Lazy evaluation is a programming technique that delays the computation of values until they are actually needed. This approach contrasts with eager evaluation, where expressions are evaluated immediately. Lazy evaluation can enhance performance by minimizing resource usage, as it avoids unnecessary calculations and memory consumption. It is particularly useful for working with large datasets or infinite data structures, allowing programs to generate values on-the-fly. Languages like Haskell utilize lazy evaluation by default, while others, such as Python, implement it through constructs like generators. Understanding lazy evaluation can lead to more efficient and responsive code.