Skip to content

← All Q&A

Why doesn't training a Large Language Model create a lasting competitive advantage?

LLM MoatsInvestmentAI Product

Drawn from Lutz Finger's Forbes column, LinkedIn writing, and Cornell teaching. Sources are cited inline so you can read the originals.

Three reasons LLM training costs fail to protect market position

Training costs for LLMs don’t create a business moat for three reasons. First, training costs are decreasing as data growth plateaus and software improves, so competitors can build similar models with lower R&D costs. Second, having a trained LLM doesn’t inherently create economic value. It’s just a new machine learning tool for sequence-to-sequence predictions. Third, economic value comes only from the application, not the algorithm itself. Algorithms become valuable when integrated into products that influence actionable outcomes. The algorithm alone isn’t sufficient for value creation.

OpenAI Isn’t Going Bankrupt, But It Has A Business Model Problem · Forbes


Have a follow-up? hello@lutzfinger.com. Or pick another question: all Q&A →