Heat limits on communication in computers

by Clarence Oxford
Los Angeles CA (SPX) Jan 01, 2026

Every task performed on a computer, from numerical calculations to video playback, depends on internal components exchanging information, and researchers are now quantifying the energy cost of that communication. Former SFI Graduate Fellow Abhishek Yadav, a Ph.D. scholar at the University of New Mexico, notes that communication is central to computation, yet the energy budget that devices devote to it has remained poorly understood.

Over the last decade, SFI Professor David Wolpert has led work on the thermodynamic costs of computation, focusing on the physical limits that govern how much energy information processing must consume. He argues that establishing thermodynamic bounds on the cost of communication is a critical but underexplored problem, relevant not only for computers but for communication systems that support much of modern society.

In a new study in Physical Review Research, co-authored by Yadav and Wolpert, the team analyzes the unavoidable heat released when information passes through a system, overturning earlier assumptions that communication could, in principle, be energetically free. The researchers combine ideas from computer science, communication theory, and stochastic thermodynamics, which studies real-world nonequilibrium systems such as smartphones and laptops.

Using a logical abstraction of generic communication channels, they derive the minimum heat that must be dissipated to send a single unit of information. The framework applies to artificial channels such as optical fibers and to biological pathways such as neurons transmitting signals in the brain, all of which operate in the presence of noise that can disrupt messages.

Their results show that the minimum heat dissipation is at least as large as the useful information, or mutual information, that successfully passes through the noisy channel. This link ties the physical energy cost of communication directly to how much information survives interference and reaches its destination.

The researchers then analyze a separate abstraction of how contemporary computers carry out computations to determine the minimum thermodynamic costs of encoding and decoding messages. These steps protect messages against channel noise, and the team finds that more accurate data transmission, achieved through improved encoding and decoding algorithms, necessarily increases heat dissipation inside the system.

Quantifying these lower bounds on energy costs for communication could guide the design of more energy-efficient systems. Yadav points to the von Neumann architecture used in most current computers, where communication between the CPU and memory contributes significantly to the overall energy budget, as a prime target for rethinking future designs.

According to the authors, the same physical constraints apply across all communication channels, suggesting a route to better understanding complex, energy-intensive systems that depend on signaling, from neural circuits to artificial logic hardware. Although the human brain consumes about 20 percent of the body's calories, it operates far more efficiently than artificial computers, raising questions about how natural computational systems manage the thermodynamic costs of communication.

Research Report:Minimal thermodynamic cost of communication