Dogecoin (DOGE): a full reference on the coin and its mining

Dogecoin is a proof of work network launched on 6 December 2013. Its authors: Billy Markus, an IBM programmer, and Jackson Palmer, an Adobe employee. The coin was meant as a parody of the crypto frenzy of the time, yet the engineering behind it was quite serious: fast blocks, fees worth fractions of a cent and a very low barrier to entry for an ordinary user.

By the origin of its code Dogecoin is a fork of Luckycoin, and Luckycoin in turn is a fork of Litecoin. The unusual scheme with a random block reward, which ran during the first months of the network, came from Luckycoin. Dogecoin is not a direct fork of Litecoin, although the Scrypt algorithm did come from there along that chain.

For a hardware owner Dogecoin is interesting not on its own but paired with Litecoin. Since September 2014 both networks have been mined by the same work through AuxPoW, that is, merged mining. At the end of July 2026, at the current price ratio, DOGE accounts for roughly 86 percent of the total revenue of a Scrypt miner, and Litecoin for about 14. Planning farm income around the LTC price means being wrong by a wide margin.

This page is built as a reference work rather than a product card: protocol parameters, a breakdown of an emission schedule with no halving, an honest explanation of why the network has no hashrate of its own, a ranking of Scrypt ASIC models from our catalogue with break even thresholds, pools, wallets and a separate list of figures that the large directories still publish in outdated or plainly wrong form.

• 2013The network launches on 6 December. Early blocks pay a random reward from 0 to 1,000,000 DOGE, a scheme inherited from Luckycoin. Already on 25 December the Dogewallet service is breached and about 21 million DOGE are stolen. The community raises compensation for the victims within days.
• 2014January: the community raises about 30 thousand dollars to send the Jamaican bobsleigh team to the Sochi Olympics. March: sponsorship of Josh Wise NASCAR car for roughly 55 thousand dollars. These campaigns cement the coin reputation as a tool for tipping and fundraising.
• 2014At block 145,000 the network moves to DigiShield and drops random rewards in favour of a deterministic schedule. Until then difficulty was recalculated once every 240 blocks, which left the network defenceless against large amounts of hashrate arriving and leaving.
• 2014By mid year the Dogecoin hashrate sags relative to Litecoin and the risk of a 51 percent attack becomes real. In April Charlie Lee proposes merged mining as the fix.
• 2014The key date: on 11 September, at block 371,337, AuxPoW is activated. Litecoin miners start closing Dogecoin blocks along the way at no extra cost. Network security is restored, but from that moment on it is borrowed from the neighbouring chain.
• 2015The reward reaches its fixed value of 10,000 DOGE at block 600,000 and reward cuts stop for good. Circulating supply passes the 100 billion coin mark, showing that the emission has no limit at all. Both founders leave the project.
• 2019The Dogecoin Core 1.14 client branch is released, and the network still runs on it today. Elon Musk speaks about the coin publicly for the first time, starting years of social media mentions.
• 2021The year of maximum publicity. In January the coin is driven up by the retail trading wave, and on 4 May it sets an all time price high of about 0.73 dollars. On 8 May, during Elon Musk appearance on a television show, the price drops by roughly 30 percent live on air.
• 2021An infrastructure breakthrough: on 3 June the coin arrives on Coinbase, in August the Dogecoin Foundation is relaunched with Vitalik Buterin among its advisors. In December Tesla announces that it accepts DOGE for branded merchandise.
• 2023In May Doginals and the DRC-20 token standard appear. Daily transaction counts hit all time records, the mempool fills with inscription data and fees rise several fold for a while. The episode shows that with a 1 MB block the network hits its ceiling sooner than people assume.
• 2024Three client releases in a row: 1.14.7 in February, 1.14.8 in August and 1.14.9 on 1 December. Fee policy takes its current shape: a recommended rate of 0.01 DOGE per kilobyte and a minimum relay rate of 0.001 DOGE per kilobyte. Version 1.14.9 is still current in 2026.
• 2025The year of exchange traded funds. On 18 September DOJE from REX Shares and Osprey launches, the first US exchange traded product on a meme coin. On 20 November TXXD from 21Shares arrives with double leverage and a fee of 1.89 percent. In November Grayscale launches the spot GDOG on NYSE Arca, followed by BWOW from Bitwise.
• 2026On 23 February the network hashrate reaches an all time high of about 8.72 PH/s, after which it falls by more than two thirds by the end of July. On 11 March Elon Musk announces the launch of the X Money payment service in April, and the price reacts with a rally, yet there is no official confirmation that DOGE is accepted in that service as of the end of July 2026.

The mechanics of the network are simple and have not changed since 2015. Miners run through nonce values, computing the Scrypt function over the block header until the result falls below the current target. Whoever finds it takes the fixed reward of 10,000 DOGE plus the transaction fees in the block. Target block time is 60 seconds, and the actual figure at the end of July 2026 is about 1 minute 3 seconds.

The main difference from most proof of work networks is that the vast majority of blocks are found not by separate work, but along the way. A miner solves the Litecoin task, and because Dogecoin difficulty is lower, the very same hash found counts in parallel as a valid Dogecoin block. No extra electricity, no separate hardware and no second fleet of machines are needed for this.

Difficulty is recalculated after every block under the DigiShield scheme: the algorithm looks at the time of the previous block and adjusts the target with damping, so that a single outlier does not shake the network. That kind of response was vital for a coin with a one minute block and migrating hashrate. For a miner it means that difficulty here breathes constantly and has none of the familiar two week steps.

1. 1. TransactionThe user signs a transfer and broadcasts it to the network. The recommended fee is 0.01 DOGE per kilobyte, that is fractions of a cent.
2. 2. MempoolNodes verify the signature and the available funds, and the transaction waits in the queue. Under normal load that queue is nearly empty: the network runs at roughly one percent of its capacity.
3. 3. Block assemblyThe pool builds a Dogecoin block and a Litecoin block at the same time, embedding a reference to the DOGE block into the LTC job.
4. 4. HashingThe ASIC computes Scrypt over the header. The same work is checked against two difficulty targets at once, and no separate accounting is kept for DOGE.
5. 5. Block foundIf the hash clears the Dogecoin target, the block goes out to the network together with the Litecoin proof of work. The reward of 10,000 DOGE plus fees is credited to the pool.
6. 6. Difficulty retargetDigiShield immediately adjusts the target for the next block, using the actual time of the previous one.
7. 7. PayoutThe pool distributes DOGE and LTC among participants under its own scheme. Most operators pay both coins in a single credit.

Scrypt was chosen for Litecoin and then inherited by Dogecoin along the chain through Luckycoin, for the sake of resistance to specialised chips. The function requires filling a memory buffer with pseudorandom data and then reading from it in an unpredictable order. In theory that makes the advantage of an ASIC minimal, because the bottleneck becomes memory rather than arithmetic.

In practice the plan did not work. In the Scrypt implementation used by Litecoin and Dogecoin the memory parameter was left small, about 128 kilobytes per thread, and that fitted on a die. The first Scrypt ASIC devices appeared as early as 2014, and since then graphics cards and processors have been pushed out completely. Today an attempt to mine DOGE on a GPU does not even cover the electricity, and the gap against an ASIC is measured in thousands of times.

The upside of the algorithm for a hardware owner is that one unit of work serves several chains. AuxPoW lets a single Scrypt ASIC close blocks in Litecoin, Dogecoin and a number of smaller Scrypt networks at the same time. That is a rare situation: the extra coins reach the miner effectively for free, with no rise in consumption and no second farm.

The downside is that the Scrypt hardware market is narrow. There are an order of magnitude fewer models than on SHA-256, the second hand market is thinner, and the spread in efficiency in our catalogue reaches twenty five fold: from 0.149 J/MH on the top machines to 3.75 J/MH on veterans such as the Innosilicon A4 Dominator. Picking the wrong model costs more here than it does in bitcoin mining.

The second downside is structural: since Scrypt economics rest on two coins at once and the bulk of the revenue comes from Dogecoin, price risk concentrates in an asset that has neither a supply cap nor any fundamental link to network usage. That makes Scrypt mining less predictable than mining Bitcoin.

Full description of the Scrypt algorithm →

The block reward is 10,000 DOGE and has not changed since block 600,000, passed in 2015. The early scheme looked different: up to block 145,000 the reward was a random value, then it became deterministic and halved every 100,000 blocks. That ladder ended at block 600,000, and the reward has been fixed ever since. There will be no more halvings in Dogecoin: none are written into the protocol and none are on the table as a plan.

That gives simple issuance arithmetic. Each day the network produces 1,440 blocks of 10,000 coins, that is 14.4 million DOGE per day and about 5.26 billion DOGE per year. This figure is constant in absolute terms and depends neither on price nor on hashrate. By comparison, Bitcoin issuance is cut in half every four years, while here it is perpetual and flat.

There is no supply ceiling. The widely repeated claim about a limit of 100 billion coins is wrong: the cap was removed early on, and the network itself passed the 100 billion mark back in 2015. At the end of July 2026 about 155 billion DOGE are in circulation, and that number grows linearly every day. The protocol has no burn mechanism either, so fees go to the miner rather than being destroyed.

Inflation still falls even though issuance is constant. Because the numerator is fixed while the denominator grows, the annual increase in money supply at the end of July 2026 is about 3.4 percent, against roughly 5 percent in the middle of the previous decade. By the same arithmetic the figure will drop to about 3 percent by the early 2030s and then keep sliding slowly, never reaching zero.

The practical conclusion for a miner is twofold. On one hand there is no income cliff every four years: you do not have to build into your model the moment when revenue suddenly halves. On the other hand 14.4 million new coins reach the market every day indefinitely, and that is constant selling pressure with nothing on the protocol side to offset it.

The full table of Scrypt models from our catalogue, sorted by efficiency: the lower the joules per megahash, the longer a machine stays afloat as difficulty rises and the DOGE price sags.

Browse miners in the catalogue →

The ranking is built on actual efficiency from the catalogue, not on how well known a model is. Break even thresholds are calculated from total network revenue of roughly 0.43 to 0.50 dollars per gigahash per day at the end of July 2026, and they already include the Litecoin payout from merged mining.

Third party firmware builds for Scrypt machines do exist, but coverage is narrow. In our catalogue compatibility is marked for only two models: the Antminer L7 and the Antminer L9. All ElphaPex, Bitdeer SealMiner, Goldshell and Innosilicon hardware runs on factory builds, and those machines are absent from the HashCore Toolkit support lists.

What an open build gives you on the L7 and L9. First and foremost, undervolting: lowering voltage and frequency so that hashrate drops by a few percent while consumption drops by tens. On Scrypt this is critical, because the break even threshold here is set entirely by joules per megahash. Second, per hashboard tuning. Third, proper telemetry with logs of temperatures and chip errors.

The AsicBoost profile does not apply to Scrypt machines: it is an optimisation for SHA-256 and has no effect on Dogecoin or Litecoin. If a supplier promises a gain from AsicBoost on an Antminer L7 or L9, that is a sign they do not understand the subject. On Scrypt the real gain comes only from working with the voltage and frequency curve.

A word about risk. Installing a third party build voids the vendor warranty, and a failed flash may call for recovery through an SD card or a programmer. Before installing it is worth recording the factory version, checking the control board model and making sure the build is meant for your exact revision. The economics do work out though: across a fleet of dozens of machines a gain of 10 to 15 percent in consumption flips the sign of the final profitability.

You do not need to mine Dogecoin separately, and in practical terms you cannot: any modern Scrypt ASIC connects to a pool that runs merged mining automatically and credits DOGE together with LTC. No separate connection, no separate worker and no separate configuration for the second coin are required. It is enough to enter a DOGE address in the pool dashboard.

What you need to start: the ASIC itself, a dedicated electrical circuit with a 16 to 32 amp breaker per machine, stable internet, a DOGE wallet address and a pool account. A wired connection is enough, and the link needs to be narrow but stable. No powerful computer is needed nearby, the ASIC runs on its own.

Noise and heat deserve plain speaking. An industrial Scrypt machine drawing 3 to 4 kilowatts puts out 75 to 85 decibels and a stream of hot air comparable to a fan heater of the same rating. In a flat or a house that is unacceptable without a separate room with intake and exhaust ventilation. Spanish noise rules for residential areas do not allow such an installation.

A home option exists, but calling it income would be dishonest. The Goldshell Mini-DOGE II at 400 W or the ElphaPex DG Home 1 at 620 W are quiet and plug into an ordinary socket, yet their efficiency is six times worse or more than industrial gear. At a Spanish household tariff such a machine consumes more electricity in value than it mines. It is a training rig and a heat source, not an investment.

The industrial scenario calls for a site with a tariff around ten cents per kilowatt hour or below, organised heat removal and spare capacity. On those terms the top models in the catalogue clear their operating costs, but whether the machine itself pays for itself depends on the DOGE price, which is unpredictable.

The income of a Scrypt farm rests on four variables: the DOGE price, the LTC price, the difficulty of both networks and the price of electricity. The key fact at the end of July 2026 is that about 86 percent of the revenue comes from Dogecoin rather than Litecoin. The arithmetic is simple: the DOGE network issues 14.4 million coins a day and the LTC network about 3,600, and at current prices the first number outweighs the second by a wide margin.

From that follows a practical conclusion that people often miss: planning the payback of a Scrypt machine around the Litecoin price is pointless. The farm lives and dies with the DOGE price. And the DOGE price is only loosely tied to network usage and depends almost entirely on the news cycle. Any profitability model with a horizon longer than a few months is inevitably speculative here.

The second factor is hashrate dynamics. From the February 2026 peak to the end of July the power of the network fell by more than two thirds. For the miners who stayed that is a positive signal: difficulty fell along with hashrate, and income per unit of power went up. But the cause is worrying: part of the fleet became unprofitable and was switched off. A move like that usually means the hardware market is close to a price revision.

The third factor is the coming Litecoin halving. The LTC reward will be cut in half around the middle of 2027, and the Dogecoin share of Scrypt mining revenue will grow from the current 86 to roughly 92 percent at unchanged prices. In other words, the dependence on a single meme coin will get stronger. That has to be built into the calculation when buying hardware today, not a year from now.

Open the profitability calculator →

For a farm the scheme is simple: the hardware wallet address is set in the pool dashboard as the payout address, and coins land in cold storage straight away, bypassing hot wallets. Check the address format before entering it: Dogecoin addresses start with the letter D for ordinary ones and with 9 or A for multisig. A Litecoin address will not work in that field, the networks are different and the funds will be lost for good.

Another classic mistake is sending DOGE to a wrapped version of the coin on a foreign network, for example on BSC or Ethereum. The address format there is different and the transfer goes nowhere. There is no recovery in Dogecoin: transactions are irreversible and the network has no support desk. Always make a test transfer of a small amount before your first large payout.

Dogecoin liquidity is high, and that is one of its real advantages for a miner. The coin is in the top dozen of the market by capitalisation and daily turnover stays in the hundreds of millions of dollars. A farm can sell a day or a week of output with no noticeable slippage on practically any large venue.

For reference, with no recommendation: DOGE spot pairs are on Binance, Coinbase, Kraken, OKX, Bybit, KuCoin, Gate and MEXC, as well as with the brokers Robinhood and eToro. Since 2025 regulated exchange traded wrappers have appeared on the US market as well: DOJE from REX Shares and Osprey, the spot GDOG from Grayscale, BWOW from Bitwise and TXXD from 21Shares with double leverage. For a European farm owner these are more an indicator of market maturity than a working instrument.

On security there are three rules. First: do not keep farm revenue on an exchange longer than the sale requires, the keys there are not yours. Second: for regular sales set up fiat withdrawal in advance and check the limits, otherwise the money will get stuck in verification exactly when the price is sagging. Third: any offer to buy hardware or hashrate paid in DOGE through personal transfers is a classic scam, the payment is irreversible and there is nowhere to appeal.

Hardware for Dogecoin is the same Scrypt ASIC gear as for Litecoin, and machines built only for DOGE do not exist and never did. Buying any Scrypt model from our catalogue gets you both coins automatically through merged mining.

Historically the main use case for Dogecoin has been micropayments and tipping. A one minute block and a fee worth fractions of a cent made the coin handy for small transfers long before second layer solutions appeared in other networks. The early tip bots on Reddit and on streaming platforms shut down, but the niche itself remains.

Real payment cases do exist, though there are fewer of them than people assume. Tesla accepts DOGE in its branded merchandise store: that means merchandise only, cars are not sold for the coin, and this is a widespread misconception. The Dallas Mavericks have accepted DOGE for tickets and merchandise through the BitPay processor since 2021. BitPay itself remains the main gateway through which a shop can start accepting the coin without its own infrastructure.

About X, formerly Twitter, we should be honest. Native Dogecoin support in the platform payment service is not officially confirmed as of the end of July 2026. The X help pages describe tipping through third party services and do not list DOGE as a supported asset. Everything published about DOGE payments inside X still belongs to expectations rather than to a working feature.

For a miner the practical conclusion is this: the payment use of the coin has almost no effect on its price today. The network runs at about one percent of its ceiling, and the price is formed by speculative demand rather than by the turnover of goods. Building a farm payback calculation on forecasts of growing payment adoption is not advisable.

The main thing to understand about Dogecoin security is that the network has no hashrate of its own. Since 11 September 2014 the vast majority of blocks have been found by merged mining, that is, by the same work that computes Litecoin. At the end of July 2026 the hashrates of the two networks are practically identical, about 2.6 PH/s for Dogecoin against about 2.5 PH/s for Litecoin, and that is no coincidence but a consequence of the hardware fleet being one and the same.

That leads to a double edged conclusion. On one hand AuxPoW saved the coin in 2014: before it the hashrate was falling and the risk of a 51 percent attack was quite real, and that is exactly what pushed the community to accept the proposal from Charlie Lee. On the other hand the network has no independent protection. If Scrypt mining of Litecoin becomes broadly unprofitable, the hardware will leave and Dogecoin will lose its security automatically, through no fault of its own.

The second risk factor is pool concentration. The three largest operators find about 83 percent of network blocks, and F2Pool alone about 40 percent. Formally this is not an attack and participants can move to another operator at any time, but as a measure of decentralisation the number is unpleasant, and it is noticeably worse than in Bitcoin.

As for protocol level incidents, over its whole history the network has lived through two significant episodes: the breach of the third party Dogewallet service in December 2013, when about 21 million DOGE were stolen, and the period of vulnerability to a 51 percent attack in 2014, closed by the move to AuxPoW. The current client, Dogecoin Core 1.14.9 of 1 December 2024, contains fixes for bugs inherited from the Bitcoin and Namecoin code base, and the developers strongly recommend that all users upgrade.

They write about a Dogecoin halving and even name dates. There are no halvings in Dogecoin and there never will be. Reward cuts ended at block 600,000 in 2015, and the reward has been hard fixed at 10,000 DOGE ever since. Any article with the phrase the next DOGE halving in such and such a year is simply mistaken, the mechanism does not exist in the protocol.

They write that the supply is capped at 100 billion coins. There is no cap. The network passed the 100 billion mark back in 2015, and at the end of July 2026 about 155 billion DOGE are in circulation. The ceiling was removed early in the project, and today the supply is formally infinite.

They write that the reward halves every 100,000 blocks. That was only true up to block 600,000. For the first 145,000 blocks the reward was a random value from 0 to 1,000,000 DOGE, then came a deterministic halving ladder, and it ended in 2015. The schedule is flat now.

They write that Dogecoin inflation is 5 percent. An outdated number. It was close to the truth in the middle of the previous decade. Since issuance is constant in absolute coins while the base grows, at the end of July 2026 the annual increase is about 3.4 percent and it keeps falling.

They list a maximum supply of about 170.9 billion DOGE. That is an artefact of how one large aggregator computes its fully diluted estimate. In the maximum supply field on the same page it honestly says no limit. The real circulating figure is about 155.2 billion and it grows by roughly 14.4 million a day, so no ceiling of 170 billion exists.

They write that Dogecoin has a powerful mining network of its own. The network has practically no hashrate of its own. Since September 2014 blocks have been found by merged mining through AuxPoW, that is, by the same work as Litecoin. The hashrates of the two networks are nearly equal precisely because the hardware is the same, and Dogecoin security is borrowed.

They call Dogecoin a direct fork of Litecoin. Not accurate. Dogecoin is a fork of Luckycoin, and Luckycoin was a fork of Litecoin. The random block reward scheme, which Litecoin never had, came from Luckycoin. The Scrypt algorithm was indeed inherited from Litecoin, but along that chain.

They list a minimum fee of 1 DOGE per kilobyte. Years out of date. The current Dogecoin Core policy is a recommended rate of 0.01 DOGE per kilobyte and a minimum relay rate of 0.001 DOGE per kilobyte. The dust threshold is 0.001 DOGE hard and 0.01 DOGE soft.

They call Dogecoin Core 1.14.5 the current client. That version came out in November 2021. The current stable release is 1.14.9 of 1 December 2024, and it contains fixes for bugs inherited from the Bitcoin and Namecoin code. The developers recommend the upgrade to all users.

They write that Tesla sells cars for Dogecoin. It does not. DOGE acceptance is limited to the branded merchandise store and covers selected items only. Cars are not sold for the coin and no such announcement was ever made.

They claim you can already pay and receive DOGE on X. As of the end of July 2026 there is no official confirmation of native Dogecoin support in the platform payment service. The X help pages describe tipping through third party services and do not name DOGE as a supported asset. This is an expectation, not a working feature.

They publish a badly outdated DOGE price. During our check at the end of July 2026 one large blockchain explorer was serving a Dogecoin price of about 0.40 dollars while the actual level was around 0.072. For a profitability estimate take the data from the calculator or from specialised aggregators, not from the first explorer panel you find.