I was in telecom at the height of its prestige. 4G LTE was popping off, cashflow was everywhere, and our teams were all smiles. Around the tail-end of 2017, however, some new players in the wireless infrastructure industry started to make the news. Tillman Infrastructure, CitySwitch, Harmoni Towers, and several others began building new towers for the carriers to help alleviate the escalating cost of infrastructure. Meanwhile, carriers were negotiating hard with the large tower owners to try and lower rents and get some relief.

It was a leading indicator of a troubling truth: Wireless carriers had begun to see their infrastructure cost model as unsustainable. The boom of 4G was topped off. All Americans had phones, were using them to their max capacity, and the promise of 5G bringing a fresh wave of revenue began to show signs of disappointing.

How does this impact data center teams? Maybe not at all in the immediate future. But if we don’t take stock of the lessons learned in wireless, which in so many ways is a sister-industry of data centers, we may repeat the same mistakes.  

Wireless learned that eternally escalating costs made version-1 of the shared infrastructure model unsustainable.  

The data center market is currently experiencing significant growth. The global data center market was valued at USD 301.8 billion in 2023 and is projected to grow at a compound annual growth rate (CAGR) of 10.1% during 2024-2030, reaching USD 622.4 billion by 2030 (Web ID: 10).  

The demand is there and seems unlimited, but we know from the wireless boom that there will be a plateau. How we plan for that plateau will separate winners and losers.  

Canary in the Coal Mine

I remember asking a member of our C-Level team personally: “Do you see Tillman and some of these other BTS tower companies as a threat?” The answer was, with a lighthearted chuckle, unequivocally, “No.” And they were right. In a sense.

American Tower, Crown Castle, and to a lesser extent, SBA Communications, had a huge head start. A couple hundred towers being built here and there wasn’t going to put a dent in the armor of a $6.53B annual tower revenue player (per Crown Castle’s earnings release FY2024). But there was another threat that these build-to-suit tower owners were silently indicating:  

A canary, singing her heart out in a coal mine.

The carriers could not grow revenue at a speed that matched the growth pace of their infrastructure costs. The rental agreements in place between the carriers and tower owners have fixed escalators. Every year, without fail, over the 10-30 year life of the agreement, rent goes up. Back in the early 2010s, that seemed like a non-issue to carriers. AT&T, Verizon, and T-Mobile all divested their tower portfolios seeking the savings from shared infrastructure. However, when the revenue boom flat-lined, suddenly those 1.5-5% escalators started to hit like the approaching orc armies in Lord of the Rings.  

“This is far beyond any of you!” - Gandalf

In response, the carriers began building again. Not out of desire but out of necessity. This gave rise to the many great tower providers we’ve seen gain a foothold in American infrastructure in the last 15 years. This also hurt the largest infrastructure providers: their revenue, their relationships, and their long-term outlook.

Lessons Learned for the Next Technology Infrastructure Frontier

The data center boom is just beginning. I really believe that. And the optimist in me thinks there’s no end in sight! Growth for decades!  

Wireless tells us a different story.  

In 2013, no one was talking with any certainty about the revenue-growth plateau carriers would see in the next decade. No one was concerned about the financial health of behemoths like AT&T. No one thought these fixed escalators within the shared infrastructure models would be anything but a benefit for all parties. We were all mistaken.

There are lessons we can take from an industry reaching its maturity. These lessons can be applied to an industry just getting out of the gate.

1. Keep a pulse on the financial health of the hyperscalers.

It’s crazy to even imagine that these behemoths, who in some cases hold cash equivalent to a medium-sized country’s GDP, could ever be wounded. Amazon, Meta, Apple, and the like are too big to fail, right?  

It’s the same thoughts wireless teams had around the carriers. AT&T, Verizon, and T-Mobile are still profitable and certainly have a long life ahead of them, but they could not afford to do nothing when they see their most important line of business being trash-compacted by rising costs and declining revenue growth.

Hyperscalers such as Amazon, Microsoft, and Alphabet (Google) account for over 60% of global hyperscale data center capacity, making their financial health critical to the entire market.  

The rise of artificial intelligence (AI) is driving unprecedented demand for data centers, with significant investments projected in AI-ready infrastructure over the coming years.  

What will the hyperscalers do when the AI boom finds its ceiling? They will undoubtedly do the same as the wireless carriers. They will try to negotiate a better long-term deal. If data center operators hold out and say, “check the contract language”? There will be opportunities for new players and a new game.

2. Strive to be a business partner, not a landlord.

Business partners come to the table. They creatively work together to find a mutually beneficial solution. Landlords are content to enforce the terms of the contract without considering the broader business context.  

In the wireless industry, the large tower owners’ reluctance to renegotiate terms led carriers to seek alternatives, including building their own infrastructure. Data center operators should learn from this and prioritize collaboration over strict contract enforcement.

3. Embrace flexible pricing models to adapt to market changes.

The wireless industry’s fixed escalators in rental agreements became a burden when revenue growth slowed.  

Data center firms often strike incredibly similar terms with their customers. Sure, there’s some flexibility in things like power consumption costs, but underneath that is a fixed escalator that beats a drum every year.  

This may be fine for the first terms of their contracts, but where will the hyperscalers be in 2035? 2045? If wireless as a sister-industry is any indicator, there will come a time when those shared infrastructure costs pose a serious risk to these AI giants.  

Data center operators should consider more flexible pricing structures that can adjust to these inevitable market conditions. This could include performance-based pricing, revenue-sharing models, or periodic renegotiation clauses that allow for adjustments based on market realities. Some providers are already working under these partnership dynamics.  

By learning from the wireless industry’s experience, data center teams can avoid similar pitfalls and build stronger, more resilient partnerships with hyperscalers.  

How long before AI and data center industries reach maturity? Impossible to say, but until then the canary keeps singing.