skip to content

Where Did All the Fees Go? Analyzing Bitcoin’s Onchain Activity and UTXO Set

Analyzing Bitcoin's Onchain Activity and UTXO Set

Introduction 

With the emergence of digital asset treasury companies and bitcoin’s price breaking all-time highs, people have been quick to forget the core value proposition of the Bitcoin network: a decentralized, censorship-resistant monetary system designed to provide permissionless access to global value transfer.  

Since the late-2024 decline in non-monetary Runes and Ordinals activity, Bitcoin’s onchain usage has sharply diminished. We’re now seeing an increasing number of blocks that are “free” or nearly free, meaning the average fee paid is 1 satoshi (one hundred-millionth of a bitcoin) or less per virtual byte. (A virtual byte is the standard unit for measuring transaction size on Bitcoin.) This is a short-term benefit for users who want cheap, fast transfers. However, it adds strain to the economics of mining, which is already under pressure following the 2024 halving.  

This note analyzes the structure of Bitcoin’s fee market to assess what’s really happening onchain and examine what it means for the network’s economic health. It also analyzes OP_RETURN transactions and their evolving usage. This is particularly relevant given the debate surrounding Bitcoin Core’s upcoming v30 release. This version of the network's open-source software would allow much larger and multiple OP_RETURN outputs per transaction by default. The planned changes have sparked criticism from parts of the Bitcoin community concerned about spam.  

Key Takeaways 

  • Bitcoin fee pressure has collapsed. 

    • Median daily fee has dropped >80% since April 2024. 

    • As of August 2025, ~15% of daily blocks are “free blocks.” 

  • OP_RETURN activity surged and receded. 

    • During peak Runes adoption (Q2-Q3 2024), OP_RETURN txs often accounted for 40-60% of daily txs. 

    • As of August 2025, that share has declined to ~20%. 

  • Mempool activity is lacking.

    • The percent of not-full blocks has spiked to nearly 50% at times in the past few months.

    • A dormant mempool could pose long-term sustainability questions for miner revenue following the 2024 halving, which reduced block rewards to 3.125 BTC. 

  • Onchain activity may be getting displaced by alternatives. 

    • Spot BTC ETFs now hold ~1.3 million BTC, much of which doesn’t really move onchain. 

    • Trading and speculative activity are shifting to alternative L1s like Solana, especially for use cases such as memecoins and NFTs.

  • Over 1.5 million BTC are still held in legacy P2PK addresses. These are bare public key addresses and are considered immediately vulnerable to potential quantum computer attacks because the public key is exposed onchain at all times. 

  • Over 6 million BTC are still held in legacy P2PKH addresses. 

  • P2WPKH now holds the largest share of unspent BTC out of any address format. 

Methodology 

All onchain data presented in this note was sourced from Galaxy’s internal Bitcoin infrastructure, including our in-house full node. Unless otherwise stated, statistics reflect the state of the Bitcoin network as of Aug. 12. 

Fee Metrics: Calculated using block-level data, including average and median fees, the percentage of “free blocks” (blocks with an average fee of ≤ 1 sat/vbyte), and the share of blocks classified as non-full. For this analysis, we define a block as non-full if its total weight is less than 3,900,000 weight units (relative to the 4,000,000 maximum). 

OP_RETURN Analysis: Transactions were parsed to identify those containing the OP_RETURN opcode. Daily OP_RETURN share is expressed as a percentage of total transactions per day.  

Address Format Classification: Unspent outputs were categorized by script type (e.g., P2PKH, P2PK, P2SH, P2TR, P2WPKH). Totals represent balances as of Aug. 12.  

The Fee Market 

The Bitcoin fee market, where users compete to have their transactions included in the next block, has entered a period of stagnation. While virtually all Bitcoin transactions have a fee attached, users choose how much to pay, and those who pay more are generally confirmed faster. After months of congestion driven by the rise of fungible and non-fungible tokens on Bitcoin (Runes and Ordinals, respectively), the network has seen a dramatic collapse in fee pressure. Daily average transaction fees have dropped to levels not seen since early 2023. 

Median and Average Bitcoin Transaction Fee (USD)

While average and median fee metrics offer a useful summary of fee market trends, they don’t capture the full picture. The chart below shows a more nuanced view: daily transaction fee rates by percentile, measured in satoshis per virtual byte (sats/vB).  

Covering the period from January 2023 to present, the chart plots the 10th, 25th, 50th (median), 75th, and 90th percentiles of transaction fee rates. This view reveals how fee pressure has evolved across the mempool, not just at the midpoint, highlighting the sharp compression in fee variance since late 2024.  

Bitcoin Transaction Fee Percentiles (sats/vbyte)

This trend is best captured in the rise of what we are defining as “free blocks,” those where the average fee rate is 1 sat/vbyte or less. These blocks, which were virtually nonexistent in 2024, have become increasingly common. At the time of writing, free blocks make up a meaningful share of total daily block production, highlighting the collapse in competition for blockspace.  

Bitcoin Blocks: Free vs. Non-Free

Along with this is a high share of “not full” blocks. These blocks fail to reach the maximum weight limit (4,000,000 weight units) despite room to include additional transactions. In short, the mempool, Bitcoin’s waiting room for pending transactions, is frequently empty, and when it isn’t, it’s full of transactions that don’t need to pay high fees to get processed quickly.  

Bitcoin Blocks: Full vs. Non-Full

From a miner’s perspective, this is concerning. With the 2024 halving having cut per-block rewards to 3.125 BTC, transaction fees were expected to take on a more meaningful role in supporting miner revenue. Instead, they’ve dried up. The long-term economics of Bitcoin security rest on a future with a robust fee market, and right now the fee market is anything but robust.  

The rise of custodial and “paper Bitcoin” solutions, whether in the form of ETFs or other institutional derivatives, may be dampening onchain activity. At the same time, memecoin activity has users who go for that sort of thing increasingly migrating to faster, cheaper L1s like Solana. It’s simply not worth dealing with the poor UX of trading Runes when the Solana memecoin trading experience has become so clean.  

OP_RETURN 

The Bitcoin OP_RETURN operation code, or opcode, was introduced in 2014 to allow users to embed up to 80 bytes of arbitrary data into a transaction output. These outputs are provably unspendable and typically carry zero BTC value, meaning they do not add to the set of unspent transaction outputs (UTXOs). UTXOs are discrete chunks of bitcoin that a wallet can use in future transactions. They form the basis of Bitcoin’s accounting system, where each transaction consumes previous outputs and creates new ones. If any BTC is sent to an OP_RETURN output, it becomes permanently unspendable. But the tradeoff is a new use case: publishing messages or metadata permanently onchain.  

This script type has seen an explosion in usage in the past 18 months. The launch of Runes in April 2024 (block 840,000, during the last halving of miner block rewards) pushed OP_RETURN transactions to record levels. At their peak, these primarily non-monetary transactions dominated blockspace, as seen in the chart below.  

Bitcoin Daily Transaction Composition: January 2009-January 2025

Today, OP_RETURN usage has tapered off with the decline in Runes activity. Still, developers and institutions continue to use it to anchor data onchain. Here at Galaxy, we used OP_RETURN to broadcast a historic sale of 80,000 BTC for a client a few weeks ago.  

Bitcoin Daily Transaction Composition August 2024-August 2025

OP_RETURN is now being used for novel purposes. A group named after the storied Wall Street firm Salomon Brothers has begun using OP_RETURN to post legal notices to dormant Bitcoin wallets. These onchain messages cite the Doctrine of Abandonment and offer wallet owners 90 days to respond before “Salomon” (which owns the trademark but is unaffiliated with its namesake’s successor institution, Citigroup) asserts its client’s rights to reclaim funds. (It is unclear how the group would access the wallets.) 

While this may be an interesting use case for anchoring legal messages directly on Bitcoin, it has stirred debate in the context of Bitcoin Core 30.0. That upcoming release of the network’s open-source software would expand default limits on OP_RETURN data payloads, enabling larger and multiple data-bearing outputs per transaction. Critics argue that while OP_RETURN outputs do not bloat the UTXO set (because they are provably unspendable) they still contribute to overall blockspace consumption, potentially crowding out monetary transactions and raising concerns over long-term network sustainability. In response, Bitcoin Core developers defended the update, emphasizing that the choice of whether to relay or mine larger OP_RETURNs rests with individual node operators and miners. 

Bitcoin Held by Script Type 

While fee trends offer insight into Bitcoin’s short-term dynamics, the UTXO set itself provides a long-term view of how BTC is actually distributed across the network. By classifying unspent outputs by script type, or format, we can observe the adoption of address types and the implications for spendability, security, and quantum resistance.  

The P2PKH (Pay-to-PubKey-Hash) format, which rose to prominence after Bitcoin’s earliest days, still holds a significant share of BTC (>6 million coins). However, newer formats have gained traction in recent years. P2WPKH (native SegWit), introduced in 2017, now holds the largest share of unspent BTC, while P2TR (Taproot), launched in 2021, is growing steadily and supports advanced scripting use cases.  

The chart below shows balances by script type as of Aug. 11. This data also informs discussions about future security risks.  

Legacy address format P2PK (Pay-to-PubKey), used mainly in early coinbase transactions (with a small c, as in mining reward payouts, not the exchange), are inherently quantum-vulnerable, because they expose the full public key onchain even before being spent.  

In contrast, other formats do not reveal the public key until an output is spent. Once spent (and especially if reused) the public key is exposed, making those coins similarly vulnerable.  

Galaxy Research will soon do a full report on quantum’s implications for Bitcoin, which will dig into this concept deeper.  

Unspent Bitcoin by Address Type

Conclusion 

Bitcoin’s onchain activity has entered a lull, but the network’s infrastructure continues to evolve.  

In the near term, low fees benefit users looking to consolidate UTXOs or move value cheaply. But the long-term picture is murkier. A fee market in retreat raises real questions regarding network security.  

With block rewards already cut to 3.125 BTC, miner incentives are increasingly exposed to fluctuations in organic demand. If more BTC volume continues to migrate to ETFs, custodians, and fast alt-L1s, the core network risks becoming a settlement layer without sufficient settlement activity.  

As “paper Bitcoin” scales and fee revenue stalls, Bitcoin’s security model becomes more reliant on a kind of usage that’s no longer guaranteed. Fee volatility is nothing new, but Bitcoin does need real reasons to use the chain.  

You are leaving Galaxy.com

You are leaving the Galaxy website and being directed to an external third-party website that we think might be of interest to you. Third-party websites are not under the control of Galaxy, and Galaxy is not responsible for the accuracy or completeness of the contents or the proper operation of any linked site. Please note the security and privacy policies on third-party websites differ from Galaxy policies, please read third-party privacy and security policies closely. If you do not wish to continue to the third-party site, click “Cancel”. The inclusion of any linked website does not imply Galaxy’s endorsement or adoption of the statements therein and is only provided for your convenience.