The Artificial Accounting of Artificial Intelligence
AI at scale is inevitable, but so is human greed and impatience.
I will start this off by clarifying that I am not shocked, but once again disappointed as I write this, that we continue to live in what feels like a dual reality of both an incredibly innovative and creative time, built upon incredibly creative, legal, (questionably ethical) fractured and ever-more rigged financial foundation of the Federal Reserve and centralized fiat financial markets. That all being said, as the great saying goes, “Don’t hate the player, hate the game.”
The Meta–Blue Owl Louisiana deal
This recent deal is a vivid example of how “artificial intelligence” is being financed with what can fairly be called “artificial accounting.” Meta gains access to a roughly $27 billion hyperscale data center campus, critical to its AI ambitions, while most of the asset and the associated debt live somewhere other than Meta’s balance sheet. This is not fraud; it is a sophisticated use of existing accounting and financing rules to reshape how economic reality appears to investors, regulators, and the public.[11][12][13]
However, it is a sleight of hand that is conveniently buried in the fine print of the prospectus that very few passive investors (or their advisors) ever read, and even fewer would understand.
The deal in one sentence
Meta is building its largest AI-optimized data center in Richland Parish, Louisiana, via a joint venture where it owns 20% and Blue Owl Capital funds own 80%, funded by an SPV (Special Purpose Vehicle) called Beignet Investor that issued about $27.3 billion in bonds largely bought by Pimco and other institutions.
Meta will not show that $27-ish billion of project debt as its own borrowings; instead, it will recognize a lease and a right-of-use asset while the bond obligations sit at the SPV level.[12][14][11]
How Meta built a $27B data center without “borrowing.”
Meta’s Louisiana campus, sometimes called “Hyperion,” is designed to be the company’s largest data center globally, with roughly 4 million square feet of white space and up to 2 gigawatts of power capacity at full build-out. It anchors Meta’s long-term plan to train and serve AI models like Llama, and the state has retooled incentives and infrastructure to secure the project.[3][5][7]
Instead of financing the build directly on its own balance sheet, Meta set up a joint venture with Blue Owl Capital, with Meta holding around 20% and Blue Owl-managed funds holding about 80%. The Blue Owl side is organized in an SPV called Beignet Investor, which issued approximately $27.3 billion in bonds (DEBT) to investors, with Pimco cited as a major anchor buyer. Meta contributed land and prior spend, received around $3 billion of cash, and agreed to lease the facility once it is operational.[1][2][14][15][16][12]
At a high level:
The SPV (Beignet Investor) borrows ~$27B by selling bonds to asset managers like Pimco.[14][15][12]
The SPV uses that capital to fund the construction of the Louisiana data center campus.[11][12]
Meta becomes the tenant, agreeing to pay rent under a long-term lease structure and providing various guarantees.[12][11]
The key is that the SPV, not Meta, is the legal issuer of the bonds. Meta gets the AI infrastructure it needs (the stock pump off the press release), but the giant slug of debt appears on the “shadow” ledger of a project vehicle backed by private credit, not on Meta’s own balance sheet as conventional corporate debt.[13][17][12]
Operating lease vs. finance lease: the four-year trick
The accounting magic revolves around how the lease is structured. Meta’s economic intent is to use the facility for roughly 20 years starting at the end of the decade. But the contract is written as:
An initial non-cancellable lease term of about four years.[18][12][11]
Multiple renewal options that allow Meta to extend its use of the data center for up to roughly 20 years in total.[12][11]
Under lease accounting rules, the distinction between an “operating lease” and a “finance lease” hinges on substance:
Does the arrangement effectively transfer control and substantially all the risks and rewards of ownership to the lessee?
If so, it should be treated like owned, debt-financed equipment (a finance lease).
If not, it can be treated more like a service contract (an operating lease), even though there is still a lease liability and right-of-use asset.[18][12]
By baking in:
A short guaranteed term (four years), and
Renewal options that are formally “optional,”
Meta can argue that it is not “reasonably certain” to exercise all those options, and therefore, the enforceable lease term is only four years for accounting purposes. A four-year lease on a campus expected to operate for decades looks much more like an operating lease than ownership, especially when paired with a narrative that Meta does not control the joint venture’s key economic decisions.[14][18][12]
The result: Meta books a lease liability and right-of-use asset tied to a short base term rather than an asset and debt stack resembling a $27B owned data center financed by bonds.
The commitment looks lighter, even though the economic intent is to have a very long-lived AI campus.[11][12]
“We don’t control it” – the consolidation escape hatch
The second leg of the stool is the claim that Meta does not control the joint venture or the SPV that owns the data center. For an entity to be consolidated onto a company’s balance sheet, accounting rules look for control: who has power over the activities that significantly affect returns, and who is exposed to the majority of risks and rewards?[14][12]
In this case, the public narrative is:
Blue Owl funds hold 80% of the JV and control the board or formal governance of the SPV.[12][14]
Beignet Investor, not Meta, is the issuer of the bonds and the legal owner of the data center asset.[14][12]
Meta is the tenant and a minority equity holder, but purportedly does not control the JV’s economically significant decisions.[12][14]
The tension is obvious
Meta is the hyperscaler that designs, builds, and runs massive data centers as a core competence; Blue Owl is primarily a capital provider. Meta manages construction, oversees property operations, and is effectively the only realistic tenant for such a specialized facility, which gives it enormous practical leverage over how the campus is used and maintained.[2][11][12]
Yet, for consolidation purposes, the assertion is that Blue Owl’s governance position and economic rights are sufficient that Meta cannot be said to “control” the entity. That classification keeps the SPV and its $27B of bonds outside Meta’s consolidated financial statements, despite Meta’s central role in making the project viable.[14][12]
This is not unprecedented. Project finance in energy, telecom, and infrastructure has used similar structures for decades. The difference here is scale, AI-driven urgency, and the way a single hyperscaler’s economic centrality can coexist with a formal claim of “no control.”[17][19]
Residual Value Guarantee: The Invisible $27B Debt
If the short lease term and non-consolidation are the visible tricks, the residual value guarantee (RVG) is the invisible one.
Meta has reportedly provided a guarantee that, for a large portion of the project life (roughly the first 16 years), if it terminates the lease early or chooses not to renew, it will backstop the residual value of the project up to a defined cap.
In plain language: if Meta walks away, Meta will step in with cash so bondholders don’t suffer a major loss on a purpose-built data center in rural Louisiana that may be difficult to repurpose.[16][18][11]
That guarantee does three things simultaneously:
It makes the bonds much safer and therefore financeable for Pimco and its peers.[16][18][11]
It aligns the economics with the idea that Meta is effectively underwriting the project’s long-term viability.[18][11]
It creates a contingent liability for Meta that looks a lot like a form of hidden debt.[18][12]
Accounting rules require companies to recognize liabilities for guarantees when it is “probable” that an obligation will be incurred and the amount is reasonably estimable. The off-balance-sheet nature of this RVG relies on the judgment that the likelihood of Meta triggering it is not high enough to require a full liability to be booked.[18][12]
That judgment strains common sense. If Meta is investing resources, political capital, and time into a $27B data center campus that will not be fully built until near 2030, the economically realistic scenario is that Meta uses it for well beyond four years. If Meta leaves early, the guarantee is likely to come into play. If Meta stays, then the short initial term and “uncertain” renewals look artificial. Either way, the RVG exposes Meta to a material obligation that sits in the shadows relative to how investors typically think about leverage.[5][2][3]
The Three Claims That Don’t Cohere
This is where I believe that “artificial accounting” becomes an apt phrase. The structure rests on three simultaneous claims:
Meta does not control the JV’s economically significant decisions.
Meta is not reasonably certain to stay beyond the four-year initial lease term.
Meta’s residual value guarantee is unlikely enough that no large liability needs to be recorded.
Taken together, they are hard to reconcile with economic reality.
On control: Meta designs, builds, and operates the facility; it is essentially the only natural operator and off-taker for such a highly specialized campus. Governance documents may allocate formal control to Blue Owl, but the practical dependency of the facility on Meta’s presence suggests that Meta has de facto control over the asset’s fate.[2][11][12][14]
On lease term: no rational operator expects to recover the full value of a $27B hyperscale data center within four years. The campus is designed for a multi-decade life; the AI hardware inside will churn, but the shell, power, and network infrastructure will anchor Meta’s platform for a long time. Treating renewal as genuinely uncertain is convenient for classification, not economically persuasive.[8][3][12][18]
On the guarantee: if Meta might credibly leave after four years, that raises, rather than lowers, the probability that the RVG is relevant, because alternative tenants for such an asset and location are thin on the ground. If leaving is unlikely, then the lease term assumption looks contrived. Either scenario undermines at least one leg of the accounting logic.[11][18]
This is why journalists and analysts have described the transaction as “off-balance-sheet financing without consolidation” and likened it to pre-crisis structures that shifted risk into opaque vehicles while maintaining economic exposure.[13][17][12]
The Shadow Ledger of AI Infrastructure Deals
Meta is not alone. The Louisiana deal is one of several highly structured AI infrastructure financings that are rewriting Wall Street’s playbook. A recurring theme across commentary and bank research is that AI capex is so large, and the hyperscalers are so focused on preserving reported balance sheets and EPS optics, that a new “shadow system” of private credit and SPVs is emerging to carry the load.[15][19][17][13]
Analysts have highlighted several patterns:
AI data centers are being financed via project-specific entities, often backed by private credit funds, with long-term offtake agreements from hyperscalers.[15][13]
These structures are engineered to avoid consolidation, keeping multi-billion-dollar debts away from the hyperscalers’ reported leverage metrics.[17][12]
Guarantees, minimum-volume commitments, and make-whole provisions create economic exposures that may not show up as headline debt.[19][17]
One widely cited estimate from Morgan Stanley suggests that tech companies may require on the order of $800 billion of private-credit, often off-balance-sheet, financing by 2028 to fund AI data centers and related infrastructure. Given current run-rates, AI-linked funding is already accumulating at something like $100 billion per quarter across various structures. Meta’s Louisiana JV is simply one of the largest and most visible.[19][17]
The Dangerous Sleight of Hand
The AI revolution is being built in a parallel financing universe, where the true leverage sits in SPVs, private funds, and bespoke contracts rather than in the core financial statements of the Big Tech names driving the wave.[13][17]
Why this matters for investors and policymakers
From a narrow corporate-finance perspective, Meta’s move is rational. Off-balance-sheet project finance can:
Preserve headline net-debt-to-EBITDA and capex metrics.[13][12]
Free up balance sheet capacity for other strategic initiatives or buybacks.[17][13]
Attract specialized infrastructure investors willing to accept lower returns in exchange for quasi-bond-like cash flows backed by a hyperscaler.[15][11]
But at the system level, several risks emerge:
Hidden leverage: Investors, rating agencies, and regulators may underestimate how much AI-linked debt is ultimately backstopped by a handful of hyperscalers via guarantees and contractual obligations.[19][17]
Pro-cyclicality: In a downturn or AI demand shock, SPVs reliant on a single hyperscaler tenant could suddenly reveal contingent liabilities as guarantees are called or leases are restructured.[17][19]
Regulatory blind spots: Traditional bank-centric oversight may miss concentrations of risk in private credit funds and SPVs that depend heavily on the perceived invincibility of Big Tech.[19][17]
The Meta–Blue Owl deal is important not just because of its size, but because it normalizes a template. If every major hyperscaler pushes AI infrastructure into JV/SPV structures with short-term leases, residual guarantees, and non-consolidation, then much of the true AI capex burden becomes harder to track and model.
Oh, and just in case you wondered what I think they will do? They almost certainly already have copied this template across the board.
Is this wrong, or just clever?
Legally and technically, transactions like this sit inside the rules. Company management and auditors can point to specific accounting standards on leases, consolidation, and guarantees, and show how the structure threads the needle.
In my opinion, this is one of those situations, however, where one must ask, “Just because we can do something, should we?”
They can argue that investors are not truly misled because disclosures in footnotes, MD&A, and presentations describe the nature of the commitments.[12][14][18]
The problem is that economic intuition and formal classification diverge:
Economically, Meta is the anchor of a $27B AI campus that only exists because Meta wants it and will use it for decades.[3][5][2]
Formally, Meta can present itself as a minority JV partner with a four-year lease and a remote guarantee, leaving a vast amount of debt in the shadows.[11][14][12]
Calling this “artificial accounting” is not to accuse Meta of lying; it is to highlight that the map no longer looks like the territory.
The numbers that matter most to long-term systemic risk—aggregate AI infrastructure leverage, concentration of hyperscaler counterparty risk, and the durability of private-credit structures—are increasingly off the page.
For sophisticated allocators and policymakers, the response should not be panic but adaptation:
Ask how much AI capacity is backed by off-balance-sheet structures, not just reported capex.[17][19]
Scrutinize guarantees, minimum payments, and off-take agreements as carefully as bonds.[17][18][12]
Recognize that the new “rails” of AI—data centers, power, fiber—are being financed in a way that echoes pre-2008 project finance and securitization, with all the benefits and familiar pitfalls.[19][17]
Meta’s Louisiana project shows that the artificial intelligence boom is being built on multi-layered financial engineering. The code running inside the data centers may be new, but the balance-sheet alchemy funding those data centers is very old.[13][12]
Why Meta’s Structure is Allowed Under Today’s Rules (Post-Enron, But Familiar)
It is legal because it fits within current accounting and securities rules for leases, consolidation, guarantees, and SPVs, but it lives right on the edge of those rules in a way that rhymes with Enron’s structures, even if it is not the same in intent or opacity.
The core difference is that what Enron did often depended on concealment, self-dealing, and outright misclassification, whereas Meta’s structure is closer to aggressive but disclosed financial engineering that exploits judgment calls built into modern standards.[1][2][3]
Modern GAAP/IFRS gives companies structured tests for three key questions:
Should a project entity be consolidated?
Is a lease “finance” or “operating”?
When do guarantees become liabilities on the balance sheet?
Meta’s Louisiana data-center JV is deliberately designed to pass those tests in a way that keeps the project debt off Meta’s balance sheet.
Consolidation of the JV / SPV
The JV and Beignet Investor are legally owned 80% by Blue Owl funds and 20% by Meta, with governance designed so Blue Owl’s side formally controls key decisions (board majority, veto rights, etc.).[2][4]
Consolidation rules for variable-interest entities look at “power over relevant activities” and “exposure to the majority of returns.” Meta can argue that Blue Owl governs financing, distributions, and asset disposition, while Meta is “just” a tenant and minority investor.[4][2]
If Meta and its auditors conclude Meta does not have the power to direct the activities that most significantly affect the SPV’s returns, Meta does not consolidate the entity, so the ~$27B of bonds remain at the SPV level.[2][4]
Legality Here Hinges On Substance
If documents and actual practice align with Blue Owl control, the non-consolidation treatment is allowed under the rules.[4][2]
Operating vs. finance lease
The lease is written as an initial non-cancellable term of about four years, with multiple renewal options that can extend total use to roughly 20 years.[5][6][2]
Standards ask whether it is “reasonably certain” that renewal options will be exercised and whether the lease conveys control and substantially all economic benefits of the asset’s life.[6][2]
By asserting that it is not reasonably certain Meta will stay beyond the four-year base term and pointing to the JV’s formal control, Meta can classify it as an operating lease instead of a finance lease, even though economically the campus is intended as a multi-decade AI hub.[5][6][2]
Again, this is legal if management’s judgment about “reasonable certainty” is defensible and disclosed, and if auditors sign off.
Residual value guarantee and contingent liability
Meta provides a residual value guarantee that backstops bondholders if Meta terminates early or does not renew for much of the project life.[7][6][5]
Accounting requires recognizing a liability for guarantees when it is probable that an outflow will occur and the amount can be estimated.[6][2]
If Meta and its auditors decide that triggering the RVG is not “probable” in the accounting sense, they can treat it as a contingent liability disclosed in notes rather than a large on-balance-sheet liability.[2][6]
The legality comes from those probability and judgment thresholds being baked into the rules; Meta is not inventing them.
Put together, this is a textbook, if aggressive, use of:
Non-consolidated SPVs with majority outside ownership.[4][2]
Short “enforceable” lease terms plus optional renewals.[6][2]
RVGs and contingent-liability thresholds.[5][2][6]
All of that is permitted as long as Meta’s disclosures are accurate and auditors agree that the judgments are within the standards’ bounds.
How This Differs from Enron’s classic tricks
Enron’s off-balance-sheet structures often crossed the line from aggressive to fraudulent because they violated or evaded the rules that existed at the time, not just pressed them.
Key differences:
Control and equity tests: Enron used SPEs (Special Purpose Entities) that appeared independent but were often effectively controlled by Enron insiders, with sham outside equity that did not truly bear risk. Modern rules tightened those loopholes by focusing on variable interests, risk, and power, not just nominal ownership. Meta, by contrast, is partnering with a real third-party asset manager (Blue Owl) managing institutional capital at arm’s length, so the outside equity is substantive.[3][8][2][4]
Round-tripping and price support: Enron used JVs/SPEs to move bad assets off its books at inflated values and used derivatives and guarantees to secretly support those values, while treating them as independent market prices. Meta’s Louisiana structure is financing a new build, not offloading existing toxic assets at fake prices. The economics (large capex, long-term rent, real power consumption) are not fictional, even if the accounting optics are optimized.[9][3][2][5]
Disclosure quality: Enron often disclosed SPEs in opaque, minimal ways, and investors lacked a clear picture of aggregate exposure until things failed. In the Meta case, the JV terms, size, role of Blue Owl, bond issuance via Beignet Investor, and the broad lease/guarantee structure have been widely described in filings and market commentary, even if the full risk picture still requires careful reading.[8][10][1][3][2][5]
Intent and misstatement: Enron used accounting assumptions and projections that were knowingly unrealistic (e.g., expected equity values to keep SPEs unconsolidated), and internal emails later showed awareness of the deception. There is no public evidence that Meta is misrepresenting facts; instead, Meta appears to be playing by the rules as written to preserve balance-sheet optics while still disclosing the arrangement.[3][8]
So, while both use SPVs (SPEs), JVs, guarantees, and non-consolidation, Enron’s core problem was often that the “independent” entities were not independent, the risks were concealed, and the accounting violated substance-over-form principles. Meta’s structure, on current information, fits within the post-Enron rule set that was designed to prevent exactly those abuses, albeit in a way that pushes gray areas of judgment.[8][3]
Where the Resemblance is Uncomfortably Close
Despite the differences, there are echoes that concern regulators and sophisticated investors:
Economic control vs. formal control: Enron argued it did not control SPEs while effectively dictating their actions through side agreements and insider control. Meta argues it does not control the JV while being the only realistic tenant, operator, and economic engine for a bespoke $27B AI campus, which gives it enormous de facto influence even if Blue Owl nominally controls governance.[3][8][2][4][5]
Risk transfer in name only: Enron’s SPEs often purported to offload risk that Enron had quietly guaranteed back. Here, too, Meta “off-loads” the asset and debt to an SPV, but then provides a residual value guarantee that effectively keeps a large slice of downside risk with Meta if things go wrong.[2][3][5][6]
Shadow leverage: Enron’s web of SPEs obscured how levered the economic entity truly was, which only became clear when conditions turned. A growing ecosystem of AI-infrastructure SPVs financed through private credit and backed by big-tech guarantees risks creating an analogous shadow leverage problem: the system relies on a few hyperscalers always performing and honorably rolling forward their obligations.[1][8][3]
The critical line is that, after Enron, the rules intentionally allow some off-balance-sheet project finance, but only when risk and control have genuinely moved.
Critics argue that in deals like this, the form (20% equity, outside control, four-year lease) may not reflect the substance (Meta needs, uses, and backstops the asset for decades), which feels “Enron-ish” even if it is not illegal in the same way.[8][3]
The Practical Implication: Legal but Optically Distorting
The clearest way to frame it:
Legal: The structure uses permitted SPV/JV project finance, applies lease classification tests, and uses contingent-liability thresholds defined in standards. There is outside capital at risk, and relevant features are described in market-facing documents.[10][1][5][6][2]
Not Enron-level fraud: There is no evidence of sham equity, insiders secretly controlling “independent” entities, or fictitious trades manufactured solely to hit earnings targets.[3][8]
But economically misleading if taken at face value: It allows Meta and peers to present relatively low reported leverage and capex while committing to tens or hundreds of billions of AI infrastructure obligations that live in SPVs, long-term leases, and guarantees that most equity investors only see in the footnotes.[1][8][2][3]
In other words, it is legal because the post-Enron rulebook is full of judgment-based gates that sophisticated players can navigate. It differs from Enron in that the entities and risks are more real and more clearly disclosed. Yet it resembles Enron in that the economic substance—who truly bears the long-term risk and control of AI infrastructure—is increasingly diverging from what the headline balance sheet shows.[1][8][2][3]
How to Protect Your Portfoliol Against Hidden AI Infrastructure Leverage
Everyday investors can protect against hidden AI infrastructure leverage risks by diversifying away from overexposed hyperscalers, scrutinizing off-balance-sheet footnotes, and favoring hedges that profit from leverage unwind scenarios.[1][2][3]
Scrutinize hyperscaler exposure in holdings
Limit concentration in Meta, Microsoft, Amazon, Google, and Nvidia, as these firms drive AI capex and increasingly use SPVs/JVs to mask true leverage.
Review portfolio weightings with your advisor: Cap any single hyperscaler at 5% or less; aim for under 20% total Big Tech exposure if holding broad indices.[2][3]
Check ETF holdings: Funds like QQQ or VGT have heavy hyperscaler tilts; rotate some into equal-weight tech (e.g., RSPT) or non-tech growth (e.g., XLP, XLI).[2]
Favor cash-flow proxies: Look for companies with transparent capex (e.g., utilities financing data centers directly) over those relying on private-credit shadows.[1][2]
Read Footnotes For the Real Leverage Picture
Off-balance-sheet risks hide in MD&A, lease notes, and commitments sections, not headlines.
Annual/quarterly filings: Scan “Commitments and Contingencies,” “Leases,” and “Variable Interest Entities” for JV/SPV mentions, residual guarantees, or long-term offtake volumes. Meta’s next 10-K will detail the Louisiana deal’s terms.[4][5]
Key red flags: Large operating leases with renewal options, minority JV stakes in capex-heavy projects, or guarantees tied to AI infrastructure.[5][6]
Tools for everyday use: Free platforms like Yahoo Finance or Seeking Alpha aggregate note summaries; set Google Alerts for “[company] data center JV” or “residual value guarantee.”[5][1]
Build Hedges Against a Leverage Shock
AI shadow debt amplifies downturns if growth slows or rates rise, so position for volatility.
Short-duration bonds/Treasuries: Hold 20-30% in 1-3 year U.S. Treasuries (e.g., SHY ETF) to capture yield without duration risk if private credit spreads blow out.[3][2]
Self-Custody Bitcoin: Hold 10% in BTC (direct self-custody) for the asymmetric upside that is inevitable as the cycle plays out through the “A.I. Pause period” (read more on that here in my collaboration with
), since BTC has near zero probability risk of going to zero and staying there, and is the only digitally native hard asset collateral for the future economy. Its near-term price volatility is more a reflection of liquidity issues in the fiat system than anything fundamentally wrong with Bitcoin itself.Volatility protection: Allocate 5-10% to tail-risk ETFs like TAIL or SWAN, which hedge equity drops from hidden leverage realizations.[2]
Inverse/hedged tech: Use small positions (2-5%) in SQQQ or TECS for tactical shorts during AI hype peaks, but pair with stops.[3][2]
Gold/Silver/commodities: 10-15% in GLD, SIL, or DBC as non-correlated ballast, as energy/power strains from data centers could spike input costs.[3]
Diversify into AI-Adjacent Winners Less Prone to Capex Tricks
Shift toward AI beneficiaries with cleaner balance sheets or different risk profiles.
Stress-Test and Rebalance Quarterly
Run simple scenarios to quantify risk.
Use free tools like Portfolio Visualizer: Model a 20-30% hyperscaler drop if AI capex reveals $800B private-credit needs by 2028.[2][3]
Rebalance rule: Trim winners above 7% allocation; add to hedges if VIX >25.
Long-term tilt: Increase small-cap value (e.g., AVUV) and international (e.g., VEU) to 20-30% for decoupling from U.S. tech leverage cycles.[3]
This approach keeps everyday portfolios simple, defensive, and positioned to weather an AI financing unwind without needing constant monitoring.[1][2][3]
Stay sane, healthy, and wealthy out there!
~Chris J Snook
Endnotes
Meta–Blue Owl Louisiana deal, structure, and AI project finance
[1] Meta forms joint venture with Blue Owl Capital for Louisiana data center
https://www.reuters.com/technology/meta-forms-joint-venture-with-blue-owl-capital-louisiana-data-center-2025-10-21/
[2] Meta partners with Blue Owl Capital on $27 billion AI data center project
https://www.cnbc.com/2025/10/21/meta-blue-owl-capital-partner-on-27-billion-ai-data-center-project-.html
[3] Meta announces 4 million sq ft, 2GW Louisiana data center campus
https://www.datacenterdynamics.com/en/news/meta-announces-4-million-sq-ft-louisiana-data-center-campus/
[4] To land Meta’s massive $10 billion data center, Louisiana pulled out all the stops. Will it be worth it?
https://www.cnbc.com/2025/06/25/meta-massive-data-center-louisiana-cost-jobs-energy-use.html
[5] Meta - LED | Louisiana Economic Development
https://www.opportunitylouisiana.gov/metadatacenter
[6] How Louisiana made the deal for a Meta data center | Business News | nola.com
https://www.butlersnow.com/wp-content/uploads/2025/01/How-Louisiana-made-the-deal-for-a-Meta-data-center-Business-News-nola.com_.pdf
[7] Meta Selects Northeast Louisiana as Site of $10 Billion Artificial ...
https://www.opportunitylouisiana.gov/news/meta-selects-northeast-louisiana-as-site-of-10-billion-artificial-intelligence-optimized-data-center-governor-jeff-landry-calls-investment-a-new-chapter-for-state
[8] The largest Meta data center yet brings big impact to Louisiana
https://datacenters.atmeta.com/richland-parish-data-center/
[9] Meta Richland Parish Data Center
https://www.dpr.com/projects/meta-richland-parish-data-center
[10] The behind-the-scenes story of how Meta’s $10 billion data center ...
https://www.butlersnow.com/news-and-events/the-behind-the-scenes-story-of-how-metas-10-billion-data-center-came-to-louisiana
[11] Meta Offers Residual Value Guarantee for $26 Billion AI Data ...
https://www.ainvest.com/news/meta-offers-residual-guarantee-26-billion-ai-data-center-deal-2509/
[12] “Off-balance-sheet financing” without “consolidation”! Can ...
https://news.futunn.com/en/post/65359939/off-balance-sheet-financing-without-consolidation-can-meta-s-data
[13] PIMCO Anchors $27 Billion in Debt For Louisiana Meta Data Center
https://www.linkedin.com/posts/joruffolo_pimco-anchors-27-billion-in-debt-for-louisiana-activity-7386028196443570176-2f45
[14] “Implicit Guarantee” but “Not Consolidated”! Can Meta’s ...
https://longbridge.com/en/news/267256528
[15] How three AI megadeals are rewriting Wall Street’s playbook
https://www.moneycontrol.com/world/how-three-ai-megadeals-are-rewriting-wall-street-s-playbook-article-13671359.html
[16] Meta and Blue Owl close $27B data center deal with innovative ...
https://www.linkedin.com/posts/christian-schmidt-ai-cco-cro-cdo-growth_privatecredit-aicompute-hyperscalerfinance-activity-7387701821395529728-Qrfs
[17] The Shadow Ledger of AI Giants: Where Has the Debt Gone?
https://www.itiger.com/news/1133549465
[18] The coming debt deluge? - MBI Deep Dives
https://www.mbi-deepdives.com/the-coming-debt-deluge/
[19] The New Shadow System: AI Finance and the SPV
Legal structure, GAAP/IFRS treatment, and Enron comparison
[1] PIMCO Anchors $27 Billion in Debt For Louisiana Meta Data Center
https://www.linkedin.com/posts/joruffolo_pimco-anchors-27-billion-in-debt-for-louisiana-activity-7386028196443570176-2f45
[2] “Off-balance-sheet financing” without “consolidation”! Can ...
https://news.futunn.com/en/post/65359939/off-balance-sheet-financing-without-consolidation-can-meta-s-data
[3] The Shadow Ledger of AI Giants: Where Has the Debt Gone?
https://www.itiger.com/news/1133549465
[4] “Implicit Guarantee” but “Not Consolidated”! Can Meta’s ...
https://longbridge.com/en/news/267256528
[5] Meta Offers Residual Value Guarantee for $26 Billion AI Data ...
https://www.ainvest.com/news/meta-offers-residual-guarantee-26-billion-ai-data-center-deal-2509/
[6] The coming debt deluge? - MBI Deep Dives
https://www.mbi-deepdives.com/the-coming-debt-deluge/
[7] Meta and Blue Owl close $27B data center deal with innovative ...
https://www.linkedin.com/posts/christian-schmidt-ai-cco-cro-cdo-growth_privatecredit-aicompute-hyperscalerfinance-activity-7387701821395529728-Qrfs
[8] The New Shadow System: AI Finance and the SPV
[9] Meta announces 4 million sq ft, 2GW Louisiana data center campus
https://www.datacenterdynamics.com/en/news/meta-announces-4-million-sq-ft-louisiana-data-center-campus/
[10] How three AI megadeals are rewriting Wall Street’s playbook
https://www.moneycontrol.com/world/how-three-ai-megadeals-are-rewriting-wall-street-s-playbook-article-13671359.html
Sources – Everyday investor portfolio defenses and risk mitigation
[1] PIMCO Anchors $27 Billion in Debt For Louisiana Meta Data Center
https://www.linkedin.com/posts/joruffolo_pimco-anchors-27-billion-in-debt-for-louisiana-activity-7386028196443570176-2f45
[2] The Shadow Ledger of AI Giants: Where Has the Debt Gone?
https://www.itiger.com/news/1133549465
[3] The New Shadow System: AI Finance and the SPV
[4] Meta Offers Residual Value Guarantee for $26 Billion AI Data ...
https://www.ainvest.com/news/meta-offers-residual-guarantee-26-billion-ai-data-center-deal-2509/
[5] “Off-balance-sheet financing” without “consolidation”! Can ...
https://news.futunn.com/en/post/65359939/off-balance-sheet-financing-without-consolidation-can-meta-s-data
[6] The coming debt deluge? - MBI Deep Dives
https://www.mbi-deepdives.com/the-coming-debt-deluge/