4 ms·

I think you’re missing the point. If I take $10 and give it to you for 10% of your company and you give it back to me to rent a room in my building I now own 1

I think you’re missing the point. If I take $10 and give it to you for 10% of your company and you give it back to me to rent a room in my building I now own 10% of your company for $0 outlay and you get to use the room in my building for your business, and in these arrangement the room is rented at cost. Everyone wins in this arrangement. This isn’t stupid, this isn’t offloading capital into income. This is an efficient way to maximize my investment by effectively using existing infrastructure spend and boosting my investments cash flow position by reducing their infrastructure spend.

nostrademons 3y ago
That illustrates exactly the point, though. Normally when you rent a room for $10, it implies that the room is worth at least $10 to the tenant, and the opportunity cost to the landlord was < $10, otherwise the transaction wouldn't have happened. And similarly, if someone invests $10 for 10% of a company at $100 pre, they're assuming that the discounted cash flows of the company are worth at least $100. People who analyze financial statements and trade in markets assume this to be true - it's fundamental to the price discovery mechanism. But if the actual cost to the tenant was $0, the value of the service might have been any number greater than $0. And similarly, if all the money in the investment comes back to the investor as revenue, the actual cost of their $10 investment wasn't $10, it was $0. You can't make inferences about what they actually value these services and investments at, because the prices were not set by a free competitive market, they were set by companies colluding (or cooperating, depending on framing) for mutual benefit. We saw the same issue with all the food-delivery startups that were offering 6 free meals - they booked the revenue and used it to justify outlandish valuations, and they booked the promotion as a marketing expense, but the root problem is that consumers wouldn't buy the product at list price, they were only buying it because of the investor-funded promotion. Or with Groupon - it was economically unviable for the businesses that participated, but they wouldn't find that out until after their revenue could be used to justify the next funding round and cash out in an IPO. Wash-trading is just the limiting case of this, where it's apparent to everyone that the economic value of the transaction is zero because there is no actual product being traded, but a self-dealing transaction is used to set a price that other participants in the market anchor off of. The tricky part is that sometimes this tactic can be economically viable, and offering discounted or financed products is a way to get consumers to get over the activation energy of trusting a new product, and once they've converted they become regular customers of the normal product offering and go on to spend a lot more. This is the point of a sale, or a rebate, or a promotion, or a financing deal. Which I guess answers my original question about why it's legal here but not for crypto wash-trades: it can sometimes result in true economic value being provided, and it's difficult to prove malice instead of incompetence. But the caution being offered in this thread is to not take revenue figures or valuations of AI startups or cloud providers at face value, lest you be the one who is incompetent.
- fnordpiglet 3y ago
  I’m not sure I follow your logic. There is a market for, say, cloud computing services. If it helps you could see it as an exchange of cloud computing services for equity, but the reality is the cash invested likely was not 1:1 exchanged to cloud computing and they had optionality. That’s somewhat irrelevant, but it’s still a real truth. Any money used to purchase cloud services has real expenses and capital allocations associated, real economic value for the company who exchanged equity, and the investment still sits as an asset and liability. It’s hard to argue an AI model company derives no value from GPU hours, or Amazon derives no value from owning equity in and locking an AI model company into their platform and business relationships. This isn’t a wash trade, which literally has no economic value, and it isn’t fake revenues as there is a real exchange of economic value for both parties. The fact the exchange involves at cost services isn’t nothing - it magnifies the investment as otherwise they would have to purchase the services at a higher cost. The idea that reserved capital equipment and supporting services have no material cost or value to Amazon is on the surface false - they have a very real cost, both opportunity and capital, especially in a time of high end GPU shortage. The idea that reserved high end GPU at cost has no value for an AI company is absurd. The fact it was done as cash gives flexibility in how the investment is used, and the fact it counts as a tiny amount of aws revenues is not materially relevant to anyone.

nostrademons 3y ago

That illustrates exactly the point, though. Normally when you rent a room for $10, it implies that the room is worth at least $10 to the tenant, and the opportunity cost to the landlord was < $10, otherwise the transaction wouldn't have happened. And similarly, if someone invests $10 for 10% of a company at $100 pre, they're assuming that the discounted cash flows of the company are worth at least $100. People who analyze financial statements and trade in markets assume this to be true - it's fundamental to the price discovery mechanism. But if the actual cost to the tenant was $0, the value of the service might have been any number greater than $0. And similarly, if all the money in the investment comes back to the investor as revenue, the actual cost of their $10 investment wasn't $10, it was $0. You can't make inferences about what they actually value these services and investments at, because the prices were not set by a free competitive market, they were set by companies colluding (or cooperating, depending on framing) for mutual benefit. We saw the same issue with all the food-delivery startups that were offering 6 free meals - they booked the revenue and used it to justify outlandish valuations, and they booked the promotion as a marketing expense, but the root problem is that consumers wouldn't buy the product at list price, they were only buying it because of the investor-funded promotion. Or with Groupon - it was economically unviable for the businesses that participated, but they wouldn't find that out until after their revenue could be used to justify the next funding round and cash out in an IPO. Wash-trading is just the limiting case of this, where it's apparent to everyone that the economic value of the transaction is zero because there is no actual product being traded, but a self-dealing transaction is used to set a price that other participants in the market anchor off of. The tricky part is that sometimes this tactic can be economically viable, and offering discounted or financed products is a way to get consumers to get over the activation energy of trusting a new product, and once they've converted they become regular customers of the normal product offering and go on to spend a lot more. This is the point of a sale, or a rebate, or a promotion, or a financing deal. Which I guess answers my original question about why it's legal here but not for crypto wash-trades: it can sometimes result in true economic value being provided, and it's difficult to prove malice instead of incompetence. But the caution being offered in this thread is to not take revenue figures or valuations of AI startups or cloud providers at face value, lest you be the one who is incompetent.

fnordpiglet 3y ago
I’m not sure I follow your logic. There is a market for, say, cloud computing services. If it helps you could see it as an exchange of cloud computing services for equity, but the reality is the cash invested likely was not 1:1 exchanged to cloud computing and they had optionality. That’s somewhat irrelevant, but it’s still a real truth. Any money used to purchase cloud services has real expenses and capital allocations associated, real economic value for the company who exchanged equity, and the investment still sits as an asset and liability. It’s hard to argue an AI model company derives no value from GPU hours, or Amazon derives no value from owning equity in and locking an AI model company into their platform and business relationships. This isn’t a wash trade, which literally has no economic value, and it isn’t fake revenues as there is a real exchange of economic value for both parties. The fact the exchange involves at cost services isn’t nothing - it magnifies the investment as otherwise they would have to purchase the services at a higher cost. The idea that reserved capital equipment and supporting services have no material cost or value to Amazon is on the surface false - they have a very real cost, both opportunity and capital, especially in a time of high end GPU shortage. The idea that reserved high end GPU at cost has no value for an AI company is absurd. The fact it was done as cash gives flexibility in how the investment is used, and the fact it counts as a tiny amount of aws revenues is not materially relevant to anyone.