# AI Winters: Cycles of Hype and Funding Collapse

Throughout AI's history, enthusiasm has repeatedly outpaced capability, followed by funding contractions. This sub-article expands the recurring-theme discussion in [History](../history.md).

## What Is an AI Winter?
An **AI winter** is a period of reduced funding, interest, and expectations in AI research and industry, typically after a "spring" of hype.

## Major Winter Episodes
- **First winter (1970s–1980s)**: early overpromising and the LISP machine market crash (1987).
- **Second winter (1990s)**: expert-system decline and reduced corporate and government investment.
- **Pattern**: hype → underdelivery → disinvestment → later breakthrough.

## Causes
- Overstated claims ("AI will soon match human intelligence").
- Brittle systems that fail outside narrow settings.
- Economic downturns and shifting government priorities.
- Benchmarks and demos that do not generalize.

## Lessons for the Present
- Modern AI (e.g., large language models) may face similar cycles of expectation.
- Distinguish **capability** from **deployment**.
- Fund research on fundamentals, not just demos.

## Sources
- Crevier, D. (1993). *AI: The Tumultuous History of the Search for Artificial Intelligence.*
- Nilsson, N. (2009). *The Quest for Artificial Intelligence.*

*Back to [History](../history.md).*