What emerging technologies promise to displace conventional silicon chips? Future computers could run on graphene, perhaps, or the hidden powers of quantum physics or brain-like synaptic networks. Research on all these options and more is under way as it becomes clear that enhancement of silicon-chip technology is hitting serious practical obstacles: in manufacturing, connectivity and heat generation.

No emerging technology is likely to be a get-out-jail-free card. Amazing performance in one area is often accompanied by serious limits in another. Computing based on carbon nanotubes or graphene, for example, presents formidable challenges in reliable fabrication.

On page 147, information technologist Igor Markov argues that we should focus on the fundamental limits in computing, and use those to evaluate future possibilities. This approach has a rich history. Working out the maximum efficiency of steam engines, nineteenth-century physicists discovered thermodynamics. Modern information science was born in 1948 when Claude Shannon at Bell Labs considered what an ideal communication channel would look like.

Computations have limits: they take up space, time and energy. In 2000, IT researcher Seth Lloyd calculated the computing power of the ultimate laptop, which, by miraculous engineering advances, could harness all its energy for information processing (S. Lloyd Nature 406, 1047–1054; 2000). This ultimate machine could perform 1051 operations per second, 40 orders of magnitude more than computers today. That represents 250 years of progress at current rates of improvement.

Markov’s message is not to be overly optimistic or pessimistic about further progress. We should focus on the boundaries and push to see where they yield.