Moore's law

In 1965, Gordon E. Moore published a paper about the state of computing that continues to shape how we think about the field. In it he made a remarkable observation, namely that the amount of computing power that engineers could squeeze onto a single chip appeared to double at regular intervals. At the time, that doubling happened every twelve months. Before long, it was more like once every two years.

Moore’s law, as this observation came to be known, is often taken to be an immutable rule of nature. It is nothing of the kind. It was an observation about human ingenuity that became a prediction that has turned into a self-fulfilling prophecy.

Recently, the runaway growth predicted by Moore’s law has started to slow down. Computer components are getting so small, and so energy-intensive, that we are running against physical limits on how tightly they can be packed. All sorts of solutions have been proposed: from abandoning the comforts of silicon to exploring alternative computing technologies like spintronics.

There are other, more pressing reasons to get Moore’s law back on track. With the global data industry set to consume 18% of the world’s energy by 2025, developing more efficient computers is a matter of urgency. Otherwise, we run the risk of drowning in our own data.