Data centre securitisation: navigating a fast-growing asset class

Surging artificial intelligence (AI)-related capital expenditure (capex) is driving one of the fastest growing sectors in the securitised products market.

Combined issuance of asset-backed securities (ABS) and commercial mortgage-backed securities (CMBS) by data centre operators reached US$26bn in 2025, over 10 times more than in 2020 (see chart below).

This growth reflects a broader structural force reshaping the fixed income market – the AI supercycle – where demand for intelligence as a new factor of production is driving an extraordinary build-out of physical infrastructure. Data centre operators are amongst the direct beneficiaries: ‘hyperscalers’ including Microsoft Azure, Google Cloud and Amazon Web Services are investing US$700bn in 2026 alone to support their cloud and AI platforms. The scale of investment is driving substantial debt issuance.

The capex pipeline points to further acceleration in issuance. Permanent financing needs for data centre construction completions are estimated to hit US$150bn across 2026 and 2027, with a substantial portion expected to be funded through securitisation.¹ Annual issuance could reach more than US$40bn in 2026, which would represent up to 10% of total US ABS issuance.²

For credit investors, this expansion creates both opportunities and risks. Selectivity will be critical in determining risk-adjusted returns as technological obsolescence, structural complexity and refinancing risk shape the attractiveness of individual issuances.

Source: BofA Global Research, February 2026

Distinct structures and risk profiles

Securitisation affords data centre operators several advantages. Both ABS and CMBS can offer issuers competitive financing costs, off-balance-sheet execution and access to post-construction capital that complements earlier-stage financing from banks, private debt and project finance.

There are important distinctions to be made between the two structures, however. 

In terms of the structural characteristics, data centre ABS are backed by both the property interests and the corresponding lease cashflows generated by those properties. They employ a master trust structure that allows additional assets to be contributed to the collateral pool over time, with guardrails to prevent credit drift.

In contrast, data centre CMBS are backed by a mortgage loan on the underlying real estate properties. They are structured like typical single-asset, single-borrower CMBS transactions, where the collateral pool is fixed at issuance.

When it comes to risk, analysis of data centre ABS should emphasise cashflow diversity and durability and evaluate the degree to which cashflow triggers protect bondholders.

An important difference is that ABS are also structured as soft bullets – where the issuer is not strictly obligated to repay the principal on the scheduled maturity date – typically with a five-year anticipated repayment date. Data centre CMBS meanwhile are structured as fixed maturity instruments, and credit analysis centres on the borrower’s ability to refinance at maturity based on future property values and the sponsor’s willingness to recapitalise if valuations fall short.

Rising spread dispersion in a supply-constrained market

The underlying collateral across data centre issuances is becoming more homogeneous, as the shift toward larger, single-tenant occupancy accelerates.

‘Hyperscale’ operators are increasingly leasing entire large-scale campuses, concentrating the market – especially for new developments – around a narrower set of well-capitalised tenants. Existing multi-tenant co-location data centres offer more collateral diversity for investors, albeit with shorter contractual commitments.

Amid record issuance in the sector, spreads have diverged across data centre ABS issuers as investors grow more selective around tenant quality, asset diversification and operator expertise. Co-location ABS spreads are wider than that of hyperscaler-backed transactions.³

In the near term, several structural tailwinds help to mitigate credit risk. Power availability constraints and lengthy permitting timelines are limiting data centre construction, raising barriers to entry, while the capital intensity of new developments limit supply growth and protect existing operators.⁴ Tenant churn meanwhile remains low, as the cost and disruption of migrating data centre infrastructure impose meaningful costs. This incumbency bias, combined by low vacancy rates and high demand from cloud computing, supports rental income growth.

Four overarching risks

Despite the favourable technical backdrop, the asset class remains early stage and structurally complex. We believe that investors should pay attention to four risks.

1. Obsolescence risk

Credit differentiation in data centre ABS and CMBS will increasingly be driven by the strength and expertise of the underlying operator and the durability of its tenant base, as both navigate rapid technological advances.

Power, cooling, redundancy and network interconnectivity all determine a data centre’s competitiveness, and older facilities can degrade rapidly relative to evolving standards. Modern facilities can better support high-density compute and easily scale to liquid cooling and higher rack densities, whereas older data centres will face rising capex burdens to remain viable.

Tenant composition, concentration and creditworthiness also warrant close attention, particularly where a small number of tenants represent a disproportionate share of contracted revenue. This amplifies lease renewal risks and impacts the sustainability of cashflows to ABS noteholders.

2. Structural complexity

The master trust structure in data centre ABS means the collateral pool an investor underwrites at origination can evolve materially over the life of the investment. Deal terms, documentation standards and structural protections vary across issuers, and rating methodologies may continue to change as the sector develops.

Investor disclosure also remains limited: tenant identities are typically not disclosed and service level agreement language is often summarised via key terms only.

3. Refinancing risk

Unlike the amortising structures prevalent in most ABS sectors, data centre ABS concentrate the principal repayment at or around the anticipated repayment date, exposing investors to rollover risk at a point in the credit cycle that may not be favourable.

In effect, investors are underwriting the issuer’s ability to access capital markets at a specific future date. If spreads widen materially or investor appetite for the sector cools at the point of refinancing, issuers may face significantly higher funding costs or, in a stress scenario, struggle to refinance at all.

4. Secondary liquidity

While improving, the market has not yet matured to match the depth seen in more established ABS sectors. This could limit investors’ ability to exit or resize positions efficiently, particularly during periods of market stress. This dynamic is compounded by the large increase in anticipated issuance over the next two-to-three years.

Approaching a vast opportunity set, selectively

We expect recurring bouts of volatility as the data centre ABS and CMBS market grows and matures, introducing risks for ill-timed investments as well as creating compelling entry points for investors with the analytical infrastructure to distinguish between individual securities.

Within this space, we currently favour senior ABS tranches backed by long-term hyperscale tenant leases. These typically offer attractive relative value versus comparable investment grade corporate credit, with stronger structural protections, including multiple cash trap provisions that restrict borrowers’ access to free cash flow if debt service coverage and loan-to-value thresholds are not met.

We also identify some compelling spreads among multi-tenant data centre transactions – and against collateral that carries greater tenant diversity – as hyperscale ABS issuance has grown.

Meanwhile, the sustainability credentials of data centres are increasingly material to credit quality. Energy and water efficiency, as measured by power and water usage effectiveness (PUE and WUE), and renewable energy procurement, are fundamental drivers of operational resilience and cash flow stability. While efficiency gains and liquid cooling are reducing resource intensity per unit of compute, absolute power and water consumption continue to rise with AI-driven demand growth.

Data centre operators are therefore exposed to supply constraints and rising input costs. As carbon regulation tightens and hyperscale tenants integrate sustainability requirements into their lease terms, the consequences of poor sustainability performance are shifting from reputational to contractual and financial. Older or less efficient facilities face compounding risks of tenant loss, stranded asset valuation and refinancing stress that current spread levels may not adequately compensate for.

In an area where documentation standards are still evolving and structural nuances can meaningfully affect recovery rates, we believe expertise, rigour and selectivity will be crucial for credit investors.

¹ J.P. Morgan, August 2025
² BofA Global Research, February 2026
³ Morgan Stanley, January 2026
⁴ Grid capacity across most US markets is largely committed through 2030, making power access the primary site selection criterion. Construction costs have risen at a 7% CAGR to US$10.7mn per MW, with JLL forecasting US$11.3mn per MW in 2026. This figure does not include AI fit-out costs of up to US$25mn per MW

Nothing presented herein is intended to constitute investment advice and no investment decision should be made solely based on this information. Nothing presented should be construed as a recommendation to purchase or sell a particular type of security or follow any investment technique or strategy. Information presented herein reflects Impax Asset Management’s views at a particular time. Such views are subject to change at any point, and Impax Asset Management shall not be obligated to provide any notice. Forward-looking statements or forecasts herein are subject to known and unknown risks and uncertainties including inaccurate assumptions that could cause actual results to differ materially from those expected or implied by the forward-looking statements. . While Impax Asset Management has made reasonable efforts to obtain information from reliable sources, we make no representations or warranties as to the accuracy, reliability or completeness of third-party information presented herein. No guarantee of investment performance is being provided and no inference to the contrary should be made.