Assemblage: Building a Campus Parcel by Parcel Without Moving the Price
The best sites for large computing campuses are frequently too big to exist as a single parcel under a single owner. A five-hundred-megawatt campus with room for substations, setbacks, and future phases may need six hundred, eight hundred, a thousand contiguous acres, and in most of the country that footprint spans several separately owned properties, a few farms, a woodlot, an inheritance split among siblings. Assemblage is the craft of stitching those separate parcels into one developable whole. It is where a great deal of the value in this business is created and, done carelessly, destroyed.
The central problem of assemblage is the one that has haunted every land assembler from railroads to shopping malls: the holdout. If sellers learn that a single buyer needs all of their parcels to make a large valuable project work, the last few sellers realize they have enormous leverage. The buyer who has already secured eight of ten parcels cannot walk away over the ninth without stranding the entire investment, so the ninth and tenth owners can demand prices wildly above what their land is worth in isolation. Every seller who senses the assemblage tries to be the last holdout. Prices ratchet up parcel by parcel, and the economics that made the site attractive can evaporate before the last deed is signed. The entire art of assemblage is structured around preventing this dynamic from taking hold.
The first defense is discretion. Assemblers work quietly, often through intermediaries, buyers' agents or single-purpose entities with anonymous-sounding names, so that no individual seller realizes they are one piece of a much larger and more valuable puzzle. Each parcel is negotiated as if it were a standalone purchase at its standalone value. A farmer selling eighty acres to what appears to be an agricultural investor or a small local developer has no reason to demand a computing-campus premium, because as far as they know they are selling farmland to a farmland buyer. The moment word spreads that a hyperscaler is assembling a campus, every remaining price in the area re-rates upward, so professional assemblers guard the project's identity and purpose fiercely, sometimes using multiple shell entities so that even a curious seller cannot connect the dots between transactions.
Options are the assembler's essential tool here, for a reason specific to the holdout problem. If you buy parcels one at a time outright, you expose yourself: after the third or fourth purchase, someone notices a pattern, and the remaining owners raise their prices. Worse, if the assemblage fails, you are stuck owning a handful of non-contiguous parcels that are worth far less scattered than they would have been whole. By securing each parcel with an option instead, you control the entire footprint while committing only premiums, and, critically, you can make each option contingent on assembling the whole. Well-structured assemblage options let the developer walk from all of them if any essential parcel cannot be secured, so you never end up owning a useless fragment. You tie up the whole board quietly, confirm you can complete it, and only then exercise across all the parcels more or less at once, before anyone can hold out, because by the time the pattern is visible the deals are already locked.
Sequencing and parcel importance require judgment. Not every parcel carries equal weight. The parcel that contains the interconnection point, or the one without which the site loses road frontage or the acreage needed for the substation, is a must-have, and the assembler identifies these keystone parcels early and secures them, or firm options on them, before spending heavily elsewhere. There is little sense in optioning eight easy parcels if the one keystone parcel, the one the whole site depends on, turns out to be unobtainable because its owner will never sell. Experienced assemblers pursue the hardest and most essential parcels first, so that they learn early whether the assemblage is even possible before sinking premiums into the easy ones. Conversely, some parcels are nice-to-have expansions, and the developer may design the campus to work with or without them, preserving flexibility and reducing exposure to any single stubborn owner.
Assemblage also intersects with the power and permitting work that runs in parallel. The assembler is usually advancing the interconnection application and the entitlement process while the land is being stitched together, which means the parcels must be controlled firmly enough to file those applications but the capital exposure kept low enough to survive a failure. This is a delicate balance of legal control versus financial commitment, and it is why assemblage at this scale is practitioner-level work rather than something a first-timer should attempt. The developer is running three clocks at once, land, power, and permission, and must keep them roughly synchronized so that the assemblage completes about when the power path firms up and the permits land, without over-committing to any one of the three before the others are secure. Faraday's read: assemblage is won or lost on discretion and sequencing. Move quietly, option before you buy, secure the keystone parcels first, and make every deal contingent on the whole, so that no single owner ever learns they hold the piece you cannot do without. The assembler who tips their hand pays the holdout tax, and the holdout tax has killed more good sites than bad power ever did.