What it means
Fine-tuning is the post-training stage where a pretrained base model is trained further on a smaller, targeted dataset to specialize its behavior, teaching it to follow instructions, adopt a tone, master a domain, or align with human preferences. Techniques range from full supervised fine-tuning (SFT) to reinforcement learning from human feedback and parameter-efficient methods that update only part of the model. It sits just after pretraining in the AI workload stack: pretraining supplies broad capability, and fine-tuning shapes it into a usable product. Because it uses far less compute and data than pretraining, fine-tuning is where many companies and open-model builders differentiate cheaply. It is the lever for turning a general model into a specialized one, and its quality depends heavily on curated data and human feedback rather than raw scale.
Why it matters to investors
Fine-tuning lets teams specialize models without the cost of pretraining, so it lowers the barrier to competitive products and shifts value toward data quality and human feedback. The frontier labs and open-model builders that ship strong post-trained models set the pace, and data-labeling and platform providers like Scale AI are exposed to that spend.
Companies on this part of the chain
Named to show where the term sits in the AI supply chain — research, not advice, and never a recommendation to buy or sell.
Related terms
See Fine-tuning in the live AI chain.
THE ENTITY maps every constraint onto one live model — which part is tight now, who owns it, and who gets squeezed when it moves. Plain-English reads you can check.
THE ENTITY is an educational read on the AI supply chain — research, not investment advice. It explains how the chain works and who sits where, never price targets or buy/sell calls.