The economist Friedrich Hayek made the argument in 1945 that a price is a signal. No single person knows why tin has suddenly grown scarce, but the moment its price rises, everyone who uses tin begins to economize, and the knowledge travels through the whole system without anyone ever holding it whole. Hayek called the price mechanism a kind of telecommunications network. A price, in his account, is information about the world, compressed into one number and broadcast to strangers who never meet.
That is the famous half. The half a salesperson learns on the floor is that a price also carries information about the seller, and often that is the only information in it. The number on the tag can say almost nothing about what the thing cost to produce, and almost everything about what the person who set it is trying to do.
Consider the most disciplined price in American retail. A hot dog and a soda at Costco have cost a dollar and fifty cents since 1985. The hot dog does not cost a dollar fifty to serve, and everyone involved knows it. When a chief executive once raised the idea of lifting the price, the company’s co-founder, Jim Sinegal, told him, in words the current chief executive still repeats, “If you raise the effing hot dog, I will kill you. Figure it out.” They figured it out by building their own hot dog plants rather than touch the number. That price is not a fact about a hot dog. It is a sentence about the whole company: this line holds at a loss because the line is the brand. The rotisserie chicken sits at the far back of the store for the same kind of reason, so that the trip to reach it runs past everything the margin actually lives on. None of that is cost plus markup. All of it is strategy, printed as a number.
The frontier AI companies are writing a different sentence in the same grammar. A great deal of what they sell is priced below what it costs to run, because the object right now is not margin, it is ground. Get the tool into every workflow, make leaving it unthinkable, and wager that scale and falling compute costs settle the account later. A price set beneath its own cost is not an error in the arithmetic. It is a declaration: we are buying the market and we will pay to be early. Read the price and you have read the plan.
A high price is just as talkative as a low one. In a market served by only a few suppliers, where the product is easy to copy but the capacity to produce it took decades to build, prices tend to settle into a similar and comfortable range. There is no room where this is agreed. The lawful version even has a name, conscious parallelism, and it is not collusion; it is simply what happens when a handful of players can each see that a price war has no winner. A fat margin in a market like that is not a fact about cost either. It is information, and the information is plain: the buyer needs this, there is nowhere cheaper to go, and everyone selling it knows both things at once.
None of this is guesswork on the seller’s side, which is the part the objection-handling forgets. No company brings a product to market without studying what everyone near it charges. Even a genuinely new thing has a cousin, some adjacent product the buyer already pays for, and the price is set in reference to it and lands inside a defensible range. So a price is not only information about the seller’s strategy. It is a compressed report of the competitive research that produced it, the one number that already survived the hardest audit it will ever face, the market’s, before it was ever printed.
Which lands on the strangest habit in the trade. The standard sales playbook now demands a method to defend the price against objections, as though the number were a soft spot to be guarded. But a price is not an argument to be won or lost. It is a signal the seller chose to send, and one that has already been checked against every rival in reach. Know what strategy it encodes and holding it becomes simple, because the work is no longer defending a figure, it is stating a position. The rep who reaches for a framework to justify the price has misread the object completely. They are treating it as the answer to a sum, when it was a decision all along, and a message on top of that.
The first people who should be handed that information are the reps who have to say the number out loud. Given a price list and nothing else, a rep can only recite. Given the reasoning, whether the margin is fat or thin, whether a line is a loss-leader, where it sits against the competition, they can do the thing the frameworks only pretend to teach. They can explain it, and know exactly where there is room to move and where there is none. Consider a seller who carries a partner’s data at cost, at no margin, and says so plainly, names the partner, explains there is nothing to give because nothing is being made. That is not a script. That knowledge did not come off a pricing sheet; it came from a phone call to the partner, and then from passing it on. A rep represents the number. That is no reason to keep them in the dark about what stands behind it.
A price is the most compressed thing in any deal and the most misread. It looks like the answer to a question about cost. It is usually the answer to a different question, the one the seller was actually asking: what am I trying to make happen, and what will I say, in a single number, to make it happen. Read it that way, and the objection a rep is trained to dread becomes the most informative sentence in the room.
Touchstones: Friedrich Hayek, “The Use of Knowledge in Society” (1945), on the price system as a mechanism for transmitting dispersed knowledge; the Costco hot dog account, attributed to Jim Sinegal and rated true by Snopes; conscious parallelism, the lawful cousin of collusion in a market of few sellers; and W. Brian Arthur on increasing returns and lock-in, the logic under pricing below cost to hold a new market.