assistant now provides real-time Google search results with source page navigation, improving content transparency. Recent additions also include Perplexity’s Sonar Pro and ChatGPT o3-mini, alongside existing models like SKT’s A.X, Perplexity Sonar, ChatGPT variants, and Claude 3.5 models. Tons of choices for subscribers… I wonder how the subs use the different models? If you know, drop me a line…
Why do AI data centers require so much power? Here’s a nice explainer from RCR Wireless that details five main reasons. This massive demand is why Big Tech is turning to nuclear power to fuel AI, which I discussed with Charles Fitzgerald from Platformonomics in a recent Telco in 20 podcast episode. Check it out!
Does anyone feel sorry for Ericsson? Probably not. But maybe the telco network equipment gravy train is slowing down. The equipment market crashed by 11% in 2024—its worst decline in two decades according to Dell'Oro Group—with Optical Transport, SP Routers, and RAN plummeting 14% each. While the vendor hierarchy remains mostly intact (though Huawei leapfrogged Nokia outside China), this market tumble perfectly illustrates my long-standing point: until MNOs stop burning bajillions on new networks without clear monetization strategies, this downward trend will continue. Smart players are already flipping the script by embracing OpenRAN and the public cloud to slash costs while enabling new revenue streams through AI and automation. Don’t know how to do it? Give me a call!
