The Coin You Can't Spend

You check your wallet and see 800 satoshis sitting there, technically yours, right next to the balance that actually matters. Then fees climb. Now that output costs 1,200 satoshis to move. You own something you cannot economically spend. That's dust: not a bug, not a scam, just arithmetic working against you in slow motion.

The question worth asking is why high-fee networks manufacture dust so much faster than low-fee ones. The answer sits in the UTXO model, and once you see the mechanism, you can't unsee it.

How UTXOs Actually Work (The Part That Matters)

Bitcoin doesn't track balances the way a bank does. There's no ledger entry reading "Alice has 0.05 BTC." Instead, the network tracks discrete chunks of unspent value called Unspent Transaction Outputs, UTXOs. When Alice spends, she consumes one or more UTXOs as inputs and creates new ones as outputs: one to the recipient, usually one back to herself as change.

Every UTXO has a size, measured in bytes when you want to spend it. A standard P2PKH input (the legacy format most older wallets use) costs roughly 148 bytes to spend. At 10 sat/byte, that's a 1,480-satoshi fee just to consume that one input. At 100 sat/byte, it's 14,800 satoshis.

The UTXO itself hasn't changed. The fee environment has.

Dust, then, isn't a fixed number. It's a threshold that moves with the fee market.

The Threshold That Moves With the Market

Bitcoin Core defines a dust limit dynamically, pegged to three times the minimum relay fee for spending a given output type. The practical dust threshold for a P2PKH output sits around 546 satoshis at the base relay fee rate. But that's the floor for relay, not for economic rationality.

The economically rational threshold is higher and shifts constantly. If the mempool is clearing at 80 sat/byte, any UTXO worth less than roughly 11,840 satoshis (80 times 148 bytes) is uneconomical to spend solo. You'd pay more in fees than you'd recover in value.

Here's a scenario I keep coming back to. Maria runs a busy exchange hot wallet during a period of congested blocks, regularly receiving tiny outputs from consolidation errors and promotional airdrops. Dozens of UTXOs land between 500 and 5,000 satoshis. At 3 sat/byte, most of those are spendable at around 444 satoshis cost per input. Fine. But when fees spike to 150 sat/byte, a 5,000-satoshi UTXO now costs 22,200 satoshis to spend alone. Not dust by the relay definition, but economically inert. Maria's wallet has accumulated what I'd call functional dust: outputs stranded above the technical floor but below the practical ceiling.

Her colleague James runs the same wallet software on Litecoin, where fees have historically stayed in fractions of a cent per byte. James's equivalent outputs remain spendable through almost any realistic fee environment. His dust problem is small and slow-growing. Maria's is large and fast.

Why Accumulation Accelerates on High-Fee Chains

This is where the feedback loop gets interesting.

On a high-fee network, users and services adopt two behaviors that both create dust as a side effect. First, they batch aggressively: one transaction sweeps many recipients, generating multiple change outputs, each potentially tiny. Second, they avoid consolidating small UTXOs because the fee to combine ten small inputs into one large output is prohibitive precisely when fees are high. The small UTXOs sit and wait for fees to fall. Sometimes they wait years.

Meanwhile, the same high fees push the economic dust threshold upward. An output that was spendable last month becomes functionally stranded this month. The pool of dust grows from both ends: new small outputs arrive from normal activity, and existing medium outputs get reclassified as uneconomical by a rising fee floor.

There's a third mechanism people tend to overlook. Dust attacks. An adversary (usually one probing for privacy information) sends tiny UTXOs to thousands of addresses. On a low-fee network, recipients might consolidate these outputs quickly, defeating the probe. On a high-fee network, those outputs are too small to spend economically, so they sit permanently in wallets, unspent, exactly as the attacker intended. High fees make dust attacks cheaper to sustain and harder to clean up.

The UTXO set itself grows as a result. Bitcoin's full UTXO set, which every fully validating node must keep in memory, has grown to tens of millions of entries. A meaningful fraction of those entries are outputs that will never be spent because no rational actor will pay to move them. They are permanent fixtures, like barnacles on a hull: invisible until you haul the thing out of the water, then suddenly everywhere.

SegWit, Taproot, and the Partial Escape Valve

Protocol upgrades help. They don't solve the structural problem.

SegWit moved signature data into a separate witness field, discounted at 75% of the base weight. A native SegWit (P2WPKH) input costs roughly 68 vbytes rather than 148 bytes. At 100 sat/vbyte, that's 6,800 satoshis to spend versus 14,800 for legacy. The dust threshold drops meaningfully, and more outputs remain spendable at any given fee level.

Taproot pushes further. A key-path P2TR spend is around 57.5 vbytes, trimming the cost another 15% relative to P2WPKH. For wallets that have migrated fully to Taproot, dust accumulates more slowly because the fee cost per input is lower.

These optimizations are linear improvements on a problem that scales with fee rates, and that ceiling is worth being clear-eyed about. If fees climb 10x, a 15% efficiency gain from Taproot barely registers. The dust threshold still rises dramatically. Efficiency helps; it doesn't immunize.

Payment channels (the Lightning Network being the primary example) offer a more fundamental escape. By moving small-value transfers off-chain entirely, they prevent the creation of dust-sized UTXOs in the first place. A 200-satoshi micropayment routed over Lightning never touches the base chain, never creates a UTXO, never risks becoming dust. That said, opening and closing Lightning channels requires on-chain transactions, which themselves carry fee risk. During high-fee periods, even channel management becomes expensive, and some users end up with channel outputs too small to close economically. Dust chases you even here, just one layer up.

What You Can Actually Do With Stranded Outputs

So you've found dust in your wallet. Now what?

If you're above roughly 85% of the economically rational threshold at current fees, consolidate during low-fee periods: combine many small UTXOs into one large output in a single transaction. Most serious Bitcoin wallets (Sparrow, Electrum, Bitcoin Core) let you manually select inputs and build consolidation transactions. The optimal window is when the mempool clears and fees drop to single-digit sat/vbyte, which does happen periodically.

For outputs below the relay dust limit (under 546 satoshis for legacy), options shrink fast. Some wallets simply won't let you spend them. A few will include them as inputs in larger transactions where the fee is covered by a bigger UTXO, effectively subsidizing the dust cleanup with value from elsewhere.

The coldest take here is also the most accurate one: a significant fraction of satoshis that exist on-chain will never move again. Not because they're lost in the cryptographic sense, but because the economic cost to move them permanently exceeds their value, and that gap widens every time fees rise. They're not lost coins. They're retired ones.

High-fee networks don't just charge more to transact. They slowly convert a portion of their supply into permanent ballast, outputs that exist in the ledger forever but contribute nothing to economic activity. Whether that's a bug or an emergent property of a system with a fixed block size and variable demand is, genuinely, a question worth sitting with.