The Sat-Denominated Grid

A different unit, a different story

The U.S. Energy Information Administration reports that the average residential electricity price has climbed from 12.89 cents per kilowatt-hour in 2017 to 18.05 cents in early 2026. That is a 40 percent increase in nine years. In dollar terms, the story is straightforward and familiar: energy is getting more expensive. Fuel costs rise. Grid infrastructure ages. Demand for compute expands. Ratepayers absorb the bill.

But what happens when you denominate the same kilowatt-hour in a different unit of account — one with a fixed supply and a transparent issuance schedule? Price the grid in satoshis (sats), the smallest unit of Bitcoin at one hundred-millionth of a coin, and the narrative inverts. In 2017, one kilowatt-hour of residential electricity cost roughly 3,223 sats. In March 2026, that same kilowatt-hour costs about 254 sats. That is a 92 percent decline.

The dollar price went up. The sat price went down. Both statements are factually correct, measured from the same underlying commodity: a unit of electrical energy delivered to an American household. The divergence between these two curves is not a curiosity. It is a window into the relationship between energy, computation, and what it means for money to hold its value.

The data

The calculation is simple. Take the EIA's average residential electricity price in cents per kilowatt-hour, convert to dollars, divide by the annual average Bitcoin price, and multiply by one hundred million to express the result in sats. The chart below shows the result for each year from 2017 through early 2026.

Two features stand out immediately. First, the sat-denominated price is dominated by Bitcoin's price cycle, not by changes in the underlying cost of generation and delivery. The 2022 spike to 773 sats per kilowatt-hour does not reflect a grid emergency; it reflects a bear market in Bitcoin following the collapse of leveraged lending platforms. Second, even with that cyclical volatility, the trend is unmistakably downward. The floor of each cycle is lower than the last.

Energy as the base layer

Why does this framing matter? Because energy is not merely one commodity among many. It is the *meta-commodity* — the input required to produce every other good and service in an industrial economy. Steel, aluminum, silicon, fertilizer, water treatment, transportation, data processing: every link in the value chain is ultimately a transformation of energy. The price of energy, denominated in your unit of account, is therefore a measure of how much productive capacity that unit of account commands.

When you hold dollars and energy gets more expensive in dollar terms, your money buys less of the thing that makes everything else possible. When you hold bitcoin and energy gets cheaper in sat terms, your money buys more. This is not an abstraction. It has direct implications for anyone making long-horizon capital allocation decisions — whether that is a utility planning a twenty-year generation portfolio, a manufacturer evaluating reshoring economics, or an individual thinking about retirement purchasing power.

Compute is just energy with instructions

The connection between energy and computation is not metaphorical. Every unit of compute — whether it is a hash on the Bitcoin network, an inference pass through a neural network, or a simulation running on a cloud cluster — is a direct conversion of kilowatt-hours into useful work. The cost of compute is, at its foundation, the cost of energy plus the amortized cost of the silicon and infrastructure that channels it.

This is why the sat-denominated price of electricity tells a story about compute as well. If one kilowatt-hour buys a fixed amount of computational work (holding hardware efficiency constant), and that kilowatt-hour is getting cheaper in sat terms, then computation is getting cheaper in sat terms. For a world in which the demand for compute — mining, inference, simulation, rendering — is growing exponentially, the unit of account in which you measure energy inputs matters enormously.

Consider the current trajectory. The EIA projects continued increases in electricity demand driven substantially by the expansion of compute-intensive workloads. Commercial and industrial consumption is rising, not falling. The grid is under pressure. Yet in sat-denominated terms, even as the grid strains, the cost of the energy flowing through it continues to fall relative to the monetary network that is itself the single largest peaceful consumer of energy on earth.

Bitcoin's energy feedback loop

There is an elegant circularity here that is easy to miss. Bitcoin mining is the process of converting electricity into monetary security. Miners purchase kilowatt-hours, run SHA-256 computations, and in return receive newly issued bitcoin and transaction fees. The network's difficulty adjustment ensures that as more energy is directed at mining, the computational barrier to producing a block rises, which in turn increases the cost (in energy terms) of attacking the network.

This means Bitcoin's value is, in a very real sense, backed by the cumulative energy expenditure required to sustain it. When the dollar price of bitcoin rises and the sat-denominated price of energy falls, what the market is expressing is that it values the security, scarcity, and settlement assurance of the Bitcoin network *more than* the underlying energy used to produce it. The spread between the cost of energy input and the value of the monetary output is the thermodynamic premium the market places on sound money.

This feedback loop — energy secures the network, the network's value makes energy relatively cheaper for holders, cheaper energy enables more computation — is not available to any other monetary system. Fiat currencies degrade purchasing power over time precisely because their supply is elastic. Gold requires energy to mine but offers no computational utility in return. Bitcoin is the only monetary asset where the energy expenditure is not a cost to be minimized but a feature to be valued, because it directly produces the security that makes the asset worth holding.

The two curves and what they mean

Look at the chart one more time. The blue line (dollar-denominated electricity prices) slopes gently upward. The orange bars (sat-denominated electricity prices) collapse downward. These two curves will continue to diverge for as long as two conditions hold: first, that the dollar continues to inflate its supply faster than productivity improvements can offset; and second, that Bitcoin's monetary premium continues to grow as adoption, infrastructure, and institutional allocation expand.

Neither condition is guaranteed to persist indefinitely. But the structural pressures are clear. The federal government runs persistent deficits financed in part by monetary expansion. Grid costs are rising due to aging infrastructure, extreme weather hardening, and the unprecedented growth of compute demand. Meanwhile, Bitcoin's supply issuance was cut in half again at the April 2024 halving and will halve again around 2028. The asymmetry between an inflating cost denominator and a deflating monetary denominator is not accidental. It is arithmetic.

Implications for builders and operators

For those who build and operate energy infrastructure — power plants, transmission systems, compute facilities — this dual-denominated view of electricity costs is more than a thought experiment. It reframes the economics of long-duration capital investment.

A compute facility operator who finances a build in dollars and earns revenue in dollars faces steadily rising energy input costs. The same operator who holds a portion of treasury reserves in bitcoin is, in effect, hedging against the very input cost that dominates their operating expense structure. The kilowatt-hours they must purchase become relatively cheaper against the asset they hold in reserve. This is not speculative alchemy. It is balance sheet management informed by the directional relationship between energy costs and monetary debasement.

The same logic applies to grid operators planning decade-scale capital programs, to industrial manufacturers evaluating the total cost of electrification, and to sovereign wealth funds assessing the long-term purchasing power of their energy-importing economies. In each case, the question is the same: in what unit of account are you measuring the most important input to your productive capacity?

Conclusion: the grid priced in truth

The chart of U.S. electricity denominated in sats per kilowatt-hour is a small thing — ten data points, two axes, a simple conversion. But it encodes a large idea. It shows that the rising cost of energy in dollar terms is not primarily a story about energy becoming scarcer or harder to produce. It is a story about the dollar becoming a worse measuring stick. Energy, in absolute terms, is as abundant and as technologically accessible as it has ever been. What has changed is the denominator.

Bitcoin offers an alternative denominator — one anchored to thermodynamic reality through proof-of-work, capped at 21 million units, and secured by the very energy it helps to price. When you denominate electricity in sats, you are not performing an arbitrary unit conversion. You are measuring the cost of the physical world in a monetary unit that is itself a product of the physical world. Energy in, security out, purchasing power preserved.

The grid is getting more expensive. Or the dollar is getting weaker. The chart tells both stories at once. Which one you see depends on what you choose to count in.

Data sources: U.S. Energy Information Administration, Average Retail Price of Electricity (Residential). Bitcoin annual average prices derived from CoinGecko and Messari historical data. 2026 figures use current spot values as of mid-March 2026: BTC ~$71,000, residential electricity 18.05¢/kWh (EIA/BLS).