Grandfathering

When a company introduces new rules, this usually applies immediately to all new customers. Customers who were already on board beforehand sometimes get an exception: their old terms remain in place, even though everything around them has changed. This principle is called grandfathering. The term originates from American English and derives from the so-called Grandfather Clause — a contractual clause that exempts existing parties from new obligations. Anyone who falls under this exception is referred to as “grandfathered.”

Grandfathering as a customer retention tool

Companies use grandfathering to avoid losing loyal customers. Anyone who has used a service for years reacts particularly sensitively to price increases or limited features. A grandfathering arrangement reduces this resistance: the customer feels rewarded instead of punished.

At the same time, it allows the company to test new conditions on the market. New customers pay the new price, existing customers do not. If the new model proves successful, the grandfathering arrangement may eventually be phased out — or it may remain permanently in place because it was communicated as a sales promise. Both variants occur in practice.

A common misunderstanding: grandfathering is not a right, but a promise. Whether and for how long it applies is usually stated in the fine print. Under certain circumstances, companies can revoke it unilaterally — this too is governed by the terms and conditions.

What grandfathering looks like in practice

The most common scenario: a service raises its monthly price from 10 to 15 euros. Anyone who signed up before a certain cutoff date continues to pay 10 euros — as long as they don’t cancel their subscription and sign up again. As soon as they do, they lose their grandfathered status and pay the new price.

It works similarly with API access to AI models. A provider can freeze old pricing plans for existing developers, while new users only get access to more expensive or more limited tiers. This can result in two developers using the same model but having very different terms — depending on when they registered.

Grandfathering also applies to usage rules. A company introduces a data protection policy that prohibits certain data processing — but it only applies to new projects. Existing applications that already use this data are allowed to keep running. This is often politically more convenient than an immediate ban, which would shut down many running systems all at once.

Grandfathering in tech news and AI regulation

In reports about platforms like OpenAI, Google, or Adobe, grandfathering comes up regularly. When a service changes its terms of use for AI training data, the question often arises: does this apply retroactively, or are existing users protected? The answer to this determines how much economic damage developers suffer.

The concept also plays a role in European AI regulation. The EU AI Act, the European Union’s AI regulatory framework, contains transitional periods. Systems already on the market only have to meet the new requirements at a later point in time than new systems. This is grandfathering at the legislative level.

For consumers, it’s worth paying attention to grandfathering promises when signing a contract. Sometimes an older, cheaper plan is worth more than a new offer with more features — because it acts as a shield against future changes. Knowing and preserving one’s status can save real money in the long run.