The Cascade, and the Payer of Last Resort.

Yesterday the market measured the exit. The Nasdaq fell five percent, closed in a confirmed correction, and found the door exactly as wide as this series said it was. Today the harder question arrives, the one the first five papers were built to make unavoidable: if the loss keeps moving, whose name is on the last line?

The objection has been waiting since the series began, and it is correct in the way that matters. They will have to bail it out. The sovereign rescued the actors in 2008; it will rescue them again; no structure this large is allowed to finish falling. This paper does not rebut the objection. It files it — as the map’s final node.

A bailout does not create the missing counterparty retroactively. It drafts one.

Terms, fixed before the argument. A repricing is a claim re-marked: holders bear losses in order, and no one has yet failed to perform. A cascade is different in kind — a holder who cannot sell meets a liability that cannot wait, and a valuation loss converts into a payment failure that travels. A lender of last resort supplies liquidity against collateral to the illiquid but presumed solvent. A payer of last resort absorbs the loss itself. Liquidity can be repaid. A payer cannot. The distance between those two offices is the distance this paper measures.

The objection belongs to Dr. Estabrook — the author’s father, a school psychologist — who put it the week the first paper published: the government bailed them out in 2008, and it will bail them out again. The grade given it by this desk, from the first week: probably right.

What the objection misses is not its fact but its filing. A rescue is not an outside event that defeats the series. It is the series’ terminal state, arriving on schedule. The private structure has no counterparty — that was Parts One through Five. Therefore, if the loss becomes systemic, the counterparty is manufactured politically, after the failure, at public expense, under another name.

The rescue will not be called a bailout of artificial intelligence. It will be called financial stability, liquidity support, strategic compute, defense capacity, energy security, pension protection, industrial policy — whatever the season’s costume is. And what gets saved will be chosen precisely. Not shareholders, not marks, not bonuses, not the story. Liabilities and functions: payments, deposits, pensions, power, the funding markets themselves. The equity can be crushed to the floor while the obligations are carried out of the building. Both things can be true at once. In 2008, both were.

A cascade is a loss changing address. The route, in the order the mail arrives:

1. Equity takes the first mark. Households hold 48.2 percent of their wealth in equities, through the index funds and pensions that own the build-out whether they chose it or not. The terminal holder was inside the structure from the first paper. The first loss and the last one have the same address.

2. Credit re-prices the promise. Hyperscaler and supplier bonds, leveraged loans, and credit protection stop believing the equity story. Watch Oracle first among the landlords: investment-grade by one notch’s grace in this series’ receipts, carrying the build-out’s most borrowed balance sheet.

3. The shadow layer gaps. Neoclouds and SPVs — an estimated $662 billion of shadow borrowing in the BIS figures carried across this series — cannot refinance presumed demand. This is the layer with no deposit base, no regulator of its own, and no name the public knows. It is where a repricing learns to run.

4. Contracts come home. Leases, take-or-pay floors — $430.6 billion of Anthropic floors itemized from the counterparties’ own filings, against roughly $895.4 billion of aggregate commitments in the amended S-1; two measures, never merged — power purchase agreements, and residual-value guarantees convert a demand shortfall into payment obligations at exactly the wrong moment: when the paper value of everything pledged against them is falling.

5. The quiet holders discover their exposure. Pensions, insurers, private-credit funds, and banks find the build-out inside portfolios sold to them as infrastructure, cloud growth, diversified credit. None of them bought a speculation. All of them own one.

6. Funding decides. If repo and the money markets stay calm, the loss remains a repricing — painful, orderly, survivable. If funding refuses, every holder who cannot sell meets a liability that cannot wait, at the same time, in the same week.

7. The sovereign is presented with the bill — framed not as rescuing the build-out but as defending everything wrapped around it. Which, by then, is true. That is the trap’s final mechanism: by the time the bill arrives, paying it is the reasonable thing to do.

Against all of this stands one comforting number: the companies are worth trillions. It is not an answer. Market capitalization is not a cash reserve. It is a price opinion, and it falls fastest exactly when capital would need to be raised against it. Issuance into a falling market is dilutive, slow, and self-announcing — the attempt to raise the rescue confirms the need for it. Buybacks are procyclical, funded from the same balance sheets that borrowed for the build-out; they are largest when least needed and unavailable as a standing bid. The equity is the first loss absorber, not the rescuer. Its destruction is the event, not the cure.

When the bill is presented, it will not arrive as a single act. It will arrive as a menu, most of it pre-existing authorities, each item priced in a different costume:

Liquidity facilities — central-bank lending against widening collateral classes, if the funding markets seize. Guarantees — Treasury or agency backing for the refinancing of data-center, power, or semiconductor obligations, dressed as infrastructure or strategic capacity. Anchor tenancy — the state becomes the customer the prospectus always presumed was standing behind the demand: defense, intelligence, and agency compute contracts, at sustaining prices, for sustaining durations.

Regulatory time — forbearance on marks, capital, and accounting recognition, for the banks, insurers, and pensions holding the exposure. The cheapest bailout is being allowed not to look. Tax and rate channels — accelerated depreciation, credits, and utility rate structures that move the cost onto taxpayers and ratepayers in slices too thin to organize against. Currency — the residual channel. If the loss is monetized, the public pays as dollar-holders. The rescue is printed, and the printing is the payment.

Part One closed with the public in three costumes: the gambler, holding the equities; the casino’s creditor, holding the bonds, the pensions, the private credit; and the landlord’s mortgage holder, through the housing-and-finance chain under the landlords. Part Six completes the wardrobe:

4. The taxpayer — guarantees, facilities, procurement, and tax expenditures, voted or administered in the language of stability. 5. The dollar-holder — inflation and debasement, if the loss is absorbed by the currency: the one bill that arrives at every address simultaneously and cannot be unsubscribed from. 6. The ratepayer — power costs socialized through capacity auctions and rate classes: the meter from Part Five, arriving as a bill, for the load that rendered the feed.

The same household, billed at every layer of the structure, under six different names — none of them the name of the thing being rescued.

Filled after the close, November 6, 2026. This section was published before the bell with its decision rules fixed in advance; the closing statistics for Friday, November 6, 2026 are entered below and graded against those rules, unchanged. Nothing in the rules was revised. The data slot is dated, as promised.

The series began with a map of obligations that terminate in a presumption. It ends with the only name that was never on the map and was underneath it the whole time.

No counterparty signed. No counterparty priced it. No counterparty consented.

Receipts carried forward. No Counterparty dataset (25 nodes, 58 edges, 39 of 58 graded filed/SEC) · Anthropic take-or-pay floors $430.6B (counterparties’ filings) · Anthropic S-1/A aggregate commitments ≈ $895.4B (separate measure — never merged with the floors) · BIS ≈ $662B SPV/shadow borrowing · household equity share 48.2% (Federal Reserve Financial Accounts) · November 5, 2026 closes (Nasdaq ≈ −5%, confirmed correction; S&P 500 ≈ −3.4%; Dow ≈ −2.6%; Russell 2000 ≈ −4.6%) · November 6, 2026 payrolls (+165,000 vs ≈65,000 expected; unemployment 4.1%). Open slots: November 6 closes and credit read (above); Anthropic roadshow and pricing, week of November 9.