Today’s infrastructure financings rarely fit neatly into traditional legal or capital markets categories: What was once viewed as a relatively straightforward real estate or project finance story has expanded into the broader capital markets ecosystem.
Against this backdrop, 9fin sat down with S Alexandra de Padua (partner, Global Finance, New York), Sean Darling (partner, Global Finance, London), and Agnès Rossi (partner, Private Funds, Paris), of Proskauer Rose, to talk about the evolution of infrastructure debt. Proskauer’s infrastructure debt team advises capital providers on complex US, European, and cross-border infrastructure investments. Recent transactions include Ares’ $2.4bn financing for Vantage Data Centers, and lenders to nexfibre’s £2bn acquisition of Substantial Group.
We discussed the complexities of financing AI-focused data centers, the growing influence of secondaries markets, and how sponsors and lenders are adapting underwriting, restructuring, and liability management strategies as infrastructure assets become more technologically complex and operationally interconnected.
1. Tell us about Proskauer’s infrastructure debt initiative. Both of you came from a background advising clients on complex financing transactions. As you moved to Proskauer, what are you doing differently?
de Padua: Proskauer’s infrastructure debt practice starts from a fairly basic observation: our clients don’t think in terms of legal silos, so neither should we. To reflect that, our team deliberately brings together lawyers with different product backgrounds—real estate finance, leveraged finance, project finance and fund-level solutions—so we can assess a transaction holistically and bring the right expertise into the room from the outset. Increasingly, that also means helping clients rationalize capital structures that are becoming more layered and complex as the asset class evolves. In practice, that means we can engage earlier in the process to help shape the capital structure itself, rather than stepping in once it’s already set.
Darling: From a London perspective, what attracted me to Proskauer was the ambition to grow and to focus energies and resources on high value, niche and tricky-to-execute deals. Sourcing and structuring infrastructure debt in Europe today rarely involves a clean choice between “project finance” and “leveraged finance”—most of the interesting situations sit somewhere in the middle and, in that middle, there is complexity, nuance and the requirement for thoughtful lawyers who are able to understand the pedigree of both and deliver infrastructure financings.
2. At the loan level, what concepts are you borrowing from leverage finance or private credit when it comes to underwriting debt documents?
de Padua: I’d probably reframe that a bit, because I don’t think it’s simply a matter of terms being imported wholesale from one market (levfin) to another, rather, documentation terms for infrastructure debt deals are actively being shaped by competition between markets. Borrowers with genuinely defensive, asset-backed cash flows have the credibility to move between syndicated high yield, private credit, and project finance capital solutions. The result is a convergence of terms—project finance lenders in some cases accepting lighter maintenance covenant regimes, HY investors tolerating more structural protection than they’d typically demand—driven less by consensus than by each market’s need to compete for the same credits.
Darling: I’d add that this thoughtful, tailored approach is particularly visible in the European sponsor‑led market where we have multiple pockets of national banks and champions that produce a rich regional framework for financing and for how the assets can be delivered, taking into account local laws, economics and priorities. For those sponsors and lenders, the answer cannot simply be to copy‑paste leveraged loan terms into an infrastructure context as they are.
3. In terms of placing this in context of larger trends that we are seeing in the private investments funds space, how is the growth of the secondaries market changing infrastructure investing in private capital?
Rossi: Secondaries are no longer just about exits—they are becoming capital solutions embedded throughout the lifecycle of infrastructure investments. What we are seeing—quite clearly—is the importation of private equity-style liquidity solutions into a traditionally long-duration, yield-oriented asset class, but with infrastructure-specific adaptations.
One of the most notable developments is the increasing use of GP-led secondary transactions by infrastructure asset managers. These are becoming a core portfolio management tool rather than a last-resort liquidity option. In practice, they are being used in two principal ways.
First, to transfer mature, de-risked assets into core or yield vehicles better aligned with their long-term cash flow profile, addressing the structural mismatch between finite fund lives and very long asset duration. Second, to build or scale thematic platforms—aggregating assets within a defined vertical and bringing in new capital to support growth.
4. Asset managers are raising funds for AI strategies to build AI-focused data centers. What makes them fundamentally different from traditional data centers, both in terms of fund formation and financing?
de Padua: AI-focused data centers have characteristics that do not cleanly fit the traditional infrastructure playbook – to take only the most obvious examples, the counterparty mix is more concentrated and the underlying compute economics evolve faster than in traditional digital infrastructure. You can see that in how these deals are being structured – more staged capital deployment, tighter linkage to power availability, and a greater focus on flexibility around buildout and ramp-up. Managers raising AI strategies aren’t just raising another digital infrastructure fund. They are underwriting a more development-heavy, power-constrained, and operationally complex opportunity set, and that requires more flexibility in mandate and a greater focus on control over land, power, and interconnection.
At the fund formation stage, sponsors are addressing the issues raised by AI-focused data centers by leaning more heavily on (1) single asset funds that operate more like private equity than traditional blind pool funds in terms of providing preemptive rights and liquidity—potentially even governance for their most significant investors; single asset investment vehicles also lend themselves naturally to single-asset continuation funds, which dominated GP-led secondary volume in 2025 and are increasingly used in digital infrastructure and the data center sector and (2) data center-focused funds with substantial seed (or target) portfolios, which provide a competitive advantage over blind pools by giving prospective investors the opportunity to understand better their exposure to this asset class before making an investment decision (although seeded and semi-blind structures are now increasingly standard in digital infrastructure fundraising more generally, not unique to AI strategies).
5. How does restructuring actually work in practice when different parts of a data center—land, operations, GPUs—are financed separately?
Darling: Often these financings will be completely ring-fenced from each other, have recourse to and credit support from different entities in the capital structure and, as such, have limited direct legal crossover. However, given the operational overlay of the various asset classes, and the continuing evolution of different financing tactics in the space, it may well become a commercial imperative to deal with some or all of these arrangements in the event of a default in one.
As the financing markets for these assets evolve, a realistic whole asset / whole business understanding of what happens if and when deals might go wrong becomes even more critical. Do the documents allow for a coordinated sale of targeted assets? If so, what is the intrinsic value of such assets outside of the specific ecosystem? Are there information-sharing and standstill mechanisms that make a multi-creditor discussion workable in practice? These are the kinds of questions we are asking with clients at the structuring stage, not just when a restructuring is already on the table.
6. According to an AlixPartners report from last November, market participants were anticipating potential industry distress for data centers over the next five-10 years. What’s your take on that, and how are you preparing for this?
Darling: We still see significant appetite for investment in this sector both from a sponsor and creditor perspective. It is worth noting that the finance documents funding these transactions continue to include valuable lender protections both in the standalone space and in the group financing context. For example, the need for borrowers to meet drawdown conditions, extensive reporting requirements and the provision of more comprehensive security when compared with more standard leveraged finance documents all remain standard elements of the financing terms.
de Padua: I do think the AlixPartners report is picking up on an important underwriting question for the sector: are the durability assumptions that the capital is underwriting keeping pace as technology and demand continue to evolve? You’re deploying capital against infrastructure-style return expectations, but the nature of the underlying assets themselves is that they need to continue to evolve in complexity and capability to ensure their continued relevance in their market. The retrofit-versus-new-build question is really just a symptom of that—at some point you’re not simply refinancing an asset, you’re making a fresh capital allocation decision against a technology landscape that may have moved on since origination.
My sense is that, where we do see stress or underperformance, it is more likely to be asset-specific than a sector-wide reset. The harder question is how financing structures continue to evolve alongside what the industry itself is anticipating—and how that gets reflected in how deals are being structured today.
7. Is there a chance that an LME-type situation could occur in this space? If so, why?
de Padua: People draw comparisons between leveraged finance and infrastructure fairly regularly, and it’s worth taking seriously, but infrastructure debt transactions typically benefit from strong structural protections that constrain what sponsors can do with a capital structure in ways that don’t really exist in leveraged finance. Add a lender base that is predominantly institutional, long-dated, and relationship-oriented, and you’re missing most of the ingredients that make these situations possible in the leveraged finance market in the first place.
However, as competition drives terms migrate across high yield, private credit, and project finance you start to see a broader range of covenant packages and structural flexibility being introduced into deals that historically would have been more tightly constrained. To my above point about convergence of market terms - the question is whether, in combination, they introduce degrees of flexibility that weren’t necessarily intended in an infrastructure context.
Darling: I would also add that, from a commercial perspective, the analysis is not limited to the debt documents. In many infrastructure and digital infrastructure assets, you also have to consider whether offtakers, customers or other key contractual counterparties would be required to consent to structural changes, or whether those changes could give rise to termination or other rights. That adds a layer of complexity that is often less visible if the analysis is approached purely through a leveraged finance lens.
8. Proskauer has intentionally built out the practice as a global business. What unique challenges or innovations are emerging in non-US markets?
Darling: The short answer is that the fundamentals of good infrastructure probably don’t change all that much by jurisdiction, but the unique types of legal and business frictions do.
In some markets and in some assets you have more overt regulatory and political risk. Some investors have more FX risk (e.g. fleet wide railcar businesses with USD denominated debt and local currency revenues) while other deals are more exposed to inflation (e.g. through fixed price construction risks which hit contractors’ margins and working cashflows hard).
At the same time there is still a high degree of novelty in thinking about how infrastructure assets operate and how they should be financed and run – you only have to look at the original hyper-scale DCs financed in Europe in the early 00’s on standard REF terms vs the targeted flexibility that the group financings have now. New approaches criss-cross the Atlantic given that the underlying fundamentals of the assets remain the same but dealing with financing them evolves.
9. Outside of AI, what is the newest frontier of growth for infrastructure debt?
de Padua: Outside of AI, the opportunities I find most interesting are the ones that don’t fit cleanly into any single vertical. What’s sometimes called cross-vertical infrastructure—where energy, digital, and traditional asset classes are starting to overlap in ways that weren’t really possible five years ago.
The renewable energy and data center linkage is the obvious example—power access is increasingly the binding constraint on data center development, and that’s pushing investment decisions across both verticals simultaneously. It’s driving a much more integrated approach to underwriting and structuring, because the value of one asset is increasingly dependent on the other.
Darling: To underscore that, the “cross‑vertical” and the expanding definition of infrastructure is exactly where I see some of the most interesting credit work happening.
Whether that is hybrid structures combining different pools of capital, or documentation that blends elements of project, corporate and asset‑based lending, our job—particularly out of London, where we see a wide mix of sponsors and capital providers in a business centre that looks out to the world—is to help clients navigate that complexity without losing the discipline that has made infrastructure debt such a resilient asset class in the first place.
Reproduced with permission. Originally published May 20, 2026 by Yiwen Lu, "9Questions — S Alexandra de Padua, Sean Darling, and Agnès Rossi, Proskauer — Infrastructure debt at the crossroads," 9fin.
