live · 271 entities · 174 companies
intermediate

AI vendor risk assessment using Blast Radius Score

How to evaluate AI vendor dependency risk before it becomes a production problem — using sourc.dev data.

48% of tracked AI tools depend on two providers — OpenAI and Anthropic combined. If either raises prices significantly, changes their API, or has a significant outage, nearly half of the AI tool ecosystem is affected.

This is not a risk prediction. It is a verified measurement computed from active integration relations tracked on sourc.dev.

What Blast Radius Score measures

The BRS measures how many downstream entities are affected if a given entity has a significant incident. A high BRS means many other entities depend on this one.

How to use this for vendor decisions

Before adopting an AI tool, three questions worth answering:

What is the entity’s BRS? High BRS means the entity is a critical dependency for many others — less likely to disappear, but more affected when it has issues.

How many certified integrations does it have? More certified integrations = more embedded in the ecosystem.

What is the provider concentration? If a tool only supports one model provider, you inherit that provider’s risk entirely.

Building a stack with lower concentration risk

For each AI component in your stack, identify the alternative that could replace it within 30 days. If no alternative exists, that component is a single point of failure.

Start here: Find your most critical AI dependency on sourc.dev. Check its BRS and certified integration count.

This article is part of a growing knowledge track. More depth, examples, and detail added continuously.