Skip to content
AI supply chain term

Model distillation

Model distillation is a training technique in which a smaller student model learns to reproduce the behavior of a larger teacher model.

What it means

In distillation, a large, capable teacher model generates outputs, and often soft probabilities, that a smaller student model is trained to imitate. The student ends up capturing much of the teacher's skill at a fraction of the size, so it runs faster and cheaper at inference time. Distillation is part of the model-building process that follows pretraining, and it is one of the main ways labs convert an expensive frontier model into practical, deployable versions. As a constraint and a lever, it lowers serving cost and spreads capability widely, which can erode the moat of the original model, but it usually depends on access to a strong teacher and to enough compute to run the student training.

Why it matters to investors

Distillation turns costly frontier capability into cheaper, smaller models, lowering inference cost and pressuring the pricing power of whoever trained the original. Labs such as DeepSeek, Moonshot AI and data providers like Scale AI sit close to this dynamic.

See Model distillation in the AI value chainIts live model score, why it matters, and every company exposed to it.

Companies on this part of the chain

DeepSeekMoonshot AIZhipu AI (Z.ai)Scale AI

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

PretrainingPretraining is the first and most compute-intensive stage of building an AI model, teaching it general skills from huge amounts of data.Fine-tuningFine-tuning is the post-training stage that specializes a pretrained model on a smaller, targeted dataset to shape its behavior.QuantizationQuantization is a technique that shrinks a model by storing and computing its numbers at lower numerical precision, such as FP8 or INT8.InferenceRunning a trained model to answer requests — the recurring cost of AI in production.

See Model distillation 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.

Explore the features
All glossary terms

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.