API Economy
The API economy describes a business model in which companies rent out their software functions to other firms via technical interfaces instead of selling them to end customers. Payment is usually made per call, which is why large parts of today's apps consist of purchased building blocks from other providers.
Today, software can not only be sold to people, but also to other software. To make this possible, a company provides what is known as an interface, or API for short: a fixed point on the internet to which an external program can send a request and receive a response. An online shop, for example, sends card data to a payment service and receives confirmation back. The shop doesn’t need to build its own payment system for this—it rents the function. When many companies depend on one another in this way and money changes hands as a result, this is referred to as the API economy. It is therefore not a technical procedure, but a way of doing business.
Why hardly anyone builds everything themselves anymore
The economic appeal lies in the division of labor. A startup with five people could never develop its own mapping application, its own payment system, and its own shipping process. Instead, it buys these components individually and focuses on what really makes its app unique. What used to take years can now be accomplished in just a few months.
This also pays off for the interface provider. They have already built the software once and can sell it as many times as they like, without needing new factories or stores. Each additional customer costs almost nothing yet still generates revenue. This is why interface providers are among the most profitable software companies of all.
However, there is a downside that is often underestimated. Anyone who buys in their most important functions becomes dependent on someone else’s prices and rules. If a provider raises its fees overnight or blocks certain uses, one’s own product grinds to a halt. This has happened repeatedly in recent years, for instance when social networks suddenly started charging for their interfaces.
What gets billed per call
Technically, every use follows the same pattern. The requesting program sends a message to a fixed internet address, along with a personal key. This key is something like a customer ID: it tells the provider who is asking and whom to bill for the usage. The provider processes the request, sends back the response, and logs the transaction.
Billing is almost always based on quantity, not time. With language models, for example, you pay per token—that is, per word fragment you send and receive. A payment service takes a percentage of each transaction, a mapping service counts the number of calls. Typical amounts range from fractions of a cent per request. This sounds harmless, but with millions of users it quickly adds up to monthly bills in the six-figure range.
To prevent anyone from overloading the system, there are upper limits per minute or per day. Anyone who sends too many requests is throttled or rejected. The API economy is frequently confused with cloud computing, i.e., renting computing power. The difference: in the cloud, you rent empty machines; with an interface, you rent a ready-made capability.
From the weather app to OpenAI’s numbers
This principle comes up constantly in everyday life, even if it’s invisible. The weather display on your phone pulls its data from a weather service. The map excerpt on a restaurant’s website comes from a mapping provider. Logging in with a Google account on a third-party site is also an interface call.
In business news, the term mainly comes up in connection with AI. Companies like OpenAI or Anthropic hardly sell their models as downloadable programs at all, but almost exclusively as an interface. A large part of their revenue comes from other companies building the models into their own products. When analysts talk about the growth of such providers, they usually mean exactly these per-call revenues.
Politics is also getting involved these days. The European payment services directive requires banks to open up account data via interfaces to other providers, provided the customer consents. This is precisely what made it possible for financial apps to emerge that display multiple accounts from different banks together. The API economy is therefore not just a trend in the tech industry, but in part a legal requirement.
