Elon Musk spent $225 million buying the government that regulates him. That figure is The Wall Street Journal’s accounting of what the richest man in the world put into this month’s elections in Texas, and it is an invoice, not a figure of speech. Roughly $160 million of it went to a super PAC — the campaign-finance instrument constructed precisely so that a single donor may give without limit — built to carry Ken Paxton, the state’s attorney general, into the United States Senate. Paxton lost to James Talarico, who accused Musk of trying to “buy his own personal U.S. Senator,” a sentence more precise than campaign rhetoric usually manages and a charge the spending itself corroborates — a backfire my colleagues had already tallied as the arithmetic predicted. The election did not deliver a receipt, exactly. It delivered the auditors: Gina Hinojosa, who defeated the departing governor, Greg Abbott, has sued Abbott for the correspondence between his office and Musk, and has asked that the state’s discussions about the land deal at Boca Chica State Park involving SpaceX wait until she is seated to conduct them.

What stood to be renewed was not small. Austin holds Tesla’s largest American electric-vehicle factory and, the company hopes, a future site of mass humanoid-robot production; SpaceX’s manufacturing and launch footprint covers much of South and Central Texas; there is a planned multibillion-dollar semiconductor plant called Terafab, and a hundred-million-square-foot facility on the books for Grimes County. The purchase did not clear every department this cycle. It had cleared enough of them, for long enough, that the incoming government’s first order of business is to establish what was signed while nobody was reading.

It is true, in the narrow sense in which these things are always narrowly true, that Musk never wrote a cheque payable to the State of Texas; the money went to candidates and committees. The trouble is that a state is not bought by cheque anyway. It is bought in instalments of land, law, and silence, and Texas had been accepting instalments for years before this cycle’s invoice arrived.

Musk relocated to Texas in 2020, and the state set about proving a point about deference. Laws were rewritten — including legislation that makes his companies difficult to sue. A new business court was created, and its critics’ account of its design — friendliness to controlling shareholders such as Musk — is the account that matters. For all its faults, Canada’s Investment Canada Act at least requires that somebody in Ottawa be notified when a foreign buyer takes control of a Canadian business — a notification, nothing more, but on the books; Texas had no equivalent form for what was happening to its own government.

Consider the land. In September, a federal judge approved a Trump administration decision to transfer about seven hundred acres of a federal wildlife refuge to SpaceX, over the objections of conservation groups. The state remains in discussions over a land deal involving SpaceX at Boca Chica State Park — the arrangement in which the public sells its own park to a tenant — and Hinojosa has asked that the discussions be postponed until there is someone at the table who did not take the money. None of these items was a favour to “economic development” in the abstract. Each was a favour to a specific tenant.

Read what the machinery does rather than what it is called. A business court is a specialized forum in which disputes over corporate governance are heard by judges who hear little else; building one to the specifications of a controlling shareholder amounts to building a courthouse where the controlling shareholder arrives with home-field advantage. And the open-records desk is the quietest and most load-bearing item in the inventory. As the reporting records it, the governor’s office developed a practice of accommodating requests from Musk’s companies to censor public records containing information about their business. The mechanism deserves translation. A citizen asks the government for the government’s paper. The affected company marks the pages it would prefer the public not read. The governor’s office concurs, and the public receives a document with the interesting parts removed. That is not an accommodation. That is a service performed by the custodian of the public’s own memory for the state’s most favoured applicant.

Consider the money. Tax incentives and infrastructure benefits; no corporate income tax; no tax on individuals’ capital gains; and, at the Texas Commission on Environmental Quality — the agency that approves large industrial projects — a commission whose members the governor appoints. “His projects are being approved in a way that other projects have not been,” as the state representative Mary González put it. It is true that every state courts large employers, in the narrow sense in which states compete for investment. The trouble is that Texas did not court this one. It incorporated it. Cory Doctorow’s observation about concentrated power — “When a company is bigger than the government, it gets damned hard to credibly threaten to punish that company” — describes a failure mode of regulation. What the record documents in Texas is the success mode: a government that did not need to be threatened, because it had already been retained. The incoming comptroller, Sarah Eckhardt — the officer whose job is to count — puts it plainly: “very big money has gotten a rubber stamp in Texas for a long time.”

It is true, and better conceded here than left to his defenders, that Texas received things for all this. Eckhardt, who dealt with Musk’s projects as a county official in Austin and later as a state senator, credits the companies with “good-paying career opportunities for people without college degrees,” and she is not wrong. The trouble is that the concessions were priced by the buyer. Texas gave up not only taxes but the surrounding apparatus that makes a development deal a deal: zoning, environmental review, litigation risk, and the public’s ability to read the correspondence that arranged it all. Subtract those and what remains is not a partnership. It is a purchase, with a civic discount. A productive partnership — which is what Eckhardt says the state is about to have — requires that both parties be able to say no, and the architecture of these years was constructed so that Texas could not. Work delivered under one buyer’s terms is not a market. It is a franchise.

Follow the benefit, because the purchase had a price and the purchaser did not pay it. It was paid in Grimes County, where residents showed up at the county commission in June to say, in the careful way rural Texans say such things, that a hundred million square feet would end a way of life they had not been consulted about — performing, in the smallest available venue, the review the state had already declined to perform. It was paid by small county governments, of the kind Eckhardt knows from the inside, pressed by Musk’s companies to move quickly on massive projects with few resources to negotiate the terms. It was paid by the school districts and the hospitals, which must be funded by taxpayers in a state where some of the largest employers pay no corporate income tax at all — the grievance the state representative Gene Wu named when he said his constituents are finished with “the ultrarich continuing to get handouts.” And it was paid by every Texan who is not the buyer, which is the constituency at the bottom of the transaction. Louis Brandeis put the arithmetic: we may have democracy, or we may have wealth concentrated in the hands of a few, but we cannot have both. Texas has spent years running the experiment. The results are in.

One observation from the trade, offered once because the structure is the same: my father’s mill in Selkirk was bought in 1995 by a company from Porto Alegre, and the town that held it spent the years after learning what it means to have an asset the size of your town owned by someone who does not live in it. Several of my uncles did not keep their jobs, and no one consulted them on the restructuring, because the contract had been written in Porto Alegre. The difference between a development deal and a purchase is who writes the contract. Selkirk did not write its contract. What the buyers never managed to purchase was the government of the town: the mill could be stripped, but the mill’s town could still, eventually, vote. Selkirk never sold the town hall to the buyer. Texas threw it in.

Harold Innis built Canadian political economy on the observation that a hinterland arranges itself around whatever staple it exports — fur, cod, timber, wheat — and that the staple’s requirements quietly rewrite the hinterland’s laws, its roads, and eventually its politics. The staple Texas is being organized around now is not wheat. It is land, electricity, and silence: acreage for launch pads and a planned semiconductor fab, grid connections for factories and data centres, and a public record kept conveniently thin. The surplus flows out of the hinterland; the institutions reshape themselves around the extraction. Musk’s own post-election contribution to the discussion — “data centers in space looking pretty appealing right now,” posted to the platform he owns — reads, in this frame, as a staples exporter reminding the colony that other colonies exist; in plainer terms, it is the sound of a tenant reminding a landlord that the building is not the only address in the market.

This month, the purchase met a counterparty. Texas was the largest line item in a national bet that lost everywhere it was placed, and the scrutiny of Musk’s contracts was already being drafted before the votes were counted, which is as it should be: contracts do not become secret because the administration that signed them lost an election. Musk told the Journal in 2020, explaining his flight from California, that “if a team has been winning for too long, they do tend to get a little complacent, a little entitled, then they don’t win the championship anymore.” He described the mechanism accurately. He assumed he was describing someone else’s team. For four years, one man’s companies spanned launch infrastructure, vehicle manufacturing, a planned fab, and a satellite-internet constellation, and the state’s bargaining position was a loading dock left open. The election was, in the only terms that matter for the next two years, an enforcement event: for a while, the government is bigger than one company. Governments turn over, and no oversight is permanent — which is why the durable remedies are the ones written into law and into the record, not the ones that depend on who occupies the office this term.

It is true that all of this was lawful — the state’s lawyers had a point, in the narrow sense in which lawyers usually do: the refuge transfer survived a federal judge’s review, the incentives were legislated, the court was created by statute. The trouble is that lawfulness was the product being purchased. The purpose of rewriting the rules and building a friendly court is to move the transaction inside the boundary of the law, so that no one can object to it without being accused of objecting to law itself. And the records practice does not enjoy even that defence, since it proceeds, as the record describes it, “despite the state’s open records laws” — an accommodation of a private party’s preferences by the office that is supposed to be the public’s counterparty in every one of those requests.

The principle Tim Wu argues in The Curse of Bigness — that sheer size, whichever company happens to hold it, is itself the thing a republic must manage — does not require punishing Musk’s companies for building things; it requires that they build them under the same rules as anyone else, enforced by a government that has read its own correspondence. Eckhardt put the second half in one sentence: on competition between Amazon and Starlink, “we will not put our thumb on the scales for one or the other.” The specific work follows from that line and is mostly reading: release the records the lawsuit seeks, let the Boca Chica deal lapse rather than complete it, seat commissioners at the Texas Commission on Environmental Quality who treat the agency’s name as a description of its job rather than as its marketing, and give counties the state backstop Eckhardt has asked for when they ask for it. The new government has already begun putting his contracts and influence under scrutiny; the records lawsuit is that work’s founding document.

Records lawsuits are slow, and discovery moves at the speed of the thing being discovered — here, four years of a governor’s correspondence. Let it move. Open-records laws exist so that the public can audit its own government’s receipts; for these years, the reporting indicates, the buyer’s preferences governed what the public was allowed to see. A wire transfer can buy a government. It does not survive discovery. A buyer who spent $225 million already knows what the paperwork says. It is the seller who needs the copy.