A Bloomberg graphic mapping AI financial interconnections went viral. It shows NVIDIA investing up to $100 billion in OpenAI, which has committed to buying millions of NVIDIA chips, while inking a $300 billion Oracle deal for data centers filled with... more NVIDIA chips. (Here’s the paywalled Bloomberg article.) Morgan Stanley noted OpenAI represents $330 billion of the $880 billion in future contracts tied to Microsoft, Oracle, and CoreWeave—meaning 66% of Oracle’s and 40% of CoreWeave’s growth depends on OpenAI’s stability. The parallels to Cisco’s 1990s vendor financing—lending money to telecom equipment buyers before the dot-com crash—are hard to ignore. There’s real technology solving real problems here, but when your customers are also your investors and suppliers, one failure could cascade. For telcos doing joint ventures with NVIDIA, investing in OpenAI/Anthropic, or building sovereign AI data centers: when you buy this infrastructure, you’re also buying into the circular financing web. If OpenAI implodes from pending copyright litigation, your AI partnership strategy implodes with it. Diversify your AI vendor relationships now, or prepare to explain to your board why your big bet was on a house of cards.
Google announced a $15 billion investment over five years to establish its first AI hub in India—its largest investment in the country and biggest AI hub outside the US. The 1-gigawatt facility will feature Google’s custom TPUs, fiber-optic networks, and partnerships with AdaniConneX and Bharti Airtel. This signals what's coming globally: sovereign AI data centers will proliferate as countries demand local infrastructure. If you're a telco exploring this space, partnering with hyperscalers who actually know how to run massive data centers at scale makes more strategic sense than jumping into bed with NVIDIA and figuring out operations as you go. Hyperscalers have been doing this for decades—let them handle the infrastructure complexity while you focus on connectivity and localization.
The AI-will-destroy-the-planet narrative needs an update. Wharton professor Ethan Mollick notes Google reports AI energy efficiency improved 33x in a year. Training models still burn power, but per-prompt usage is now on par with a 2008 Google search. It’s a similar story with water: US data centers use 50-628 million gallons daily, which Mollick points out is “a lot less than golf.” We’ve normalized watering desert fairways for recreation while panicking about infrastructure generating trillions in economic value. If you’ve been holding off on AI implementations because of environmental concerns, maybe you’re using sustainability as cover for transformation anxiety. The efficiency curve is bending fast, with per-prompt costs dropping while your legacy systems still guzzle power running batch jobs from 2015. Stop waiting for perfect environmental optics and start asking whether delaying AI adoption is costing you more than the electricity ever will.
Verizon abruptly replaced CEO Hans Vestberg with Dan Schulman effective immediately, shocking analysts who expected Consumer Group chief Sowmyanarayan Sampath—who’d been doing investor conferences as the presumed heir—to get the job. Schulman takes over with a mandate to “revive growth” at a carrier that’s “fallen behind rivals” and must “redefine our trajectory.” Here’s the lesson for telco execs: you can't play it safe, avoid bold decisions, and expect to keep your job. Leadership means taking risks and making changes, not status-quo management while competitors pass you by.
Bermuda-based Paradise Mobile announced its Net Promoter Score reached +47 in Q3 2025, more than tripling from +15 at launch in early 2024. The operator credits the network expansion, simplified pricing, and improved support it has done at the request of subscribers. Not only is this solid progress, but +47 beats the global telecom average of +31. And Paradise still has runway. Apple—considered best-in-class for customer experience—scores +61 to +72. Could it be that customers reward carriers who actually listen to them instead of forcing Byzantine pricing and IVR hell? Paradise proves you don’t need massive scale to improve. Just stop making things complicated and start fixing what customers complain about.
Former FCC Chairman Ajit Pai commemorated a recent milestone: On October 13, 1983, Ameritech Mobile launched America’s first commercial cellular service from Chicago’s Soldier Field. Early adopters paid $3,000 for handsets plus $50 monthly and 40 cents per minute for daytime calls. Adjusted for inflation, that’s roughly $9,750 for the phone and $162 monthly in today’s dollars—and that’s before making a single call. Back then, cellular was genuinely innovative and priced for early adopters who needed mobility enough to pay these insane prices. Fast forward 42 years: we’ve achieved true democratization with nearly everyone able to afford a phone and plan, but we’ve also become completely commoditized. The innovation premium is long gone. You’re now competing on price and experience alone. Paradise Mobile gets it. How do you compare?
