Angle

Percent

Core

Bitcoin

when

<p>You sent bitcoin out of your wallet, the wallet shows the transaction, and for hours it has displayed the same word: unconfirmed. The short answer first: in the vast majority of cases nothing is lost, and you have two tools to sort the matter out yourself. They are called Replace-by-Fee and Child Pays For Parent, and which one you need depends on whether you are allowed to replace the transaction yourself or have to push it along from behind.</p>
<p>The reason this issue is hitting so many users right now lies in a technical change that entered the Bitcoin software in October 2025, and whose consequences have only become fully visible across the network this year. Since then, transactions paying fees below one satoshi per vByte can be relayed at all. Those transactions form today's backlog, the one many users are stuck in.</p>
<h2 id="bitcoin-transaction-stuck-what-is-really-happening-in-the-mempool">Bitcoin Transaction Stuck: What Is Really Happening in the Mempool</h2>
<p>The <a href="https://cryptoticker.io/en/prediction/bitcoin-price-prediction/" rel="noopener noreferrer nofollow">Bitcoin</a> price stood at roughly $76,700, or about 66,100 euros, on September 13, 2026 (CoinGecko, retrieved 09:52 UTC). The price is a side issue for this topic, but the market situation is not: when a lot of investors move their holdings off trading venues into self-custody after a pullback, the number of transfers on the network rises, and competition for space in the blocks gets tougher.</p>
<p>We measured instead of guessing. At the time of the survey, the Bitcoin network's waiting area held 75,903 unconfirmed transactions totalling 38.6 million vBytes. A block holds roughly one million vBytes. The backlog therefore amounted to about 38 blocks, or some six to seven hours of work for the miners if nothing new arrived. Something new arrives constantly.</p>
<h3 id="why-a-full-mempool-still-does-not-mean-expensive-fees">Why a Full Mempool Still Does Not Mean Expensive Fees</h3>
<p>The usual rule of thumb says: full mempool equals high fees. That rule no longer holds at the moment. The rate recommended by the common fee estimators for a prompt confirmation was a single satoshi per vByte, the lowest value those estimators ever output. A full mempool and a minimum fee are no longer mutually exclusive today, and anyone unaware of that draws the wrong conclusions.</p>
<h2 id="mempool-feerate-and-sat-vbyte-the-three-terms-you-need">Mempool, Feerate and sat/vByte: The Three Terms You Need</h2>
<p>Before this gets practical, three definitions the rest of it rests on.</p>
<p><strong>Mempool:</strong> The mempool is the waiting area in which every Bitcoin node holds the transactions that have been broadcast but not yet included in a block. It is not a central location; it exists thousands of times over in parallel on all the nodes of the network, which is why different providers display slightly different figures.</p>
<p><strong>Feerate:</strong> The feerate is the price you pay per unit of data in your transaction, not per amount transferred. A transfer of 20 euros and one of 20,000 euros cost exactly the same if their data size is identical.</p>
<p><strong>sat/vByte:</strong> A satoshi is one hundred-millionth of a bitcoin. The vByte is the measure of a transaction's size. The figure sat/vByte therefore says: this many hundred-millionths of a bitcoin per unit of size. Miners sort the waiting transactions by that number and take from the top, because their space in the block is limited. Bid too little and you slide backwards, a little further with every new transaction that bids more.</p>
<h3 id="what-makes-a-typical-transaction-large">What Makes a Typical Transaction Large</h3>
<p>The data size depends on how many earlier inputs your payment is assembled from. Anyone who has received many small amounts over the years drags all of those fragments along with every outgoing payment and pays accordingly more. How to bundle such holdings cheaply during a quiet phase is described in our piece on <a href="https://cryptoticker.io/en/bitcoin-network-fee-utxo-consolidation/" rel="noopener noreferrer nofollow">consolidating UTXOs while the network fee is low</a>. That no longer helps with an acute stuck payment, but it helps a great deal in preparing the next one.</p>
<h2 id="our-measurement-99-7-percent-of-the-backlog-sits-below-one-satoshi-per-vbyte">Our Measurement: 99.7 Percent of the Backlog Sits Below One Satoshi per vByte</h2>
<p>cryptoticker.io carried out this analysis itself on September 13, 2026. Method: retrieval of the public interfaces of mempool.space on September 13, 2026 between 09:50 and 09:55 UTC; we evaluated the fee distribution of the entire waiting area as well as the fifteen most recently found blocks, from height 966,789 to 966,803. Objects examined: 75,903 waiting transactions spread across 167 fee brackets, and 15 blocks.</p>
<p>The result is clearer than we had expected:</p>
<ul>
 <li>Of the 38.6 million vBytes in the waiting area, 38.4 million vBytes sat below one satoshi per vByte. That is 99.7 percent of the entire backlog.</li>
 <li>Above the mark of one satoshi per vByte, only 109,893 vBytes were waiting, or 0.3 percent. That is a good tenth of a single block.</li>
 <li>The highest fee bracket in the entire waiting area was 1.31 satoshis per vByte. Above that, the mempool was practically empty.</li>
 <li>All 15 blocks examined were more than 99 percent full, with an average of 5,281 transactions per block.</li>
 <li>In 11 of those 15 blocks, transactions below one satoshi per vByte were confirmed. The lowest floor observed was 0.286 satoshis per vByte, in block 966,802.</li>
</ul>
<p>For you as a sender, that leads to a statement which contradicts appearances: an apparently overcrowded waiting area barely stands in your way as long as you bid above one satoshi per vByte. Only those 0.3 percent are ahead of you. Bid less and you place yourself behind almost forty blocks' worth of competition that keeps renewing itself.</p>
<p>What we could not verify: the measurement shows a point in time, not a trend. It rests on a single provider's view of the network, and other nodes may hold differing inventories, because every node is free to set its own acceptance rules. We were also unable to establish what share of the backlog originates from individual large senders.</p>
<figure><img src="https://cryptoticker-strapi-media.s3.eu-central-1.amazonaws.com/bitcoin_transaktion_haengt_rbf_cpfp_2_94b17e6fca.png" alt="Metal funnel packed with jammed matte metal discs, while a single gold coin bearing the Bitcoin symbol slips freely through at the bottom" width="1200" height="675">
 <figcaption>
  Almost the entire backlog consists of transactions below one satoshi per vByte; anyone bidding above that passes them by.
 </figcaption>
</figure>
<h2 id="why-sub-satoshi-fees-have-been-possible-at-all-since-bitcoin-core-30-0">Why Sub-Satoshi Fees Have Been Possible at All Since Bitcoin Core 30.0</h2>
<p>Until recently there was an invisible floor. The default setting of the most widely used node software simply rejected transactions below one satoshi per vByte and did not relay them. Version 30.0, released on October 10, 2025, changed that. The release notes state verbatim that the default values for the minimum relay fee and the incremental fee have been changed to 0.1 satoshis per vByte; the minimum block fee has stood at 0.001 satoshis per vByte since then. You can read it in the <a href="https://bitcoincore.org/en/releases/30.0/" rel="noopener noreferrer nofollow">official release notes for Bitcoin Core 30.0</a>.</p>
<p>The developers placed a warning right next to it: as long as these lower values are not adopted network-wide, neither relay nor confirmation is guaranteed for transactions paying such low fees. That sentence describes precisely the problem now landing on many users' screens. The wallet is allowed to offer a very low fee, the network accepts it, and then nothing happens for a long time.</p>
<h3 id="what-that-means-for-your-wallet-setting">What That Means for Your Wallet Setting</h3>
<p>If your wallet has a fee slider and proposes something in the range of 0.2 to 0.5 satoshis per vByte as its lowest step, that is no malfunction. It is the new reality. That step is meant for transfers where a confirmation within days is good enough. For anything that should still arrive today, it is currently the wrong choice.</p>
<h2 id="is-your-transaction-really-stuck-how-to-check-in-two-minutes">Is Your Transaction Really Stuck? How to Check in Two Minutes</h2>
<p>Before you repair anything, establish whether anything is broken at all. You need the transaction ID for that, a long string your wallet displays under details and which can usually be copied with a single tap.</p>
<p>Enter that ID into a public block explorer. Three pieces of information matter to you there:</p>
<ol>
 <li><strong>The feerate of your transaction in sat/vByte.</strong> If it is above one satoshi per vByte, our measurement says you have barely any competition ahead of you and should simply wait.</li>
 <li><strong>The currently recommended feerate.</strong> If your rate is well below it, that is the reason for the delay.</li>
 <li><strong>The replaceability flag.</strong> Explorers show whether a transaction is marked as replaceable. Which of the two tools you use in the next step depends on it.</li>
</ol>
<p>A transaction does not simply vanish, by the way. If it goes unconfirmed for days, it eventually drops out of the nodes' waiting areas, and the bitcoin show up as available in your wallet again. Nothing is lost in the process, because an unconfirmed transaction never truly left your balance.</p>
<h2 id="replace-by-fee-rbf-replacing-the-transaction-with-a-more-expensive-one">Replace-by-Fee (RBF): Replacing the Transaction With a More Expensive One</h2>
<p><strong>Replace-by-Fee</strong> is a node rule under which an unconfirmed transaction in the waiting area may be replaced by another one that spends at least one of its inputs and pays a higher fee. This is not about a second payment. You submit the same payment once more with a better offer, and the old version is discarded.</p>
<p>Two points are worth knowing before you press the button. First, under the widely used rule from BIP 125, the new version must pay both a higher feerate and a higher absolute fee, as the <a href="https://bitcoinops.org/en/topics/replace-by-fee/" rel="noopener noreferrer nofollow">Optech compendium on Replace-by-Fee</a> records. Making the transaction smaller is therefore not enough. Second, replacement has not been tied to a special flag since 2024: change set 30493 made general replaceability the default in August 2024, and in November 2024 the corresponding switch disappeared entirely.</p>
<h3 id="how-to-proceed-in-practice">How to Proceed in Practice</h3>
<p>In most self-custody wallets you will find an entry such as Increase Fee, Accelerate or Bump Fee on an unconfirmed transaction. The wallet builds the replacement version itself and proposes a new rate. Following our measurement, set it above one satoshi per vByte; in the current environment a normal transfer needs no more than two satoshis per vByte. The additional fee is usually deducted from the change, and the recipient's amount stays the same.</p>
<p>A word on security: for this procedure your wallet has to sign the transaction again. On a hardware wallet that means confirming on the device once more. Check the recipient address just as carefully as you did the first time. Which devices handle this process cleanly and which leave you in the dark is shown by our <a href="https://cryptoticker.io/en/comparison/hardware-wallets/" rel="noopener noreferrer nofollow">comparison of crypto hardware wallets</a>. If this is the first time in a while that you are handling your recovery words, it is a good moment to check that they are still complete and legible.</p>
<h2 id="child-pays-for-parent-cpfp-when-the-second-transaction-drags-the-first-along">Child Pays For Parent (CPFP): When the Second Transaction Drags the First Along</h2>
<p><strong>Child Pays For Parent</strong> is a procedure in which you spend an output of the stuck transaction onward in a new transaction carrying a high fee, so that miners take both into a block together. The trick lies in the ordering rule of the blockchain: a transaction can only be confirmed if the transaction before it is in the block as well. So whoever wants the child has to take the parent along.</p>
<p>Miners therefore calculate with the combined fee rate of parent and child. If that average is attractive, both move into the block together. This sounds more cumbersome than RBF, yet it has one decisive advantage: you do not need to be allowed to touch the original transaction.</p>
<h3 id="when-cpfp-is-the-tool-of-choice">When CPFP Is the Tool of Choice</h3>
<p>Two cases come up often in practice. The first: you are the recipient. Someone sent you bitcoin with too low a fee, and you are waiting for it. You may not replace other people's transactions, because that would require the sender's keys. You can, however, spend the output addressed to you onward and push the payment along that way. The second case: your wallet does not support raising the fee but does offer to spend an unconfirmed input.</p>
<p>CPFP has its limits too. Nodes cap how many connected unconfirmed transactions they keep in their memory; a long chain of parents and children eventually runs into those barriers. And if your child pays a high fee while the parent is very large, the child has to bring correspondingly more to lift the combined average.</p>
<figure><img src="https://cryptoticker-strapi-media.s3.eu-central-1.amazonaws.com/bitcoin_transaktion_haengt_rbf_cpfp_3_045a2362b9.png" alt="A small metal coin pulls a larger gold coin bearing the Bitcoin symbol out of a breaking sheet of ice on a taut steel chain" width="1200" height="675">
 <figcaption>
  Child Pays For Parent in a picture: the new transaction pays enough to drag the stuck one along with it.
 </figcaption>
</figure>
<h2 id="exchange-withdrawal-stuck-why-rbf-and-cpfp-will-not-help-you-there">Exchange Withdrawal Stuck: Why RBF and CPFP Will Not Help You There</h2>
<p>If you have triggered a withdrawal at a trading venue and it is stuck, the situation looks different. The transaction belongs to the provider, not to you. The keys sit there, so only the provider can replace it. CPFP is out as well, as long as the bitcoin have not yet reached you and you cannot spend the output.</p>
<p>What remains is knowing the mechanics. Many trading venues bundle withdrawals into batch transactions and set their fee at their own discretion. The fee deducted from you at withdrawal often has little to do with the network charge actually paid; how far apart those two figures can be is something we looked at in our analysis of <a href="https://cryptoticker.io/en/crypto-withdrawal-fees-network-fee-compared/" rel="noopener noreferrer nofollow">withdrawal fees and the real network fee</a>. If a withdrawal makes no progress for hours, support is the right address, and the transaction ID belongs in the first message. Which providers handle their withdrawals promptly and transparently is one of the criteria in our <a href="https://cryptoticker.io/en/comparison/exchange-comparison/" rel="noopener noreferrer nofollow">comparison of the best crypto exchanges</a>.</p>
<h2 id="which-fee-you-should-really-set-today">Which Fee You Should Really Set Today</h2>
<p>A simple orientation can be derived from the measurement, although it applies only to the situation measured and you should re-check it before every larger transfer.</p>
<p>For a payment that should arrive promptly, a value just above one satoshi per vByte is currently enough. That places you ahead of 99.7 percent of the waiting volume. Two satoshis per vByte is generously judged and costs only a few cents on a simple transfer of around 140 vBytes in size.</p>
<p>For a payment where days are good enough for you, you may use the new low steps. But then expect it to genuinely take days, and choose a wallet that lets you raise the fee later. Without that option you sit the waiting time out.</p>
<h3 id="the-mistake-that-costs-the-most">The Mistake That Costs the Most</h3>
<p>The most expensive mistake is the panic that follows a low fee, rather than the low fee itself. Anyone who sees a stuck transaction and promptly sends a second payment to the same address risks both being confirmed in the end, leaving the recipient with double the amount. Check first, then raise the fee or push the payment along, and under no circumstances send blindly again.</p>
<h2 id="what-rbf-and-cpfp-cannot-do-the-limits-of-both-tools">What RBF and CPFP Cannot Do: The Limits of Both Tools</h2>
<p>Both procedures accelerate a transaction that is already on the network. They do not reverse it. You can indeed replace a payment with another via RBF and in theory change the recipient too, as long as nothing is confirmed; once a confirmation exists, the process is final. No technology brings back a transfer that was sent to the wrong address and confirmed.</p>
<p>They are equally useless for a transaction that your wallet displays but that never reached the network. If the block explorer cannot find the ID at all, it was not relayed. The right step is then to reconnect the wallet and repeat the send, instead of fiddling with fees.</p>
<h3 id="the-tax-angle-in-one-sentence">The Tax Angle in One Sentence</h3>
<p>A transfer between two of your own wallets is not a sale and triggers no taxable event in Germany. The network fee paid is not a deductible item in this case either. The details are in our article on whether the network fee counts for tax purposes when sending between wallets.</p>
<h2 id="how-to-speed-up-a-bitcoin-transaction-what-to-take-away">How to Speed Up a Bitcoin Transaction: What to Take Away</h2>
<ol>
 <li><strong>Check your feerate first, before you do anything at all.</strong> If it is above one satoshi per vByte, our measurement puts only 0.3 percent of the backlog ahead of you, and waiting is the right decision. If it is below, the delay is explained. Anyone who wants to run such checks regularly on their own device will find the differences between the models in our <a href="https://cryptoticker.io/en/comparison/hardware-wallets/" rel="noopener noreferrer nofollow">comparison of crypto hardware wallets</a>.</li>
 <li><strong>Use RBF when the transaction is yours, and CPFP when it comes from someone else.</strong> When raising the fee, the new version has to pay both a higher rate and a higher total fee. When pushing a payment along, the combined rate of parent and child is what counts. Whether your wallet offers fee bumping at all is a selection criterion like any other, and it likewise appears in the <a href="https://cryptoticker.io/en/comparison/hardware-wallets/" rel="noopener noreferrer nofollow">comparison of crypto hardware wallets</a>.</li>
 <li><strong>For a stuck withdrawal from a trading venue, turn to the provider.</strong> The keys are there, so the leverage is there too. How quickly and how transparently the individual houses process withdrawals can be looked up in the <a href="https://cryptoticker.io/en/comparison/exchange-comparison/" rel="noopener noreferrer nofollow">comparison of the best crypto exchanges</a> before you open your next account.</li>
</ol>
<p><em>(As of September 13, 2026. This article is not investment advice. Prices and fee structures change; check the terms with the provider before you buy.)</em></p>

Bitcoin Transaction Stuck: How to Free It Again With RBF and CPFP

Table of Contents

CryptoTicker ENG

Other articles published on Sep 13, 2026

Join our free newsletter for daily crypto updates!