HSBC's reported sale of approximately $300 million in construction debt associated with Project Jupiter, a major AI data-center development in New Mexico, at prices in the mid-80s has brought an uncomfortable aspect of AI infrastructure financing into focus. The disposal, reported by Bloomberg on 16 November 2026, involved part of an $18 billion financing package arranged by approximately 20 banks in late 2025. Assuming an indicative transaction price of 85 cents on the dollar, the discount to nominal principal amounts to $45 million. This is not necessarily HSBC's realized accounting loss, but it represents a material concession on a financing position established less than a year earlier.
The transaction is not isolated. ING is marketing approximately $500 million of data-center loans, Société Générale has reportedly sought buyers for up to $300 million of financing connected to other Stargate developments, and several European banks are exploring credit insurance and significant risk transfer arrangements. Their motivations and the quality of the underlying exposures differ, but these developments suggest that the willingness to retain large concentrations of AI infrastructure debt is becoming more selective.
Public discussion has concentrated heavily on technology valuations and whether demand for artificial intelligence can justify the scale of investment being committed. From a banking perspective, I consider the financing behind that expansion a more immediate issue. Data centers remain capital-intensive infrastructure projects, exposed to construction delays, regulatory approvals, energy availability and refinancing conditions. Strong demand for computing capacity does not eliminate those risks, nor does the involvement of a financially substantial technology company automatically make every project financing associated with it a sound credit.
The recent transactions do not establish a developing non-performing loan crisis. They do, however, provide evidence that investors and lenders are reconsidering the price at which they are prepared to assume certain risks. For banks that committed substantial capital during the rapid expansion of AI infrastructure, the consequences could extend beyond secondary-market valuations to the original underwriting assumptions and the economics of retaining those exposures.
Project Jupiter: When the Tenant Is Stronger Than the Project
Project Jupiter is a planned data-center campus with approximately 2.45 gigawatts of intended capacity in Doña Ana County, New Mexico. The development is associated with Oracle's expansion of computing infrastructure for the Stargate initiative. Its financing, known as Project Miner, involves project-level borrowing structures rather than a conventional unsecured corporate loan to Oracle.
The distinction has practical consequences for credit assessment. A major technology company may provide a valuable source of contractual support, particularly through long-term capacity commitments, lease payments or other financing obligations. However, the borrower must still deliver an operational asset capable of supporting the repayment assumptions on which the financing was approved. The financial standing of the tenant and the creditworthiness of the project are related, but they cannot be treated as interchangeable.
The project's own disclosures provide useful context. In its consolidated annual report submitted to Doña Ana County on 31 July 2026, Project Jupiter reported estimated investment of approximately $4.7 billion through 30 June. At that reporting date, construction of the data-center facilities was approximately 9% complete, while the associated microgrid was approximately 10% complete. These are historical figures rather than an assessment of the project's present construction status. They nevertheless illustrate the scale of investment already committed while substantial development work remained outstanding.
More recent difficulties have centered on the infrastructure required to make the campus operational. Oracle has maintained that the data-center buildings and the proposed microgrid are legally separate facilities with different permitting requirements. That distinction is relevant from a regulatory perspective. From a lender's perspective, however, the economic relationship is considerably more difficult to separate. A completed facility without sufficient electricity cannot deliver the operating performance assumed in its financing model, irrespective of whether the buildings themselves have received the necessary construction approvals.
On 6 November, the public hearing concerning the microgrid's air-quality permit was postponed again, without a replacement date being immediately established. The postponement does not establish that the project will fail or that the required approvals cannot ultimately be obtained. It does, however, leave a critical operational dependency unresolved at a time when substantial financing commitments have already been made.
There are contractual protections that may reduce the consequences for lenders. According to the Financial Times, Oracle could remain responsible for specified financing carry costs for up to three years if the facilities are not operational as scheduled, potentially covering lenders' interest and an agreed return to equity investors. Such commitments can offer meaningful protection against delays. Their effectiveness depends on the contractual conditions, the period covered, possible exclusions and the ability of the responsible counterparty to meet its obligations.
The credit assessment must therefore go beyond Oracle's financial standing. A bank financing a project of this magnitude should be able to demonstrate that its exposure remains serviceable and recoverable under credible adverse scenarios involving delayed electricity supply, higher completion costs and postponed commissioning. The quality of contractual support matters, but so does its adequacy when the original project timetable no longer holds.
A strong tenant can improve the credit profile of a project. It cannot compensate for every weakness in its financing structure.
European Banks Are Reassessing Their Exposure
HSBC's disposal forms part of wider activity among European lenders seeking to reduce or redistribute financing exposures associated with AI infrastructure. The reported transactions include outright loan sales, competitive auctions and proposed credit protection arrangements. These instruments serve different economic purposes and should not be interpreted as evidence of identical credit concerns.
ING's auction includes exposures associated with Vantage Data Centers, CyrusOne and a $25 million position linked to Project Jupiter. The bank is seeking competitive bids for 12 loans, with the deadline set for 19 November. The resulting prices may provide additional evidence of investor appetite, although individual positions within the portfolio have different borrowers, contractual protections and risk characteristics.
| Bank | Reported exposure | Transaction or initiative | Status at 17 November 2026 |
|---|---|---|---|
| HSBC | Approx. $300m | Project Jupiter construction-loan disposal at prices in the mid-80s | Reported completed sale |
| ING | Approx. $500m | Competitive sale process involving 12 data-center loans | Bids due 19 November |
| Société Générale | Up to $300m | Proposed sale of Stargate-related loans in Texas and Wisconsin | Seeking buyers; completion unconfirmed |
| Deutsche Bank | Approx. €2bn | Proposed significant risk transfer involving a broader project-finance portfolio, including data-center exposures | Structuring discussions reported |
Sources: Bloomberg, Reuters and 9fin, September–November 2026. Amounts represent reported nominal loan positions or reference portfolios, not completed disposal proceeds. Deutsche Bank's portfolio is not exclusively data-center financing.
Société Générale appears to have adopted a more selective approach. According to specialist credit publication 9fin, the bank has sought to sell up to $300 million of financing associated with Stargate developments in Texas and Wisconsin. The reported asking price was approximately 95 cents on the dollar, while an offer in the 80s was rejected. This suggests that lenders are distinguishing between individual projects rather than indiscriminately reducing all exposure to AI infrastructure.
Elsewhere, Barclays has reportedly explored default protection on more than $10 billion of AI infrastructure financing, while Deutsche Bank has considered a significant risk transfer transaction involving approximately €2 billion of project-finance debt, including data-center loans. These initiatives are not distressed asset disposals. Credit insurance and risk transfer structures can allow banks to manage concentration limits, capital consumption and available lending capacity while continuing to originate business.
Nevertheless, I would be cautious about dismissing the recent activity as entirely routine portfolio management. A bank selling a recently originated exposure at a material discount is making an economic decision that deserves examination. The motivation may be concentration management, a change in expected recoveries, liquidity considerations or the availability of better risk-adjusted opportunities elsewhere. Without the underlying credit files and transaction terms, it is impossible to determine which factor predominated in HSBC's decision.
The willingness to accept a discount does not necessarily indicate that the original credit approval was flawed. Market conditions and project circumstances may have changed in ways that were not reasonably foreseeable at origination. An effective credit risk framework must nevertheless distinguish between genuinely unexpected developments and risks that should have been identified, quantified and accommodated in the original financing structure.
That distinction becomes particularly relevant when several institutions attempt to distribute comparable exposures within a relatively short period. The question is not simply why banks are selling, but whether the returns originally agreed remain adequate for the risks they are now being asked to retain.
The Credit Market Is Already Repricing Risk
Secondary-market transactions are only one source of evidence. Recent data on US data-center securitizations indicate that investors are also demanding higher compensation for certain new exposures.
According to the CRE Finance Council's Data Center New Issue Snapshot, published on 16 November 2026, the average spread on five-year, single-A-rated data-center asset-backed securities increased from 195 basis points above US Treasury benchmarks in the first quarter to 257 basis points for issues priced during the latest 13-week period. The increase of 62 basis points represents approximately 31.8% of the earlier spread. The averages cover eight and five different transactions, respectively, rather than an unchanged portfolio of securities.
The movement is also visible among repeat issuers. Comparable Switch securities were priced at 225 basis points over Treasuries in March and 290 basis points in October, while DataBank's spreads increased from 205 to 220 basis points between January and October. Differences in collateral, contractual arrangements and transaction structure still limit direct comparability, but these observations provide additional evidence that the widening is not solely attributable to changes in the composition of issuers.
Issuance volumes present a more differentiated picture. US data-center ABS and CMBS issuance totaled approximately to $21.87 billion through 13 November 2026, 13.1% below the corresponding period of 2025. However, ABS issuance increased by 9.4% to $15.32 billion, while CMBS issuance declined by 41.4% to $6.55 billion. The divergence suggests that investors are discriminating between financing structures and collateral rather than withdrawing uniformly from the sector.
The year-on-year comparison also requires care. One week earlier, cumulative issuance was still 1.2% ahead of the previous year. Much of the subsequent deterioration reflected the inclusion of a large November 2025 transaction in the comparative base, rather than a sudden contraction in new lending during the intervening week. Nevertheless, issuance during the latest three-month period was 48.0% below the corresponding period of 2025, with no new data-center ABS or CMBS transactions priced for three consecutive weeks.
By comparison, overall US private-label CMBS and commercial real-estate CLO issuance was approximately 13.3% higher year-on-year. Financing conditions are therefore becoming more differentiated, even while other segments of structured credit continue to attract capital.
These observations should not be applied mechanically to individual construction loans. ABS spreads, private lending margins and secondary-market disposal prices reflect different contractual rights, security packages and investor requirements. Their collective direction is still relevant: certain categories of AI infrastructure debt are becoming more expensive to finance, and investors are placing greater emphasis on the quality of the underlying structures.
For a highly leveraged development, higher financing costs can weaken equity returns and refinancing capacity without causing an immediate payment default. A project that remains viable under its original financing terms may require additional sponsor capital or revised repayment arrangements if debt markets become less accommodating before construction is completed.
The Financial Cost of Delayed Completion
Construction delays are often discussed as operational setbacks. In project finance, they can materially change the economics of the transaction. Debt continues to accrue interest, capital remains committed, liquidity reserves may be consumed and the commencement of operating cash flows can move further into the future.
Consider a hypothetical data-center development financed with $1 billion of fully drawn, non-amortizing construction debt. At an annual financing cost of 8%, the outstanding debt would accrue $80 million of interest over a twelve-month period. If the financing cost increased to 9%, that amount would rise to $90 million. The additional 100 basis points would therefore create $10 million of additional annual interest, assuming unchanged principal and no compounding.
The full $80 million or $90 million should not automatically be treated as an incremental loss caused by delayed completion. Under the original financing model, interest would generally continue to accrue after commissioning, although it might then be serviced from operating income or refinanced on different terms. The correct analysis compares the originally expected cash flows and financing arrangements with the revised position following the delay.
That comparison may be considerably less favorable than the simple interest calculation suggests. Rental commencement may be postponed, construction expenditure may increase, reserve accounts may be depleted and debt maturities may approach before the project has demonstrated stable operating performance. Some of these costs may be absorbed through capitalized interest, additional sponsor equity or enforceable contractual payments from a tenant. Each mechanism affects the financing structure differently.
Capitalized interest increases outstanding debt, reserve utilization reduces liquidity and further equity contributions dilute sponsor returns. A tenant's commitment to cover financing costs can protect lenders during construction, but the protection must be assessed against its contractual limits and the circumstances under which payment becomes due.
In a conventional credit review, I would expect particular attention to the relationship between remaining completion costs, available liquidity, contractual support and the maturity profile of the debt. The objective is not merely to establish that the project can eventually become operational. The financing must remain sufficiently resilient throughout the period required to reach that point.
This also determines when an operational delay becomes a restructuring issue. A temporary extension supported by adequate reserves and binding contractual commitments may be manageable within the original financing framework. A prolonged delay that exhausts liquidity, requires repeated maturity extensions or makes repayment dependent on additional sponsor support represents a materially different credit position.
For a secondary-market investor, the economics are equally sensitive to recovery assumptions. A loan purchased at 85 cents and repaid at par after two years would generate approximately 8.5% annualized capital appreciation before interest income, funding costs, transaction expenses and taxes. If the eventual principal recovery were only 90 cents, the annualized capital appreciation would fall to approximately 2.9% on the same simplified basis. These are illustrative calculations, not estimates of Project Jupiter's recovery prospects.
The discount may create an attractive investment opportunity, particularly where contractual payments remain protected and the underlying project is expected to reach completion. However, a lower entry price does not make an exposure fundamentally safer. The investor must still establish whether the expected recovery justifies the purchase price, capital commitment and remaining execution risk.
For the selling bank, the decision is the reverse. Retaining an exposure solely to avoid crystallizing a discount may be economically inferior to selling it if the expected risk-adjusted return no longer compensates for the capital and concentration risk involved. Equally, disposing of a well-protected exposure at a substantial discount may sacrifice value unnecessarily.
The rational decision depends on the revised credit assessment, not on the psychological attraction of recovering nominal principal.
What Reported Credit Quality May Not Yet Capture
A discounted loan sale does not establish that a loan has become non-performing. Under IFRS 9, Stage 2 classification generally requires recognition of lifetime expected credit losses following a significant increase in credit risk since initial recognition, while Stage 3 applies to credit-impaired exposures. Neither classification can be inferred solely from an observed secondary-market price.
Similarly, the approximately $45 million discount implied by HSBC's reported disposal at an illustrative 85 cents on the dollar cannot be presented as a confirmed accounting loss. The actual financial effect depends on the loan's carrying value, any impairment already recognized, accrued interest and other transaction-specific adjustments. These details have not been publicly established.
Market prices can nevertheless provide information that conventional asset-quality indicators do not immediately reveal. A borrower may continue servicing its obligations while investors require materially higher returns to assume its debt. Refinancing conditions may deteriorate before financial covenants are breached, just as collateral values may change before the lender recognizes a credit impairment.
The regulatory debate is beginning to reflect these concerns. In its Semiannual Risk Perspective, Fall 2026, the US Office of the Comptroller of the Currency identified rapid growth in financing requirements, construction dependencies and relatively untested financing structures as risks requiring attention. Although banks often occupy senior and secured positions, the effectiveness of those protections depends on the actual recoverability of the underlying assets and contractual claims.
The European Central Bank has also examined the sector. Using AnaCredit data for the second quarter of 2026, ECB Banking Supervision estimated approximately €62 billion of AI-related committed corporate credit exposures among euro-area banks, equivalent to 1.2% of their total committed corporate exposures. The figure covers a broader range of activities, including software development, semiconductors and data centers, while excluding certain property-classified facilities and indirect financing through special-purpose vehicles.
That limitation has consequences for understanding the wider distribution of risk. Large infrastructure developments are increasingly financed through combinations of bank lending, project companies, private credit funds, insurers and securitization vehicles. Direct corporate lending figures may capture only part of the financial relationships supporting those projects.
On 16 November, ECB Vice-President Boris Vujčić warned that AI-related financial stability risks had expanded beyond equity-market valuations to include greater reliance on debt financing and concentrated exposures. The ECB has also indicated that it is monitoring underwriting standards, potential credit-quality deterioration and banks' connections with private credit markets.
Credit insurance and significant risk transfer transactions can reduce a bank's retained exposure and may provide regulatory capital benefits where the applicable recognition requirements are satisfied. The underlying risk, however, remains within the financial system. Its distribution becomes more complex when potential losses depend on insurers, institutional investors or other counterparties whose capacity to perform may itself be affected by adverse market conditions.
The concern is not simply that risks are being transferred outside bank balance sheets. Such transactions can improve risk diversification and support additional lending capacity. The concern arises when the distribution of exposure becomes less transparent than the original financing and when assumptions about the effectiveness of credit protection have not been adequately tested.
For supervisors and bank management, understanding the ultimate allocation of risk is therefore becoming as important as monitoring the nominal amount of AI infrastructure financing.
Conclusion: The Price of Credit Discipline
The recent transactions do not establish that AI infrastructure financing is entering a broad non-performing loan cycle. HSBC's reported disposal concerns a project with specific permitting and infrastructure challenges, while other banks are pursuing transactions that may reflect concentration management, capital efficiency or ordinary portfolio distribution. There is insufficient evidence to conclude that comparable deterioration is occurring across the entire sector.
However, I do not consider the emerging repricing a development that banks can dismiss as routine market volatility. Some lenders are accepting meaningful discounts, others are seeking additional credit protection, and investors are demanding higher returns on certain securitized exposures. These developments suggest that the market is reassessing risks that may have been overshadowed by the scale of investment commitments and the financial standing of major technology companies.
My concern is not that banks have financed the expansion of AI infrastructure. Demand for computing capacity remains substantial, and well-structured projects supported by strong contractual arrangements can represent sound lending opportunities. The concern is whether the pace of development encouraged some lenders to place too much emphasis on expected tenant demand while giving insufficient weight to the financial resilience of the individual projects they were funding.
That question cannot be answered conclusively without examining the original financing terms, credit approvals and contractual protections. An adverse development does not automatically prove that a bank made a poor lending decision. Even a well-structured transaction can suffer from circumstances beyond the reasonable expectations of its lenders.
The quality of underwriting is demonstrated by how those possibilities were addressed at origination. A sound credit decision should incorporate credible downside scenarios, adequate liquidity, enforceable contractual support and a repayment structure that does not depend on the uninterrupted execution of every major milestone.
For banks now reviewing their exposures, the appropriate response is not necessarily to sell. Some credits may remain attractive despite lower secondary-market prices, particularly where contractual protections and expected recoveries remain sound. Others may no longer offer sufficient compensation for the risks involved. In those circumstances, accepting a discount can be more prudent than retaining an exposure whose economics have materially deteriorated.
I consider the current repricing a necessary adjustment in credit discipline. It reflects a market beginning to distinguish more carefully between the investment potential of artificial intelligence and the creditworthiness of the infrastructure projects built to support it. That distinction should have formed part of the original lending assessment, but it is now becoming increasingly visible in market prices.
The consequences will not be uniform. Projects supported by resilient financing structures, adequate reserves and strong contractual commitments should be better placed to withstand delays and tighter refinancing conditions. Others may require additional sponsor capital, amendments to financing arrangements or, in more difficult cases, formal restructuring before their assets become fully operational.
For lenders, the measure of success will not be the volume of AI infrastructure financing arranged or the reputation of the companies expected to occupy the completed facilities. It will be the ability of those projects to service, refinance and repay their debt when the assumptions supporting the original credit decision are tested. If a financing structure remains viable only when construction, power supply and refinancing proceed precisely as planned, the lender has accepted a margin of safety that is difficult to reconcile with the scale and complexity of the underlying risk.
Daniel Vnuk, univ.mag.oec., MBA
Corporate Credit Risk | Restructuring | Special Assets | Distressed Assets
Restructuring.Insights
E: dvnuk@restructuring-insights.com
W: www.restructuring-insights.com
LinkedIn: linkedin.com/in/dvnuk
Sources and References
Primary Sources
1. CRE Finance Council. CMBS/ACMBS and Data Center New Issue Weekly, 16 November 2026. Data Center New Issue Snapshot, information through 13 November 2026, p. 3.
2. Yucca Growth Infrastructure LLC and Oracle America Cloud Services LLC. Project Jupiter 2026 Consolidated Annual Report, 31 July 2026, p. 4. Supplemental project status through 30 June 2026.
3. Oracle Corporation. Project Jupiter Statement on Construction Permitting, 14 September 2026.
4. Office of the Comptroller of the Currency. Semiannual Risk Perspective, Fall 2026.
5. European Central Bank. Written Overview for the Exchange of Views of the Chair of the Supervisory Board with the Eurogroup, 9 November 2026, p. 2.
https://www.bankingsupervision.europa.eu/ecb/pub/pdf/ssm.written_overview261109.en.pdf
6. European Central Bank. Boris Vujčić, Risks and Resilience in the Euro Area Financial System, 16 November 2026.
https://www.ecb.europa.eu/press/key/date/2026/html/ecb.sp261116~db374b56ab.en.html
Market and comparative sources
7. Bloomberg. Reshmi Basu and Jeannine Amodeo, HSBC Sold $300 Million Troubled Data Center Loan at Discount, 16 November 2026.
8. Reuters. ING Puts $500 Million of Data-Center Loans Up for Sale, Bloomberg News Reports, 6 November 2026.
9. 9fin. Daniel Geiger, Max Frumes and Maria Heeter, SocGen Looks to Offload Stargate Data Center Debt at Discount, 11 November 2026.
https://beta.9fin.com/insights/socgen-stargate-data-center-debt
10. Bloomberg. Esteban Duarte and Arno Schütze, Deutsche Bank Hedges Data Center Risk in SRT on €2 Billion Book, 7 September 2026.
11. Bloomberg. Esteban Duarte and William Shaw, Barclays Turns to Insurers to Hedge Billions of Data Center Debt, 28 October 2026.
12. Financial Times. Oracle on the Hook to Pay Data Centre Investors Even if Site Has No Electricity, October 2026.
https://www.ft.com/content/a96bf05a-a299-4d6a-a753-b298dd0f4016
13. Albuquerque Journal. Algernon D'Ammassa, Project Jupiter Air Permit Faces New Delay, 6 November 2026.
https://www.abqjournal.com/news/project-jupiter-air-permit-faces-new-delay/3157572
14. Financial Times. Banks Seek to Curb AI Debt Exposure as Backlash Mounts, 17 November 2026.
https://www.ft.com/content/2351d9db-e938-4102-a449-318b1037765a
Valuation and Currency Methodology
This analysis is based on publicly available information through 17 November 2026. Transaction details concerning HSBC, ING, Société Générale and other lenders are drawn primarily from Bloomberg, Reuters, 9fin and Financial Times reporting. Certain transactions have not been formally confirmed by the institutions concerned, while proposed disposals and risk transfer arrangements remain subject to execution. The analysis does not reconstruct undisclosed loan-level financial information.
HSBC's illustrative discount is calculated by applying an assumed transaction price of 85% to approximately $300 million of nominal principal. The resulting $45 million difference between face value and indicative proceeds is not a verified measure of realized accounting loss or expected credit impairment. Actual pricing was reported as being in the mid-80s. The calculation excludes the carrying amount, previously recognized impairment allowances, accrued interest, transaction costs and other contractual adjustments.
Securitization data are taken from the CRE Finance Council's publication of 16 November 2026, covering information through 13 November. The spread comparison uses simple averages of different five-year, single-A-rated data-center ABS transactions priced during the relevant periods. Differences in collateral, issuer characteristics, security arrangements and transaction structure limit direct comparability. These market observations should not be used as standalone estimates of fair value, expected credit losses or recovery rates for individual construction loans.
The illustrative financing scenario assumes fully drawn, non-amortizing debt of $1 billion and annual financing costs of 8% and 9%. Interest is calculated on a simple basis without compounding, fees or principal adjustments. The figures represent financing carry during the assumed period rather than the full incremental economic cost of delayed completion. That cost would require comparison of the original and revised project cash flows, including operating income, contractual payments, construction expenditure, reserve utilization and refinancing terms. The secondary-market return illustrations assume purchase at 85% of principal and repayment after two years at the stated recovery amounts, excluding coupon income, funding costs and other expenses.
All financial amounts are presented in their original reported currencies. No foreign-exchange conversions or cross-currency aggregations have been performed. The ECB's €62 billion estimate refers to committed corporate exposures under its supervisory definition, rather than drawn balances or a complete measure of indirect AI infrastructure financing.
Disclaimer
This article has been prepared for informational, educational, analytical and professional discussion purposes. It represents the author's independent assessment of developments in AI infrastructure financing, bank credit risk, secondary debt markets and related regulatory considerations, based on publicly available information as at 17 November 2026.
The views, interpretations and conclusions expressed are solely those of the author and do not represent the official position of any financial institution, company, investor, regulatory or supervisory authority mentioned in this publication. References to institutional names, corporate identities and trademarks are made for identification and editorial purposes and do not imply affiliation, sponsorship, approval or endorsement.
As at the publication date, the author holds no direct or indirect financial interests in any of the banks discussed in this article.
The analysis draws on publicly available regulatory publications, project disclosures, market data and financial media reporting. Certain transactions discussed have not been independently confirmed by the institutions concerned, while specific financing terms, accounting classifications and contractual arrangements remain undisclosed. Although reasonable care has been taken in reviewing the available information, it has not been independently audited, and no representation or warranty is made regarding its completeness, accuracy or continued validity.
Certain financial calculations, secondary-market pricing interpretations and illustrative financing scenarios have been independently prepared using the assumptions identified in the accompanying methodology. They should not be interpreted as verified transaction outcomes, recognized accounting losses, formal credit assessments or forecasts of future financial performance. Actual results may differ materially from the illustrative scenarios presented.
The discussion of credit underwriting, project-finance structures, risk management decisions and secondary-market transactions is financial and economic in nature. It does not constitute an allegation or finding of misconduct, negligence, regulatory non-compliance or improper conduct by any institution or individual. The author's professional assessments are based on the information available and do not imply access to confidential credit files, internal approval procedures or undisclosed contractual arrangements.
Forward-looking assessments are subject to uncertainty, including changes in project execution, regulatory approvals, energy availability, credit conditions, refinancing markets and other factors beyond the author's knowledge or control. Events occurring after the stated information cut-off date may alter the circumstances examined. The author undertakes no obligation to update or revise this analysis.
Nothing in this publication constitutes regulated investment research, an independent valuation, a fairness opinion or personalized investment, financial, accounting, regulatory or legal advice. It is not an offer, solicitation or recommendation to purchase, sell, hold or otherwise transact in any security or financial instrument, nor does it establish an advisory or professional relationship with any reader.
Financial instruments, including loans and debt securities, involve risks of loss, potentially including the entire capital invested. Readers should conduct their own due diligence, independently assess the information and assumptions presented and obtain advice from appropriately qualified professionals before making investment or commercial decisions. To the fullest extent permitted by applicable law, neither the author nor Restructuring.Insights accepts liability for losses arising from reliance on this publication.
The limitations and qualifications stated in this disclaimer form an integral part of the analysis and should be read together with the accompanying methodology and conclusions.
© 2026 Restructuring.Insights. All rights reserved. Original analysis and commentary may not be reproduced or commercially distributed without prior written permission, except as permitted by applicable law. Third-party trademarks and source materials remain the property of their respective owners.