What Realized Cap Can and Can’t Tell You

Market capitalization — price times circulating supply — has an obvious flaw: it assumes every coin could sell at the current price simultaneously, which is never true for any asset. Realized capitalization was developed as an alternative that tries to fix this, and it is worth understanding both its genuine cleverness and its real limits.

The idea

Instead of valuing every coin at today’s price, realized cap values each coin at the price it was last moved on-chain, then sums across the entire supply. A coin that has not moved since it was mined in 2011 is valued at its 2011 price; a coin that changed hands yesterday is valued near today’s price. Summed across all coins, this approximates the aggregate cost basis of the network — roughly, what the market as a whole “paid” for the coins it holds, rather than what it could theoretically get for all of them right now.

This makes realized cap useful for a specific question: is the network, in aggregate, sitting on unrealized profit or loss relative to where coins last changed hands? When market cap is well above realized cap, most holders are sitting on paper gains; when the two converge or invert, a large share of holders are near or below their cost basis. That is a genuinely different and often more informative signal than spot price alone.

Where the assumptions creep in

The basic realized-cap calculation is derived directly and transparently from public chain data — it requires no clustering or guesswork, just tracking when each UTXO last moved and at what price. The more elaborate metrics built on top of it are a different story. MVRV ratio, various “profit and loss” cohort breakdowns, and most HODL-wave-adjacent metrics typically layer in assumptions about which addresses belong to the same entity, in order to estimate things like exchange-held supply or long-term-holder behavior specifically.

Those entity-clustering heuristics — common-input-ownership heuristics, known-exchange-address tagging, change-address detection — are genuinely useful, and genuinely uncertain. Different data providers use different heuristics, rarely publish them in full, and can disagree with each other on the same underlying chain data as a result. A number quoted to two decimal places can create a false impression of precision that the underlying method does not support.

Why we don’t put it on the live board

The base realized-cap figure is defensible and we discuss it in our analysis. The more refined derivatives — the ones that make for the most dramatic headlines — depend on clustering assumptions we cannot verify independently and that vary by provider. Rather than display a number we could not defend if asked exactly how it was computed, The Bitcoin Signal omits this family of metrics entirely. This is a deliberate choice, not an oversight, and it is documented at our methodology page.

How to read it responsibly

When you encounter a realized-cap-derived metric, ask what entity-clustering assumption sits underneath it, whether the provider discloses their methodology, and whether the same figure is corroborated by an independent source. Treat convergence across sources as stronger evidence than any single provider’s number, however precisely stated.