台账

The long bond, short — the one position that pays when the thing that marked the whole book keeps going. Every name on the board shares an assumption about the price of money; this makes it explicit and gradeable. The mechanism is the week's story: a hot inflation print, a Fed priced to hike into it, an August thirty-year auction that cleared at its highest yield in a quarter century and still tailed, and a deficit near two trillion. The fund's duration is just under fifteen years, so a quarter point on the long end is about three and three-quarters percent of price either way. The ten-year is already at the cusp of five and much of this is priced — the position is a hedge on the book's shared assumption, not a fresh bet against it.

Memory is a commodity cycle wearing a secular costume, and the cycle is turning where the wall says it should. The mechanism runs through the power bill: hyperscalers who cannot secure the megawatts defer the accelerators, the accelerator vendor holding thin inventory cuts its forward memory orders, and those cuts land just as the memory makers' new capacity arrives — margins built at peak pricing do not shrink in that world, they collapse. Micron sits at the receiving end of every link in that chain: it must spend heavily to build out production against orders it has not yet filled, its input is energy in an energy shock, and its product is priced by whichever competitor blinks first. The tells were on the tape by mid-July — the AI-capex rout ran straight through Asian semiconductor names, a Korean memory maker that listed in the United States at the top of the cycle fell more than eight percent in a week, and by the end of the month memory had sold hard with the sell-side turning. The expression is the stance itself, graded on the prints and deliberately not an options clock: a structural view on a theta clock is a donation, and the book buried that wrapper before it filed this thesis. The kill is pre-registered on the company's own numbers: two consecutive quarterly prints in which gross margin holds or widens quarter over quarter while next-quarter revenue guidance steps above the quarter just reported — a cycle that is not rolling over. The first scored test is the late-September print. Verdict due June 30, 2027.

Same cycle as Micron, without the secular story to hide behind. Storage is a commodity: the price is set by the marginal competitor's willingness to cut, the capacity comes on in lumps, and the customer is the same power-constrained buildout that is learning to defer deliveries. SanDisk had its run alongside memory and carries the same three burdens into the turn — heavy spending to build production against unfilled orders, an energy input repricing under it, and pricing it does not control — with none of the artificial-intelligence-demand narrative that lets a memory maker argue this time is different. That absence is the point of filing it separately: if the cycle read is right, storage shows it first and cleanest. It did on the tape — the week the AI trade cracked, this was the strongest grade on the demand side, down eleven percent on a single Friday. The expression is the stance, graded on the prints, and it shares one bet with the Micron filing under the book's concentration law: one thesis in two wrappers, sized jointly if ever sized at all. The kill is the same pair, on this company's own numbers: two consecutive quarterly prints in which gross margin holds or widens quarter over quarter while next-quarter revenue guidance steps above the quarter just reported. A cycle that stops rolling over ends the thesis. Verdict due June 30, 2027.

Volume-and-margin short on eastern rail, NOT a fuel trade — surcharges pass fuel through. Mechanism: distillate levy taxes discretionary goods volume; operating ratio deteriorates on falling carloads.

Server assembly is a toll paid, not collected; −6.8% on record AI-server news (Sep 1). Margin compresses as hyperscalers squeeze assemblers. Extension of the SNOW/CRM node.

The largest builder, short, as the price of money's second-order effect. The mechanism runs through the mortgage, not the house: a five-percent ten-year sets the thirty-year fixed near seven and a half, the builder's answer to that is the rate buydown, and buydowns are margin — so a sustained long end first thins gross margin on the orders a volume builder keeps, then raises cancellations when the buyer's payment math fails at closing. D.R. Horton is the meter on both: it reports net orders, cancellation rate and homebuilding gross margin every quarter, and it sells to exactly the buyer a five-percent ten-year prices out. The book already owns the rate leg through the long-bond short; this is the same thesis one step down the chain, where the rate becomes a household. The honest tension is printed: the packet's eight-percent mortgage is not on any screen the desk carries, claims fell to a quarterly low the week this was written, and a builder with a land-light model can hold volume by buying margin for a long time. The expression is defined-risk — a January 2028 put structure. The kill is the company printing the opposite twice, or the mortgage rate falling through six for a month. Verdict due January 21, 2028.

A volume-and-margin short — deliberately not a fuel trade, because fuel surcharges pass fuel costs straight through, and that leg was killed in review before the position existed. The mechanism runs through the customer instead: refined products have outrun crude all year — up sixty-six to seventy-eight percent against crude's thirty-two — and that gap is a tax collected at the pump from a consumer saving near three percent of income. Parcels are where the discretionary spending that tax crowds out becomes freight. FedEx's network is a meter on exactly that spending — package volumes are the demand line, adjusted operating margin is the operating-leverage line that compresses when volumes decelerate, and the trans-Pacific franchise adds a second exposure to the same slowdown arriving from the Asian side. The two lines this short lives on are the two lines the company must report every quarter. The expression is defined-risk by construction: a put structure with a January 2028 tenor, so the clock is bounded and the premium is the whole downside — a short whose worst case is known on entry. Falsified if volumes and adjusted margin both hold across two consecutive quarterly prints, or if the consumer trigger stays un-fired through mid-2027 — the book does not keep a short whose fuse never lights. Verdict due January 21, 2028.

Volume-and-margin short on LTL bellwether, NOT a fuel trade. Cass shipments −4.8% y/y; tonnage declines are the mechanism; GRIs/surcharges are the counter-mechanism to grade against.

A cruise ship is a consumer-discretionary purchase that burns fuel and floats on borrowed money, and the wall reprices all three at once. The mechanism has three legs: refined products have outrun crude all year — a tax collected at the pump from a household saving near three percent of income, which is the household that books a cabin; the fuel complex that moves a ship printed records in July; and the long end of the curve went through five percent in mid-July, repricing every dollar of the debt a fleet is financed with. Royal Caribbean carries all three — its demand is discretionary and booked in advance, it burns the fuel, and it carries the debt — so when the consumer thins out it must cut ticket prices into rising operating costs, and the squeeze shows up in net yields. The tells came in one week in late July: consumer discretionary fell five percent in a session, Brent touched a hundred, and the long bond sat at highs last seen in 2007 — the debt leg doing the work. The expression is long-dated puts, because booked demand is a slow fuse and the clock must outlast the fuse. The kill is pre-registered on the company's own line and on the fuse: two consecutive quarterly prints with net yields up year over year and full-year earnings guidance held or raised, or a consumer credit trigger that never fires by mid-2027 — the book does not keep a short whose fuse never lights. Verdict due January 21, 2028.

Legacy auto sits where three of the wall's forces cross, and it is the one name on the demand side that was struck once for saying so too loosely — so this filing says it precisely. The mechanism: energy costs land on a legacy manufacturer and on the buyer of what it sells; the electric transition is being won on price by Chinese producers whose margins no legacy maker can match, and that pressure reroutes to wherever the tariff walls do not reach; and the marginal car buyer is the subprime borrower whose delinquency printed a record for the series since 1994 in January and whose cooling since is all that stands between this thesis and its trigger. Stellantis is exposed on all three at once — a legacy footprint, an electric transition it must fund against the Chinese cost curve, and a customer financed at the bottom of the credit stack. The three legs are mechanism, not verdict: the thesis is graded on two of the company's own series, read together, and the China-share leg is carried as color until a series is named for it. The tells so far are sector tape, not company prints — discretionary down six percent the week the AI trade cracked, and every payrolls Friday a read on the credit leg. The expression is long-dated puts, one bet shared with the other consumer wrappers under the concentration law. The kill: two consecutive results reports with adjusted operating income margin flat or better year over year and shipments up year over year, or a consumer credit trigger that never fires by mid-2027. Verdict due January 21, 2028.

The foundry at the end of every forward order. The mechanism is the order book, not the wafer: leading-edge capacity is sold years ahead against the labs' and hyperscalers' training plans, and if those plans are paced — by a regulator, a grid that cannot deliver the gigawatts, or a budget that funds debt service first — the cancellation shows up in the next quarter's advanced-node revenue and in the margin a full fab earns and a half-full one does not. TSM is the single point where every merchant AI-hardware order clears, which is why it is the cleanest expression of hardware urgency dying — and the most dangerous, because the same position was in effect on the week this was written and lost: the memory and GPU names rose, and the primary source for the pacing call disclaims any halt. The book files it dated, because a short with a printed falsifier is how the room learns whether the regulatory-freeze story ever reaches a fab. The expression is defined-risk by construction — a January 2028 put structure, worst case known on entry. The kill is two prints of the company saying the opposite: advanced-node revenue growing year over year with gross margin held or better, twice in a row. Verdict due January 21, 2028.