The noisy-channel coding theorem is one of the two great results in Claude Shannon’s 1948 paper A Mathematical Theory of Communication, and it overturned the prevailing intuition about communication. Engineers had assumed that sending information over a noisy line meant accepting a trade-off: push for higher speed and you must tolerate more errors, or slow down to keep errors low. Shannon proved this is false up to a precise limit.
He showed that every channel has a fixed number called its capacity, equal to the maximum mutual information achievable between its input and output. The theorem states that for any transmission rate below the capacity, there exist encoding and decoding schemes that make the probability of error arbitrarily small, no matter how noisy the channel is. Above the capacity, reliable communication is impossible.
The surprising part is that you can drive errors toward zero without slowing below capacity, by spreading information cleverly across long blocks of symbols using error-correcting codes. Shannon proved such codes must exist, although he did not say how to build them, which launched decades of work in coding theory.
This theorem is the reason modern digital communication is possible at all. The deep-space links that return images from distant probes, mobile phone networks, and the storage of data on disks all rely on the guarantee that, with the right coding, noisy physical media can carry information essentially error-free right up to a calculable limit.