Where are VCs investing their money? From a VC perspective

Almost everyone is building an AI application at the moment, but quietly a lot of VC funds have decided to stop investing at the application layer entirely. Now, one or two years ago, if you were an application layer business and you just put a wrapper on an LLM and you had a couple customers, and the chances are you could probably raise a decent round. But that's just not the case anymore. What changed is that the underlying models are just getting too good at generic tasks, and in fact, quite good at quite intelligent tasks. So what are VCs doing with their money? They're actually going a layer below the applications. The infrastructure level. These are the companies building the tooling for everyday application layer needs. Think data pipelines, evaluation frameworks, or security and compliance tooling, fine-tuning infrastructure. Essentially the stuff that's boring but mission critical. They're also still putting money into vertical AI. So this isn't general purpose AI like AI for everything, but applications that are built for specific industry needs. Normally what you need as a founder is really deep domain experience to actually understand how the application can be tailored for that specific niche. Think AI for construction or for insurance. Very specific workflows within those industries. These are defensible domain knowledge and the data training the models. So if you're an application layer business, you need to be asking yourself, could open AI, Anthropic, Google, Gemini, etc. would they release this and if they do, do you still have a business? That's the same question that a lot of VCs are asking themselves whenever they're considering investing in a founder.