Investment into digital infrastructure has gone into overdrive this year, alongside the continuing growth of AI. DigitalBridge Group, a longtime investor in this sector with $108 billion AUM, recently closed a $11.7 billion fund. Yet, even for established players, raising capital in this crowded market isn’t straightforward.
Leslie Golden, the global head of capital formation and investor relations at DigitalBridge, sat down with PitchBook to talk about how the firm is positioning itself in the face of increased competition and navigating AI hype.
The following interview has been edited for length and clarity.
PitchBook: What was the fundraising market like while you raised DigitalBridge Partners III, which launched in 2023?
Golden: This was a record capital formation event for DigitalBridge across fund and co-investment commitments. Another record for us is how early in the fund’s life we received co-investment commitments. We expect co-investment participation will continue to grow as we deploy capital, similar to what we saw with DigitalBridge Partners II.
That said, it has not been an easy two years of fundraising generally. Many LPs have been overallocated to private markets—in particular, infrastructure. About a year ago, we also saw more entrants into the market. So, while we had a strong re-up rate, many of those allocations were a bit smaller than in prior vintages—a fundraising trend we are seeing across the industry. Fortunately, investor interest in our strategies has remained high. More than 65% of commitments came from existing investors.
Who are the newcomers entering this market, and how does DigitalBridge set itself apart?
Newcomers to the digital infrastructure sector are primarily existing infrastructure GPs that had been generalists but are now making larger allocations to digital infrastructure—in particular, data centers.
DigitalBridge has always focused entirely on digital infrastructure. We have been in the digital space for over 10 years. Many of my partners have owned and operated data center platforms or tower platforms, so we bring deep operational expertise, which enables us to build, as opposed to buy, for greater value creation.
We offer a robust co-invest program and bring co-investors in at the very start of an investment. Many large funds limit LP access to direct co-invest. That’s why our fund is sized the way it is—we’re typically able to offer at least a one-to-one ratio of fund to co-investment.
Recent News
Meet the 300+ investors who bet on Anthropic before everyone else
Anthropic’s historic IPO is the AI industry’s validation moment
What opportunities are you targeting with DigitalBridge Partners III?
The fund has already committed about 40% to 50% across five assets, two of which are hyperscale and AI data centers—Vantage Data Centers North America and Yondr Group. Those are larger positions in the portfolio. We also have a tower platform, a digital outdoor infrastructure platform and a fiber investment. We’re looking to build a diversified portfolio across the digital infrastructure ecosystem.
There has been demand at data centers, at our hyperscale platforms, as well as at our private cloud and edge platforms, driven by cloud deployment. Over the past year or so, AI demand has driven increased capacity needs.
We’re not the only game in town, by any means. This is a very large market, and this is going to be a very large opportunity for many years to come. We think our portfolio companies are uniquely well-positioned because of their access to secured power, land and capital. We’re able to enter very large contracts with our counterparties and give them the assurance that we can deliver capacity to them.
Today, AI training continues to drive substantial demand for massive, centralized data centers. But as the technology matures, the demand is expected to shift to inference workloads, which will likely spark the need for smaller data centers that are near population hubs for minimized latency. How does DigitalBridge prepare for this?
We are very discerning about the contracts that we enter, and the capacity we build is multipurpose. When we build AI training data centers, they tend to be near cloud availability zones so, if necessary, the data center could be repurposed from serving AI loads to serving cloud loads. We don’t want to be left with a single-use asset when we reach the end of a lease—whether it has a term of 10, 15 or 20 years. Therefore, while there’s some opportunity for us to build out on the training side, we need to be very cautious to ensure we are not overbuilding and we’re not taking on re-leasing risk.
Additionally, DigitalBridge supports several data center platforms that focus primarily on developing edge data centers. These edge data centers are where the majority of inferencing will take place, and DigitalBridge has invested in a global portfolio focused on this segment of the data center market.
What are the major challenges currently facing the data center market?
The biggest challenges are access to capital and access to power. For the industry, if we continue at the same pace of demand growth, power will become—and already has become—a gating item. We’ll need to continue to be very resourceful in terms of securing power.
Access to capital is also a gating item. The debt markets have been very open, particularly to hyperscale and AI data centers. We’re seeing a lot of financing, both in the bank market, [asset-backed securities] and other capital markets. At the same time, we’re starting to see more stringent underwriting as lenders become more cautious around deals with higher complexity or perceived risk. At some point, we might reach a limit, but we’re not there yet.
Latest Research
Where to Allocate in Consumer AI: Robots, App Development, and Multimedia Asset Generation
Learn about our editorial standards.