The condition where switching away from a technology provider becomes prohibitively expensive because data, workflows, and pricing sit under the vendor's control.
Vendor Lock-In is the condition where switching away from a technology provider becomes prohibitively expensive because data, workflows, integrations, and pricing all sit under the vendor's control. Switching costs rise the longer the customer uses the product, which strengthens the vendor's position.
In AI, lock-in compounds quickly, because prompts, fine-tunes, embeddings, evaluation baselines, and the institutional knowledge of how to use the system all accrue to models the enterprise rents. The vendor can raise prices, change terms, or deprecate models, and the customer has little room to negotiate after investing that much in the rented system.
The structural alternative is ownership: Private AI running inside Secure Enterprise Boundaries, where the models, the data, and the accumulated intelligence remain assets of the enterprise. That trade-off is the core of the renting versus owning decision.
See also Private AI, Secure Enterprise Boundaries, Digital Transformation and Enterprise Architecture.