Data markets behave exactly like physical engines, with a measurable temperature that determines how much work can be extracted from a single bit of information.
April 26, 2026
Original Paper
Information Thermodynamics of Data Markets
SSRN · 6642582
The Takeaway
Information is usually treated as a social or economic product governed by human preference and supply. This model applies the second law of thermodynamics to show that buying and selling data follows the same physics as heat transfer. Every bit of data has an informational temperature that sets its value in dollars per bit based on how much it can reduce uncertainty. Treating data as a physical resource allows for precise calculations of market efficiency that social science cannot provide. This perspective reveals that information markets have physical limits that can be measured just like a steam engine or a battery.
From the abstract
We apply the formalism of information thermodynamics to the economics of data markets. Starting from a constrained maximum-entropy principle on market-belief distributions, we derive an informational temperature τ (dollars per bit) as the Lagrange multiplier conjugate to the belief entropy and an informational free energy F I = U I-τS I whose non-equilibrium excess satisfies ∆F I = τ • D KL (P informed ∥ P public). Data extraction is modeled as a Szilard engine-the minimal realization of informa