A VC due diligence expert explains why most investors can't read code and miss critical risks: hard-coded implementations that don't scale, and startups locked into expensive AI models like Claude or OpenAI whose pricing can shift overnight and break unit economics.
Okay. Just to give you some light on how we've been looking at these things. So we've been working with a lot of VCs actively where we conduct the AI due diligence for them. The reason is most VCs, you know, even if they're Harvard graduate, Columbia, or, you know, Oxford, the reason here is that they haven't read a line of code and a founder might walk in, give a very beautiful pitch that this is what the product does. This is very scalable. But if the VC has, does not understand the line of code, So just to give you an example, if it's hardcoded, it's not scalable. If it has got a lot of model dependency, then a slight change in pricing from the model, maybe you're using Claude, Sonnet, you're using Gemini, or maybe you're using DeepSeek, Groq, or OpenAI. A slight changing in pricing could, you know, take you to the ruins. So the risk here is a lot. And this is how we flag. We have a whole framework of mapping these and we do it and tell the VC, look, that if you invest in this, this is the risk factor here.
Get a mentor as soon as possible — and make sure it’s a mentor in your space.
Keith Gillispie: Why Most New Investors Fail & How to Avoid the Common Traps #250 | SaaS District
