The Invoice and the Cap Table
A vendor and an investor are supposed to be different parties, for a specific reason. One wants as much of your money as it can get. The other wants your company to be worth more. Point those two interests at each other across a negotiating table and you get something close to a price: not a true one, but a contested one, which is the only kind markets actually produce.
Collapse the two parties into one and the contest doesn't get harder, it stops happening. Say your supplier also holds equity in your company. Every dollar you pay them on the invoice is a dollar that, on paper, you were trying to minimize a minute ago. But that dollar doesn't vanish once it crosses the table. It shows up as revenue on the supplier's books, which raises the supplier's valuation, which raises the value of the stake they hold in your company, or the stake you hold in theirs, or both, if the ownership runs in both directions. The loss on the invoice and the gain on the equity line are the same event, wearing two different account names.
Nobody has to lie for this to work. The service gets delivered, the invoice gets paid, every line item is real. What's missing is the party whose job was to find the number too high. That party still exists in the org chart, still signs the purchase order, still has a budget. It just no longer has a motive, because the money it's defending and the money it's spending turned out to be the same pile.
The clean historical case is vendor financing in telecom, late 1990s. The big equipment makers lent money, sometimes billions, to the competitive carriers that were buying their switches and fiber gear, carriers nobody else would extend that kind of credit to on those terms. The equipment maker booked the sale as revenue the day the gear shipped. The loan was just the same cash moving to a different pocket of the same jacket, timed so the revenue landed before the debt came due. When a chunk of those carriers folded a year or two later, the loans went bad, but the revenue had already been recognized, already multiplied into a stock price, already paid out as bonus. The corrective mechanism, "will the customer actually be able to afford this," had been quietly replaced by "will we lend them the difference."
The same shape keeps showing up wherever capital-intensive infrastructure meets capital-hungry customers who happen to also be portfolio companies. A compute provider takes a stake in the lab that rents its machines. A chipmaker invests in the startup buying its chips. None of this requires malice, and in most individual instances nobody involved is doing anything they'd call wrong. Each side can tell a true story about why the deal makes sense on its own terms. The circularity isn't a decision anyone makes. It's a byproduct of two balance sheets agreeing to merge interests while keeping separate letterhead.
What actually breaks is downstream of the deal, not inside it. Third parties, outside investors, acquirers, credit raters, read the reported revenue as if it came out of an arm's-length negotiation, because that's the only kind of revenue number the accounting has a category for. There's no line item for "revenue generated by a party who also profits from the revenue being large." So the number gets priced as if it were discovered, when it was actually agreed to by two people who both wanted it bigger.
The fix people reach for is disclosure, label the related party, flag the circular flow. That helps an analyst who's paying attention. It doesn't restore the thing that was actually missing, which was never information. It was a party on the other side of the table with a reason to say no.