Competition for Digital Infrastructure Deals is 'Really Positive'

Interview with Mike Armstrong of GI Partners

Mike Armstrong first joined GI Partners in 2014, and remembers that when digital infrastructure assets went up for auction, only about five bidders would show up. Fast forward to today, and Armstrong, a Managing Director and Head of Technology and Science Investments, says between 20 to 30 bidders compete for assets in a sale process.

Armstrong tells Cool Vector there are positives to the increased competition, such as pricing validation. In the meantime, he says “you have to be sharp” in order to stand out from the many new entrants to the market.

Key Takeaways:

• Cap rates are the signal to watch: in a market flooded with new capital and product, where cap rates settle will reveal the ecosystem’s collective confidence in future growth.

• AI’s full economic impact remains unknowable, but the next 12–24 months will be decisive in determining what it means for data center deployment and investment returns.

• Competition has exploded — bid sheets that once had five names now carry 20 to 30 — making sharp conviction and proprietary relationships the only sustainable edge.

---

Transcript has been edited for punctuation and clarity

Mike Armstrong: I’m Mike Armstrong, managing director and head of Technology and Science Investments at GI Partners.

GI Partners has been investing in the digital infrastructure space for a quarter century, the founding firm behind Digital Realty Trust, and we’ve been active in this space from operating OpCo investments to real estate investments. That’s where I spend all of my time.

So really from every angle we’ve attacked this space for, again, over a quarter century.

Cool Vector: What are you looking for in the pricing of digital infrastructure assets?

Mike Armstrong: I’m very focused on pricing, and GI Partners finds data center insights in cap rates.

With the scale of data centers that have been built and delivered over the past five years, I will be watching where pricing goes, very specifically cap rates.

Always looking for that next marginal transaction that provides insight into the health of the marketplace.

There’s been sustained growth for quite some time and pricing is one indicator to suggest our collective confidence as to the growth to come. Really low cap rates imply our collective view that people see a lot of growth going forward. Higher cap rates suggest less growth, and just with the size and scale of product that has come to the market, that’s just a meter and a metric as to the collective health of that broader ecosystem.

The more folks chasing deals buys cap rates down, and that’s certainly a part of the story of data centers of the past half decade — a lot of new participants in the market, more people chasing the same deal. And so that pushes cap rates down.

If your cash flows have more growth in them, you’ll take a lower rate of return today because, over time, you’re gonna make a little more money. The inverse of that — obviously, if there’s not a lot of growth, you kind of need a yield going in that will reflect the risk you see in your deal over the life of the deal.

Cool Vector: What are the greatest unknowns in digital infrastructure at the moment?

Mike Armstrong: Oh, wow. So that’s a big one. The greatest unknowns — look, part of the greatest unknowns is AI’s full impact on the economy is “unknowable today.”

There’s a lot that’s just unknowable today that will be relevant and impactful and affect the realized returns for the broader ecosystem.

We are all participating and watching what’s going on with AI, and how it’s impacting the space. I think there is a lot of optimism that AI is going to be here to stay. I think we’re gonna learn a lot over the next 12, 24 months and beyond about exactly how AI will exist in the broader economy, and then what that means for data center deployments.

And again, what that means for us and how we invest our capital.

Cool Vector: How would you describe the interest that investors have for the digital infrastructure asset class?

Mike Armstrong: In the space right now, there is a lot of anxious curiosity. Folks wanna be in the space. If you’re in the space, you want a larger position. If you’re not in the space, you wanna be in the sector.

But there’s some pretty fundamental dynamics at play and at shift across the ecosystem. And so cautious optimism is maybe one theme that’s emerging across conversations.

My sense is that most people’s investment committees are asking them, if they have data center positions and if they don’t, how do they get them? And if they do, how do they increase those positions? And then that same group of people is also saying, “I’m reading about artificial intelligence.” And in some circumstances people will mention the bubble word. And so we’re all kind of in this and leaning into this, and I think everyone has a long view, but there certainly are some considerations in the near term that we’re thinking through.

Cool Vector: What makes you bullish to be investing in digital infrastructure at this stage in the growth cycle?

Mike Armstrong: I like to distill things down to their simplest form. And I would say where my bullish perspective comes from is my own use patterns and how me and my family, and our business, how we’re using technology on an everyday basis.

Because really when you boil it down, data centers are a product. AI is a service that is serving the broader economy. So I usually, in moments of reflection, think about my own increased usage of technology as part of the data center story. The way we consume media, the way my home is effectively now a physical manifestation of the internet of things.

We are using more technology at an increasing rate every day. And when I think about it in that very simple lens, it actually provides a bit of a compass.

Cool Vector: How does GI Partners maintain an edge in deal sourcing?

Mike Armstrong: My entirety of my ecosystem is real estate related. If you replace investment bank with brokerage shop, then I can give you some insight — relationships drive deal sourcing in data centers.

What I tell my team is you have to walk and chew gum at the same time.

A lot of transactions come through traditional sourcing channels, come through the large brokerage houses. And we’re going to look at those. We’re going to move those to action in circumstances where we have conviction. But that can’t be the whole story. Where we develop an edge and where I think GI can have a proprietary advantage is through relationships.

It’s quite simple. Just being out in the market, talking to people, starting to piece together your own ideas that lead to deals — that’s where the real advantage is. It’s not all that complicated, but it’s just being out talking to people and trying to distill what you hear into actionable ideas.

Cool Vector: As a veteran investor in digital infrastructure, what is GI Partners’ view on the increased competition flowing into the space?

Mike Armstrong: My perspective is nothing’s all good or all bad. I think we all can appreciate that. And the benefit of having more market participants is there’s more validation in pricing. Ultimately, long term, in any spot transaction someone may overpay for something or do something you don’t necessarily understand or agree with.

But I think long term, having more market participants is a really, really positive thing for our ecosystem. I think in the short term it provides competition. What that does for any individual player — and certainly for GI, who’s been in this space for quite some time — is you have to be sharp. You have to make sure you’ve got really strong conviction in your decisions, in the deals that you move to action, because you’ve got someone on your coattails and the margin of safety has decreased quite a bit, with just more folks playing and nipping at your heels.

The data center bidders keep multiplying.

I started at GI Partners in 2014, and a typical marketed transaction — the bid sheet — would have maybe five people on it. So we would be one of five people chasing a deal, and very often that would be one of three. Fast forward to today, to 2026, that bid sheet is probably 20 or 30 people on a marketed transaction. There are just many more organizations, and many of them very well capitalized, chasing the same product.

And I’m sure I have colleagues who’ve been doing this longer than I that can tell you extraordinary stories predating 2014. Digital infrastructure, and very specifically data centers — there just weren’t a lot of folks that seemed to have this high on their target list.

I think that speaks to just how critical and how central digital infrastructure and data centers specifically have become to our broader economy.

I think COVID obviously was a moment that helped accelerate our broader understanding of exactly what was happening in these facilities and just how important they were. My guess is we’re still in middle innings with artificial intelligence. Digital infrastructure is here to stay.