Dominic Jainy stands at the forefront of the technological evolution currently sweeping through the telecommunications landscape, possessing a rare depth of insight into how artificial intelligence, high-capacity fiber optics, and advanced wireless networks are merging into a single, cohesive ecosystem. With a professional background rooted in the complex intersections of machine learning and infrastructure, he has closely followed the industry’s pivot from the heavy, physical legacy of the 20th century toward a streamlined, software-defined future. In this discussion, we explore the strategic overhaul of AT&T as it aggressively strips out its historical copper foundations to prioritize a dual-fabric approach of dense fiber and “kick-ass” 5G wireless. We delve into the massive cost-saving measures that are funding this transition, the nuances of the “convergence innings” where services move beyond simple bundles into seamless technological management, and the pragmatic reality of satellite integration as a niche but vital tool for the final two percent of connectivity.
How is the transition away from legacy copper infrastructure fundamentally reshaping the operational DNA and financial health of the company as it moves deeper into 2026?
This is a massive, almost Herculean undertaking that involves stripping away decades of physical history to uncover what has effectively become the fifth largest copper mine in the United States. By retiring these legacy systems, the firm is doing much more than just cleaning up a balance sheet; it is systematically eliminating approximately $6 billion in legacy costs that have weighed down operations for a generation. You can almost feel the physical weight being lifted as they close down old central offices, terminate ancient leases, and finally silence the massive, power-hungry chillers that once hummed away to cool rows of aging mainframes. Every ton of copper pulled from the ground and monetized helps fuel the transition toward a “kick-ass” nationwide wireless network that will define the next decade of connectivity. The agility that comes from this clarity of purpose is palpable, as the company sheds the “other stuff” it has been hauling around since the mid-20th century to emerge as a streamlined, hyper-efficient metropolitan fiber provider.
With the aggressive targets set for fiber expansion, how do you see the roadmap to 60 million locations by 2030 impacting market share and household profitability?
The momentum is undeniable when you look at the progression from the low-20 million range for fiber locations just a few years ago to the 30 million locations we see today, which was bolstered significantly by the acquisition of Lumen’s consumer fiber assets. Now that we are in 2026, the company is laser-focused on hitting that 40 million location milestone by the end of this year, serving as a critical bridge toward the ultimate goal of 60 million by the end of the decade. This growth isn’t just about digging trenches; it’s a sophisticated combination of organic builds, strategic acquisitions, and innovative open-access ventures like Gigapower with BlackRock. By putting these specific assets into play, they are seeing a superior offering where higher household revenues and better margin performance are the natural results of lower churn and stable acquisition costs. It’s a winning formula because the physical build-out is relentless, and we won’t see the footprint wrap up until the latter part of this decade, allowing the company to ride that growth chain all the way.
You’ve mentioned that the industry is currently in the “early innings” of convergence; what does the shift from simple service bundling to deep network management actually look like for the end user?
We are moving past the era where convergence just meant getting a discount for putting your internet and cell phone on the same bill, though that initial phase certainly helped lower churn and stabilize the customer base. The real excitement begins now, as we enter the middle and later innings where the focus shifts to managing different access technologies as a single, seamless connectivity proposition. Imagine a world where your device transitions between satellite, wireless, fixed broadband, and Wi-Fi without you ever noticing a flicker in quality or a drop in your session. This is the promise of products like OneConnect, which simplifies the experience into a single billable relationship while ensuring that the “always-on” nature of the connection is backed by a sophisticated software fabric. It moves the conversation beyond price points and into the realm of a product that simply works better, providing a consistent experience that justifies the premium nature of the unified infrastructure.
In the enterprise sector, how is the focus on “advanced connectivity” and cloud access changing the way the Fortune 300 and mid-market businesses interact with their network providers?
While the company admits that its effectiveness in the enterprise space once lagged behind its consumer triumphs, we are now seeing a “second act” that is much more sophisticated and attuned to the needs of modern business. There is a concerted effort to move beyond the Fortune 300 and become more relevant to the mid-market by extending distribution channels and getting fiber into the specific locations where these businesses operate. This isn’t just about providing a pipe; it’s about allowing a customer to manage quality of service and security across multiple layers of the network, whether a session starts on a mobile device or a fixed office connection. Businesses are currently forced to think differently about how they gain access to the cloud to run AI-driven tools, and they need a partner that can support these high-bandwidth use cases without the friction of traditional telco hurdles. We are seeing low single-digit growth here now, but the confidence is high because the capabilities for this deep, multi-layered management are just about ready to be fully unleashed.
The decision to return to the data-center interconnect (DCI) market seems strategic yet cautious; how does this support the core business without becoming a distracting “mainstay”?
The re-entry into the data-center to data-center and metro-interconnect space is a tactical move designed to support the primary mission of getting traffic to end-user customers with the lowest possible latency. It’s about being “back in the game” in specific spots where it makes sense to build routes for their own traffic, particularly as AI engines in the cloud require massive, reliable pipes to connect to edge locations in metro centers. However, there is a very deliberate effort to ensure DCI remains a supporting player rather than the mainstay, a strategy that sets them apart from competitors who might be more heavily leveraged in that niche. By building enough interconnect capacity to drop traffic into any cloud provider’s infrastructure, they ensure that fixed and mobile traffic reaches the right workloads with high reliability. This is especially important for maintaining the “symmetrical” performance of fiber systems, ensuring that as AI-driven workloads grow, the underlying network can handle the specific demands of both the uplink and downlink channels.
With the recent $23 billion spectrum deal, how critical is the engineering of the “uplink” for future use cases like autonomous robotics and real-time video processing?
The massive investment in 30 MHz of 3.45 GHz mid-band spectrum and 20 MHz of 600 MHz low-band spectrum is a game-changer for the capacity and reach of the 5G network. By November 2025, the deployment had already reached nearly 23,000 cell sites across 5,300 cities, and the focus now is on using that low-band spectrum with engineered uplinks to support the next generation of industrial applications. In the wireless business, the upstream is becoming just as important as the downstream, especially when you consider autonomy, robotics, and upstream video processing that requires analysis in the cloud. Using Ericsson’s 600 MHz dual-band radios with 8RX capability allows for a much more robust connection that can handle these high-demand tasks even in challenging environments. It’s about fine-tuning the network to handle asymmetrical loads where the device is sending as much data as it receives, which is essential for the “meat and potatoes” of a modern, AI-integrated economy.
There is a lot of “noise” regarding satellite-to-device connectivity; why do you believe terrestrial fiber and wireless will continue to handle 98% of use cases?
While satellite connectivity is a valuable and niche complement, the reality is that the vast majority of mobile use cases are solved by the robust infrastructure that has been meticulously built over the last twenty years. There is a certain physics to this that cannot be ignored: a signal’s “hop” from a device to a local cellular antenna is significantly shorter and more efficient than a 300-mile journey to a satellite in low-Earth orbit. Once a signal hits the fiber, nothing else can beat its speed and reliability, which is why terrestrial networks will always be the backbone of the “kick-ass” experience people expect in high-traffic areas like stadiums or high-rise buildings. The goal of the planned three-way joint venture with T-Mobile and Verizon is not to replace terrestrial towers, but to ensure that the “last two percent” of coverage—those remote gaps where infrastructure simply doesn’t exist—is handled seamlessly. It’s about providing that 100% always-on feeling, but acknowledging that the heavy lifting will always be done by the fiber-dense networks on the ground.
What is your forecast for the role of internal AI in telecommunications over the next few years?
My forecast is that we will see a shift from the “frontier model” hype toward a very pragmatic, high-return application of AI that focuses on the “meat and potatoes” of business efficiency. We are already seeing incredible success in AI-assisted call center flows where customer satisfaction is high because the effectiveness of the interaction has been fundamentally improved. Over the next few years, the real value will come from using open-source models and last year’s chips to handle high-volume tasks like churn forecasting, dynamic customer offers, and software development productivity. By being smart about apportioning workloads—saving the expensive, high-end frontier chips for only the most complex tasks—telecom companies will realize massive returns without overspending on infrastructure. This disciplined approach to AI will allow the network to become “self-healing” and more responsive to individual user needs, ultimately making the converged fiber and wireless experience feel more like a bespoke service than a utility.