Source item

This is a general comment about software companies: If you look across the software industry, you see record levels of RPO - these are future revenues that are tied to a performance obligation before you can recognize it and before a customer has to pay you. The problem with RPO is that AI can more easily attack and devour it - the more RPO you have and the more you are expected to work to earn it, the more incentive and time you’ve created for a low end AI disruptor to do that work for a customer. The point is that any valuation model that incorporates any discounted value of RPO may be on the high side and overestimating. The safest assumption would be to value a software company by valuing their RPO at 0. This may seem ultra conservative but may be the right margin of safety unless you have a precise sense of the AI capabilities of the foundation model companies and their ecosystem.