FEATURED ARTICLE: AI Convergence, Power Constraints, and the Next Consolidation Wave
By Bernard Borghei Founder, President and CEO of Connected Infra Group
After more than 32 years in telecom infrastructure – co-founding and scaling several of the largest private tower companies, participating in some of the largest wireless infrastructure transactions, and seeing more than a few “next big thing” predictions – I’d like to think I’ve seen it all. Today, even as I build out Connected Infra Group’s national telecom infrastructure portfolio, I spend a lot of time listening to investors, operators, developers, and policy stakeholders.
When it comes to wireless connectivity, demand is always there, but I keep hearing four themes across towers, fiber, and data centers that are shaping the sector’s next chapter.
1) AI-driven convergence is no longer theoretical
For years, we treated towers, fiber, and data centers as adjacent categories with different underwriting playbooks. 5G, low latency, and higher bandwidth requirements started the first wave of convergence – a trend being further accelerated by AI. Latency and location are increasingly important to AI applications, pulling compute closer to users and network edges. That means that real estate is increasingly strategic, and sites with land, access, and expansion optionality that can support additional uses and functional evolution over time are increasingly valuable. The ability to bring compute closer to the end-user will be tested as newer models get rolled out and consumer adoption increases. Underwriting will also increasingly intersect with energy access, permitting, and industrial development timelines.
2) Power is becoming the limiting factor
The data center world is colliding with power generation and local politics and in some regions, demand for power – not compute – is the limiting factor. And when large new projects are perceived to raise utility costs for residents, the pushback is immediate. Pennsylvania is one example of the direction this is heading: policymakers want the data center investment because they see where AI is going, but they’re steering development toward locations with existing gas and water access – and increasingly expecting data center builders to bring their own power generation and delivery solutions.
This changes the math. Towers are often underwritten to a five- to seven-year return profile; but power generation projects can require massive upfront spend and a much longer runway before returns show up. For investors deploying billions up front, that means execution matters more than ever, but it also means that fuel access and water are becoming new moats to future competition. Stakeholder alignment also can’t be an afterthought when power prices and grid capacity are politically sensitive.
3) Consolidation dynamics are building across the stack
After several years of valuation whiplash, I think the industry is entering a restructuring phase – less “land grab,” more ownership consolidation driven by financing realities and the need for reliable domestic cash flow. Industry chatter around tower M&A and take-private scenarios show that the strategic logic is obvious, but the regulatory and capital-markets path can be complicated.
Capital structures are driving outcomes, and we’re seeing who can refinance, who can raise, and who can hold through volatility matters as much as a portfolio’s underlying asset quality. At the same time, we’re seeing public vs. private valuation gaps distorting deal math. Public tower multiples have compressed, while private deals can still clear at higher pricing.
A disciplined and consistent approach to underwriting is key when evaluating private deals. In a market with only three main customers (Verizon, AT&T, and T-Mobile) the current expected multiples in the private deal segment don’t really align with the growth opportunities. Yet, I continue to see surprising offers on average quality assets. The recent inflationary data and uncertainty around the interest rate direction could once again make equity and debt more expensive. That would have a serious negative impact on firms that are in the market to raise funds or secure debt. Would that force a reset on the private deals’ multiples?
4) Satellite Connectivity is Bypassing Telecom Infrastructure Entirely
Another trend impacting wireless infrastructure for both carriers and infrastructure players is the growth of direct-to-device (D2D) connectivity that enables ordinary consumer devices to directly connect to a low earth orbit satellite without going through a terrestrial antenna. Once viewed primarily as an emergency communications capability for stranded hikers or a failsafe for when normal cell service dropped, it is emerging as a compelling way for carriers to close rural coverage gaps and expand connectivity without the cost and timeline associated with building additional towers in hard-to-reach areas.
I’m hearing from industry colleagues that this new dynamic is changing the economics of coverage expansion as carriers might choose satellite-based approaches instead of investing in fiber routes, small cell deployments, or macro towers around the edges of their coverage maps. Terrestrial infrastructure players will need to watch this sector closely and ensure their investment and underwriting approaches account for the growth of the technology.
My takeaway:
AI demands are accelerating convergence, but power and capital discipline will determine who can execute. If you control the right real estate, navigate power availability, and maintain a balanced capital plan, you’ll have more strategic options than peers forced to chase the market.
I’m curious what others are seeing – especially around power interconnect timelines for data centers and the resurrection of edge compute. Is power generation and delivery now as important as spectrum, fiber, or real estate availability? Is edge architecture finally a real thing or still a fringe and soon-to-come idea? Is the growth of D2D changing your terrestrial investment strategy?
Contact:
