+34 641 10 27 16

Home/Cryptocurrencies/Monero (XMR)

Monero (XMR): a guide to the coin, the RandomX algorithm and mining it

Coin encyclopedia

Monero (XMR): RandomX mining and the Antminer X5 question

Bitmain itself describes the X5 as a machine built on RISC-V cores, and the measured lead over an ordinary desktop processor is less than 1.3 times. On Spanish rates the machine runs at a loss.

RandomXAntminer X5 is not an ASICThreshold near 0.14 EUR per kWhTail emission 0.6 XMR

Denys Klimchukupdated 7 August 202625 sections

About the coin

Monero is a proof of work blockchain in which privacy is always on and cannot be switched off. The network launched on April 18, 2014 under the name BitMonero, a fork of Bytecoin built on the CryptoNote protocol, and was renamed a few days later. There was no premine, no ICO, and no owning company has ever existed. Development is carried out by a loose community, and the money for it is raised in the open through the CCS system, where every proposal is discussed in public.

Since November 30, 2019 the network has run on RandomX, which replaced the CryptoNight family at block height 1,978,433. The point of RandomX is to make a dedicated chip pointless: every hash generates its own random program that has to be executed on a virtual machine with constant trips to memory. The best tool for that job turns out to be an ordinary central processor, and that is the whole design.

For a hardware buyer this leads to the conclusion the page was written for. The Antminer X5 is not an ASIC in the usual sense of the word. Bitmain itself describes the machine as built on RISC-V architecture processors, that is on general purpose cores, and the measured efficiency lead over an ordinary desktop processor is less than 1.3 times. A real ASIC wins by hundreds and thousands of times. This is not a quibble over wording: the economics of the purchase follow directly from it.

The page is written as a reference for a hardware owner, not as a product card. Below you will find a breakdown of the algorithm, the emission and privacy, a balanced look at the Qubic pool episode of August 2025, an honest breakeven calculation on Spanish electricity rates, and a separate list of figures that major calculators and reference sites publish for this coin in an outdated or distorted form as of early August 2026. A loss is called a loss here.

Coin card

Parameter	Value
Name	Monero, a network with mandatory transaction privacy
Ticker	XMR
Network launch	April 18, 2014, originally named BitMonero
Origin	A fork of Bytecoin on the CryptoNote protocol, the CryptoNote v2 technical paper was published in October 2013
Protocol author	Nicolas van Saberhagen, a pseudonym. The first version of the network was launched by a forum member using the handle thankful_for_today
Premine and ICO	None. Coins were distributed through mining only from the very start
Owning organisation	None. No company, no foundation, no protocol treasury
Development funding	The public CCS crowdfunding system, every proposal is discussed and funded in the open
Mining algorithm	RandomX since November 30, 2019, block height 1,978,433
What came before	The CryptoNight family, versions 0 through 3, including CryptoNight R
Consensus	Proof of work, Nakamoto consensus
Block time	2 minutes, target
Block reward	0.6 XMR, tail emission, in force since May 2022 with no end date
Emission cap	None. The main curve ended at roughly 18,132,000 XMR, after which the network issues 0.6 XMR per block forever
Privacy	Mandatory for every transaction: ring signatures, stealth addresses, RingCT
Merged mining	Not supported at the protocol level

Project history

2013In October the technical paper for CryptoNote version 2 is published under the name Nicolas van Saberhagen. It describes ring signatures and one time recipient addresses, exactly the pair of mechanisms on which all Monero privacy is later built.
2014On April 18 the network launches as BitMonero, a fork of Bytecoin. A few days later the community renames the project Monero, which means "coin" in Esperanto. There is no premine and no presale, distribution runs through mining only. The project never acquires an owning organisation.
2017In January RingCT activates, ring confidential transactions. From that moment the chain hides not only the sender and the recipient but the amount of the transfer as well. A few months later RingCT becomes mandatory for all new transactions.
2018In October Bulletproofs arrive, compact zero knowledge proofs. A typical transaction shrinks several times over, and the fee falls with it. In parallel the network runs a series of CryptoNight forks to devalue the machines that had appeared for the old algorithm.
2019On November 30, at block height 1,978,433, RandomX activates. This is a change of principle: instead of an endless race of forks against dedicated hardware, the network moves to an algorithm deliberately made comfortable for an ordinary processor and uncomfortable for a purpose built chip.
2020Dandelion plus plus arrives, a propagation scheme in which a transaction first travels along a random chain of nodes and only then goes out widely. That makes it harder to tie a transaction to the IP address of the sender. In September of the same year the United States tax authority announces a reward of 625,000 dollars for a working Monero tracing tool.
2022In May the main emission curve runs out at roughly 18,132,000 XMR. The network switches to tail emission: 0.6 XMR per block, with no end date. This is a deliberate decision rather than a side effect: a permanent reward is meant to hold up the security of the network once fees alone no longer cover it.
2022In August, at height 2,689,608, protocol version 16 activates. Ring size grows from 11 to 16, Bulletproofs plus arrive, view tags appear to speed up wallet scanning, and the block weight algorithm is adjusted. As of early August 2026 this is still the protocol version in force.
2023In September Bitmain brings the Antminer X5 for RandomX to market: 212 kH/s at 1,350 W, that is 6.368 J/kH. The machine is widely called the first ASIC for Monero, although Bitmain itself describes it as built on RISC-V architecture processors, and on efficiency it pulls ahead of an ordinary desktop processor by less than 1.3 times.
2024The year of delistings. On January 5 OKX removes the coin, on February 20 Binance drops it globally. Kraken removes XMR in Ireland and Belgium in June, and on October 31 across the whole European Economic Area. Trading shifts noticeably to decentralised venues, direct exchange between people and atomic swaps.
2025In August the Qubic pool episode unfolds. The operator publicly claims to have captured a majority of network power. The actual estimates are more modest: a share on the order of 28 to 39 percent, a short lived edge over the last hundred blocks, a reorganisation 6 blocks deep against the ten confirmations the network requires. No double spends were recorded. The full account is in the network security section.
2025On October 3 the first FCMP plus plus stress test network goes live, a public alpha. In December Bitmain announces the Antminer X9 for RandomX: 1,000 kH/s at 2,472 W, priced around 5,600 dollars, with shipments promised for July 2026. On November 19 client version 0.18.4.4 is released.
2026On May 6 the second FCMP plus plus test network goes live together with the new CARROT address scheme. This is a test network and not the main one, which matters: mistaken claims that FCMP plus plus had launched on mainnet that day spread quite widely. On May 10 the RandomX library version 2.0.1 is released.
2026In mid May Bitmain pulls the Antminer X9 from the market before the first shipments. Resellers return prepayments and the company names no official reason. As of August 1, 2026 the protocol version in force for Monero is still the 16th, from August 2022, no new hard fork is scheduled, and RandomX version 2 is not active on the main network.

How the network works

Monero is built like an ordinary proof of work chain, with one difference of principle: there is nothing to look at inside it. A block shows that a transaction happened, its size and its fee, but not who sent it, who received it, or how much was moved. This is not a mode or an option you can turn on, it is how every transaction on the network has worked without exception since 2017.

Three mechanisms close three different questions. Ring signatures hide the sender: the signature is formed so that the real spent output is indistinguishable from fifteen decoys taken from the history of the chain. Stealth addresses hide the recipient: every transfer generates a one time address that is nowhere publicly linked to the address the recipient gave to anyone. RingCT hides the amount: it is encrypted, and the correctness of the transfer is proved mathematically, without revealing any figures.

Mining, by contrast, is deliberately simple and deliberately accessible. The miner gets a block template from a node or a pool, picks a nonce value and runs RandomX on every attempt. The difference from familiar networks is in what does the running. On SHA-256 the winner is always a dedicated chip, and the only question is whose. On RandomX the winner is an ordinary processor, because the algorithm is built around the things a processor is already good at.

That produces a market picture unusual for the industry. The Monero hashrate is made up not of farms of identical machines but of a very large number of very different processors: home computers, servers, rented cloud instances and, to say it plainly, infected machines belonging to other people. That power connects and disconnects for reasons unrelated to hardware prices, and planning income around it is harder than in networks with a legible fleet of gear.

Building the jobA node or a pool assembles a block template: the header, transactions from the queue and the coinbase payout. The job goes out to the miner. Your own full node is not required here, but solo mining and P2Pool do need one.
Choosing the RandomX keyThe algorithm builds its dataset not arbitrarily but from a key taken from the chain itself, which changes roughly every 2,048 blocks. You cannot pick the key yourself, and that is one of the conditions that keeps the design resistant to dedicated hardware.
Expanding the datasetThe key expands into a cache and a dataset of 2,080 MiB for fast mode, or into a cache of 256 MiB alone for light mode. Fast mode is what a miner needs, light mode is enough to verify somebody else block.
Generating a random programEvery hash attempt creates a separate program from a random set of instructions: integer, floating point, branches and memory accesses. There is no way to lay out a circuit for it in silicon in advance, because the program is different every time.
Execution and memory accessThe program runs on a virtual machine over several passes, and every pass does random reads from the dataset. It is these accesses that make the algorithm depend on the memory subsystem and not on arithmetic alone.
Verification and broadcastThe resulting value is compared against the target. A block that fits goes out to the network, nodes verify it in light mode on 256 MiB, so validation does not need the miner amount of memory and is within reach of even a modest node.
Difficulty retargetDifficulty is adjusted every block over the last 720 blocks, roughly a day, with twenty percent of timestamp outliers discarded. The protocol target is a block every 2 minutes.

Mining algorithm

RandomX came out of an exhausting race. From 2018 Monero changed its algorithm every six months to devalue the next batch of machines, and every change hit honest miners too. In November 2019 the approach changed: instead of running away from dedicated hardware, the network moved to an algorithm where such hardware gives no advantage by its very nature, not by agreement.

The mechanics are simple in idea and awkward for silicon in consequence. Every hash computed generates a unique random program that has to be executed on a virtual machine. It mixes integer operations, floating point operations, branches and, above all, random accesses to a large array of data. A dedicated circuit wins where the operation is known in advance and repeats billions of times in a row. Here it is never known in advance.

Memory in this design is load bearing, not decorative. Fast mode, the one used for mining, needs 2,080 MiB of shared memory for the expanded dataset. Light mode gets by with 256 MiB of cache and gives the same result, but counts noticeably slower, and it is meant for verifying blocks rather than mining them. The split is deliberate: a miner needs the memory, a verifying node does not, which is why you can run a full Monero node on very modest hardware.

The same design explains the unfamiliar unit of measurement. On SHA-256 the count runs in terahashes per second, because one hash is a few passes through a fixed pipeline. On RandomX one hash is the execution of a whole program plus hundreds of random trips to RAM, and memory answers orders of magnitude slower than arithmetic. So the meaningful unit here is kilohashes per second: a good desktop processor puts out tens of kH/s rather than terahashes, and that is normal, not a sign of weak hardware.

The limits of the design deserve an honest mention too. RandomX does not forbid building dedicated machines, it only makes the gain from them small. You can pack a lot of processor cores into one chassis and sell it as a miner, which is exactly what was done. But while the efficiency gain is measured in tens of percent rather than orders of magnitude, such a machine competes not with other miners but with an ordinary server processor, which can also do something else.

Full description of the RandomX algorithm →

Economics and issuance

The main emission curve of Monero ended in May 2022 at roughly 18,132,000 XMR. The curve was smooth, with no halvings: the reward shrank a little with every block rather than being cut in half on an appointed date. That is why nobody discusses the "next halving" for Monero, there simply is no such thing in the design of the network.

From there on tail emission takes over: exactly 0.6 XMR per block, with no end date. With a two minute block that is 720 blocks and 432 XMR a day for the entire world fleet of hardware, or about 157,680 XMR a year. The network has no emission cap, and that is not an oversight but a decision: a permanent reward is meant to hold up the economics of security once fees alone are no longer enough.

Inflation, meanwhile, is low and keeps falling. Around 18.45 million XMR are in circulation, annual growth is about 0.85 percent and drops further, because the numerator is fixed while the denominator grows. For comparison it helps to keep in mind that this is below the inflation target of most central banks, and on top of that the figure is known in advance for any year ahead.

Fees are almost invisible in the overall picture. The average transaction fee sits around 0.0003 XMR, roughly a tenth of a dollar, and it makes up fractions of a percent of miner income. Monero mining returns should be worked out from the same 432 XMR a day divided across the whole network hashrate, and fees can be ignored in that calculation.

There is a flip side that has to be stated plainly. A constant daily emission means that hashrate growth does not increase total miner income, it only spreads that income across more participants. When another large source of processor power joins the network, the daily output of your hardware falls exactly in proportion, and no compensation from the protocol is coming.

Network parameters

Parameter	Value
Mining algorithm	RandomX, a random program for every hash computed
Algorithm activation date	November 30, 2019, block height 1,978,433
Memory for mining	2,080 MiB of shared memory in fast mode
Memory for verification	256 MiB of cache in light mode, same result, lower speed
Algorithm key rotation	The key is taken from the chain and changes roughly every 2,048 blocks, the miner does not choose it
Block time	2 minutes, target
Difficulty retarget	Every block, over the last 720 blocks, discarding 20 percent of timestamp outliers
Block size	Dynamic, the ceiling is twice the median size of the last hundred blocks, above it the reward is penalised
Throughput	On the order of 30 to 45 transactions per block at current load, the ceiling is elastic
Confirmations	10 blocks before funds unlock, roughly 20 minutes
Address format	A primary address starts with the digit 4, a subaddress with the digit 8
Address types	Primary, subaddress and integrated with an embedded payment identifier
Ring size	16, that is 15 decoys, in force since August 2026 under protocol version 16
Merged mining	Not supported, there is no combined mining with a second coin at the protocol level

Which ASICs work

Over the whole life of the algorithm only a handful of machines have been released for RandomX, and none of them delivers the lead over an ordinary processor that people expect from an ASIC.

Model	Hashrate	Power draw	Efficiency
Antminer X5	212 kH/s	1350 W	6.368 J/kH

Browse miners in the catalogue →

Best machines for this coin

The asic.es catalogue holds exactly one position for this algorithm, the Antminer X5, so the ranking is a single row and the comparison against ordinary processors is moved to the table below.

1Antminer X5

Pros. 212 kH/s at 1,350 W, that is 6.368 J/kH. It is a finished box with a web interface, pool setup and monitoring, where you do not have to assemble motherboards, install an operating system and look after a dozen separate machines. Hashrate density per rack unit is noticeably higher than in a build of desktop processors, and servicing is simpler: one network address, one power supply, one item on the books.

Cons. The main drawback has to be named honestly: this is not an ASIC in the usual sense. Bitmain describes the machine as built on RISC-V architecture processors, that is on general purpose cores, and the measured efficiency lead over an ordinary desktop processor is less than 1.3 times. Everything else follows from that: the premium for dedicated hardware never pays for itself, the machine does one job and switches to nothing else, there is no third party firmware for the platform, and the existence of RandomX library version 2 means the balance could shift against it at the next protocol update.

Best suited for. Someone with an industrial rate well below 0.14 euros per kilowatt hour, who buys the machine cheaply on the secondary market and values simple servicing over maximum return per euro invested. Everyone else is better off looking at ordinary server processors, which can be put to other uses if things change.

Payback. The breakeven threshold on electricity alone sits around 0.14 euros per kilowatt hour on our August 1, 2026 estimate. The final rate in Spain for most buyers lies between 0.15 and 0.20 euros. At 0.15 the machine goes into the red by about a third of a euro a day, at 0.20 the loss approaches two euros a day. We quote no payback period under those conditions: there is none, and we will not invent one.

Cooling. Air. 1,350 W of heat and industrial noise levels, a living space is out of the question. It needs a supply of cool air, organised removal of hot air and a dedicated power line

Hardware manufacturers

Bitmain, Antminer X5212 kH/s at 1,350 W, 6.368 J/kH, September 2023

The only RandomX machine in the asic.es catalogue and the only model that actually reached buyers. Bitmain itself describes the design as built on RISC-V architecture processors. The verifiable consequence: on efficiency the machine beats an ordinary desktop processor by less than 1.3 times, whereas a real SHA-256 ASIC beats general purpose hardware by thousands of times. Selling it as an ASIC is convenient in form, but it misleads the buyer about the economics.

Bitmain, Antminer X91,000 kH/s at 2,472 W as announced, about 2.47 J/kH

Announced in December 2025 at around 5,600 dollars with shipments in July 2026. Custom RISC-V cores were claimed, the same architectural idea as in the X5. In mid May 2026 the model was pulled from the market before the first shipments, resellers returned prepayments, and Bitmain named no reason publicly. This machine is not in the asic.es catalogue and never will be: there is nobody to verify the claimed specifications, because the product does not exist.

Ordinary processorsOn the order of 25 kH/s on a current top end desktop processor

The main and in practice the only real competitor on this algorithm. The network hashrate is made up of a very large number of very different processors: home machines, servers, rented cloud instances. The main advantage of this route is not efficiency but the fact that a processor stays useful without mining, while a dedicated box turns into scrap metal when conditions change.

Graphics cardsThey work, but they lose to processors on this algorithm

A rare case where GPUs do not help. RandomX is built around complex control flow, branching and random memory access, and a graphics card is strong at exactly the opposite: masses of identical simple operations. There is no practical sense in mining Monero on graphics cards, and the guides that offer it as a working option are simply carrying habits over from other algorithms.

The market overallOne working model, one cancelled, no competition

The hardware market as of early August 2026 looks like this: the X5 sells on the secondary market, the X9 was cancelled before shipping, and there are no other serious attempts. That in itself says a lot. On algorithms where an ASIC really does deliver a gain, manufacturers compete and ship new generations every year. Here there is nothing to compete for: the gain is too small to pay back the development.

Firmware and overclocking

Straight to it, with no illusions: there is no firmware for RandomX. There are no machines for this algorithm on the HashCore Toolkit support lists, and we do not build and do not plan AsicBoost releases for the Antminer X5. Our tooling targets SHA-256 and Scrypt, where the installed fleet runs to hundreds of thousands of machines and development pays for itself. One catalogue model on RandomX is not a market worth building and maintaining firmware for.

There is no third party market either, and the reason runs deeper than market size. The point of custom firmware on SHA-256 is to tune frequencies and voltages finely to specific chips and win a noticeable share of efficiency that way. On RandomX there is not much to win: the bottleneck here is memory, not arithmetic, and nudging core clocks barely moves the result. If somebody offers to sell you firmware for the X5, that is a reason for caution rather than delight.

What an owner really has instead of firmware. There are three levers, and all of them sit outside the machine: the quality and stability of the supply line, the intake air temperature and the choice of a pool with a lower fee. None of them will give the effect that competent undervolting gives on SHA-256. Run the official manufacturer firmware and put your money into your electricity rate and your cooling, not into the hunt for a miracle build.

A separate word on a risk that is underrated on this algorithm. RandomX library version 2.0.1 exists, released on May 10, 2026, and it changes virtual machine behaviour. On the Monero main network it is not activated and no hard fork is scheduled, but the mere possibility of a protocol update means a dedicated box bought today could lose its already small advantage tomorrow. An ordinary processor in that situation would simply carry on a little slower, while a single purpose machine loses its value entirely.

How to start mining

Starting to mine Monero is easier than in any other network in our reference, because the ticket in is already sitting on your desk. You need a processor, a mining program and a wallet address. No special hardware has to be bought, and that is not a rhetorical flourish: an ordinary processor here is the main tool, not the fallback. Mining needs about 2,080 MiB of free memory per miner process, preferably with large memory pages enabled, otherwise part of the performance is lost.

The Antminer X5 against ordinary processors looks like this. The machine delivers 212 kH/s at 1,350 W, that is 6.368 J/kH. A current top end desktop processor delivers on the order of 25 kH/s, and by rated heat output that lands at roughly the same just over six joules per kilohash. In other words the efficiency difference is less than 1.3 times. Eight such processors give roughly the same hashrate and the same draw as one X5 box. The difference between the options is not in mining but in ease of servicing and in the fact that a processor stays useful without it.

On noise and siting. The X5 at 1,350 W is a full size industrial machine with airflow to match. You cannot put it in a flat under any amount of enclosure trickery: it needs a separate technical space with cool air intake, organised hot air removal and a dedicated breaker on the line. Home mining of Monero is entirely realistic, but by a different route: several processors in ordinary cases are incomparably quieter and can be spread across rooms.

And the most important thing, which other networks do not have. Here you compete not with farms of identical machines but with the entire world fleet of processors, including rented cloud capacity and infected machines whose owners pay nothing for electricity. Such a participant stays in the game at any coin price and any difficulty, because their costs are zero. That is a structural factor pushing the returns of an honest miner down, and it is not going away.

Profitability and what drives it

Income is worked out in one step, because emission is constant. The network issues 432 XMR a day, which is 720 blocks of 0.6 XMR. Divide that by the current network hashrate, multiply by your own hashrate, and you get your daily output in the coin. Fees can be left out, they are fractions of a percent. Then multiply by the exchange rate and compare against your electricity bill. The whole economics of Monero is the ratio of those two numbers, with no hidden multipliers.

Four factors move the result. The first and biggest is network hashrate, which behaves differently on RandomX than in networks with ASICs: it can grow noticeably in a week simply because somebody plugged in a large fleet of rented or borrowed processors, and your share falls with no connection to the hardware market at all. The second is the coin price. The third is your electricity rate. The fourth is the pool fee, usually from zero to one percent, and against the other three it barely matters.

The honest conclusion for the Antminer X5 on Spanish rates. The breakeven threshold on electricity alone sits around 0.14 euros per kilowatt hour as of August 1, 2026. The average wholesale electricity price in Spain in July 2026 was about 104.7 euros per megawatt hour, that is roughly 0.105 euros per kilowatt hour, but that is the exchange component. The final bill with grid charges and taxes for most buyers lands between 0.15 and 0.20 euros per kilowatt hour.

What that means in practice, without softening. At a rate of 0.15 euros the machine goes into the red by about a third of a euro a day, at 0.20 euros the loss approaches two euros a day. That is a loss on electricity alone, before the cost of the machine, the premises and servicing. A payback period does not exist under those conditions and we will not calculate one. If a seller shows you a four to six month payback on the X5, ask them for the rate and the price that produced the figure.

Open the profitability calculator →

Mining pools

Pool	Payout scheme	Fee	Note
P2Pool	PPLNS on its own sidechain	Zero percent	A decentralised pool with no operator and no custodial balance. The reward is paid straight into the coinbase transaction of a Monero block, so your coins never sit in anyone else account for a minute. There is no fee, and the minimum payout is 0.00027 XMR according to the official project site. It requires your own full Monero node. At a small hashrate the first payout can take several days or a week.
SupportXMR	PPLNS	About 0.6 percent	The largest classic pool on the network, holding on the order of 52 percent of network power as of early 2026. Minimum payout 0.1 XMR. It is worth understanding the double nature of that lead: the pool grew partly because miners deliberately moved power to it in August 2025 in response to the Qubic affair. For the network such concentration is a risk in its own right.
NanoPool	PPLNS	About 1 percent	The second largest operator, on the order of 21 percent of network power as of early 2026. Minimum payout 0.11 XMR. A large multi coin pool with a long history and straightforward infrastructure, a sensible choice for anyone who wants predictable daily credits.
C3Pool	PPLNS	Zero percent	On the order of 5.6 percent of network power. Minimum payout 0.001 XMR, one of the lowest among classic pools, which is handy with a small fleet. A declared zero fee is usually made up for in other terms, so read the pool rules before connecting.
Kryptex Pool	PPS plus	About 1 percent	On the order of 5.9 percent of network power, minimum payout 0.01 XMR. The only notable option running PPS plus, that is payment for submitted shares regardless of pool luck. That reduces payout variance and suits a small hashrate, but the total fee is usually a little higher.
MoneroOcean	PPLNS	Zero percent	On the order of 4.9 percent of network power, minimum payout 0.003 XMR. The feature of this pool is automatic switching of hardware to whichever algorithm pays best at the moment, with payout in XMR. For a processor fleet that is sensible, for a single purpose box like the X5 the feature is pointless.
HashVault	PPLNS	About 0.9 percent	On the order of 3.6 percent of network power, minimum payout 0.001 XMR. A long running pool with detailed per worker statistics and a convenient monitoring interface, a sensible backup when your main pool has trouble.
Solo mode	Solo	Depends on your node	Running your own full Monero node and mining straight into it. With a network hashrate on the order of 5.7 GH/s, one X5 box at 212 kH/s finds a block on average about once every fifteen years, so solo is not a way to earn income at all. The sensible alternative for anyone who wants independence is P2Pool: it gives the decentralisation of solo and the predictability of a pool at the same time.

The main thing about choosing a pool on Monero is that the choice has a network dimension as well as an economic one. As of early 2026 SupportXMR alone held on the order of 52 percent of network power, and the two largest pools together held more than 70 percent. That is exactly the concentration that makes episodes like August 2025 possible. So on this network choosing P2Pool is not an ideological gesture but a practical contribution to the stability of the very chain whose coin pays you. A reminder on the schemes: PPS pays for submitted shares regardless of pool luck and suits a small fleet, PPLNS pays for blocks actually found and gives more variance over a short horizon. Pool shares are quoted as of early 2026 and they change, so check them before you connect.

Wallets

Category	Wallets
Official desktop	Monero GUI, the graphical wallet of the project with a built in full node. The current branch is 0.18.4.4 from November 19, 2025, with builds for Windows, macOS and Linux
Official console	Monero CLI, the monerod daemon and the command line wallet. The canonical choice for a server, for solo mining and for working with P2Pool
Light desktop	Feather, a compact wallet with Tor support that can work either with your own node or with a remote one
Mobile	Cake Wallet for iOS and Android with built in swaps, Monerujo for Android with Tor support and hardware wallet pairing
Hardware	Ledger Nano S, Nano S Plus and Nano X, plus Trezor Model T and Safe 3, all through the official Monero GUI
What does not exist	Mainstream multi coin wallets such as MetaMask do not support XMR: Monero has its own cryptography and its own accounting model, and there is no compatibility with the Ethereum environment here at all

An important quirk that sets a Monero wallet apart from the familiar kind. You cannot simply look up an address balance in an explorer: nobody can see it in the chain, including you. The wallet works the balance out itself by scanning every network transaction with your view key. That is why the first launch and sync take noticeable time, and since August 2022 the process has been sped up by view tags, introduced in the sixteenth version of the protocol.

The same design leads to a practical tip for miners. The view key can be handed to a third party, an accountant or an auditor for example, and they will see your incoming funds but will not be able to spend anything. It is a convenient way to show mining income without giving up control of the funds. For large amounts use a hardware wallet through the official Monero GUI and keep the mnemonic seed offline: restoring a Monero wallet is only possible from it.

Where to buy and how to store

This section is for reference, we do not trade the coin and we give no investment advice. On Monero the access situation has to be stated plainly, because it directly concerns a miner: what you mine has to be sold. Over two years the coin was removed from most major venues. OKX dropped it on January 5, 2024, Binance dropped it globally on February 20, 2024, and Kraken removed it first in Ireland and Belgium in June 2024, then across the whole European Economic Area on October 31, 2024.

What is left as of early August 2026. The coin trades on KuCoin, MEXC, Gate.io, TradeOgre and on Kraken outside the European Economic Area. A major aggregator counts about twenty active venues and some fifty markets with daily turnover in the region of a hundred and fifty million dollars. In parallel the weight of direct exchange between people and of atomic swaps between Bitcoin and Monero, which need no intermediary at all, has grown noticeably.

The regulatory frame in the European Union looks like this. The anti money laundering regulation adopted on May 31, 2024 bars supervised service providers from handling anonymity enhancing coins from July 10, 2027. The ban is addressed to venues, not to private individuals: owning the coin, keeping it in your own wallet and transferring it are not prohibited by this regulation. But a miner should think ahead about which channel will turn mined coins into euros after that date.

On security the rules are standard and particularly apt here. Do not leave what you mine on an exchange longer than the trade requires. Withdraw to your own wallet, keep the mnemonic seed offline, and use a hardware wallet for meaningful amounts. Venue names in the text are given for reference and are not a recommendation, and tax obligations from mining and selling the coin arise regardless of how private the network itself is.

Strengths and weaknesses

Pros

The barrier to entry is effectively zero. You can mine on a processor you already own, with no purchase of dedicated hardware and no risk of being left with a pile of single purpose gear.
Hardware does not lose its value when the algorithm changes. A processor stays a processor after any protocol update, which cannot be said of a dedicated box.
Privacy works by default and for everyone, not as an option you have to switch on. That sets Monero apart in principle from networks with optional privacy, where shielded transactions stay a minority and are therefore conspicuous.
Emission is predictable for any horizon ahead: 0.6 XMR per block with no end date, 432 XMR a day, about 0.85 percent a year and falling. There will be no halvings and no income cliffs that come with them.
The network has had no owner, foundation or treasury since 2014, and development is funded by public crowdfunding. There is nobody to shut it down or force a change of policy.
P2Pool offers a rare chance in this industry to mine without trusting a pool operator: the payout arrives straight in a network block, there is no fee, and the minimum payout is 0.00027 XMR.
Steady practical demand. Unlike many narrow algorithms, the coin is genuinely used for payments, and that supports liquidity even after the delistings.

Cons

The Antminer X5 runs at a loss on Spanish rates. The electricity breakeven threshold is about 0.14 euros per kilowatt hour against final rates of 0.15 to 0.20. There is no payback on the machine, and no calculation will change that.
The X5 does not deliver the advantage people buy dedicated hardware for. Bitmain describes it as built on RISC-V cores, and the measured lead over an ordinary processor is less than 1.3 times at an order of magnitude difference in price.
You compete with the entire world fleet of processors, including rented cloud capacity and infected machines for which electricity is free. Such a participant does not leave the network at any coin price.
The risk of an algorithm change is real. RandomX library version 2.0.1 has existed since May 10, 2026, and although it is not activated on the main network, a protocol update could devalue a dedicated purchase. The cancellation of the Antminer X9 in May 2026 before the first shipments shows the manufacturer rates that risk too.
Pool concentration is high: SupportXMR alone held on the order of 52 percent of network power as of early 2026. That is what makes episodes like August 2025 possible.
Exchange access is shrinking. OKX, Binance and Kraken removed the coin in the European Economic Area, and from July 10, 2027 supervised venues in the European Union are barred from handling it at all.
Absolute return per unit of hardware is low by design: the entire daily emission of the network is 432 XMR spread across the whole world fleet of processors, and hashrate growth does not raise that figure, it only dilutes it.

What the coin is used for

The main use of Monero is payments that outsiders have no business knowing about. This is not only about what people tend to assume: privacy of the amount and the counterparty matters to an ordinary business no less than to a private person. A public blockchain reveals turnover, suppliers and deal terms to competitors, and Monero does not. In practice the coin is accepted by shops with direct integration, subscription services, independent developers and venues where the confidentiality of a settlement matters.

Inside the ecosystem itself the coin is used literally as intended: payments to developers through CCS go out in XMR. A proposal is discussed in public, funding is raised in the open, the work is accepted by the community, and payment leaves in the coin of the network. It is a rare case where the funding method of a project and its own product coincide, and it also explains why the network has neither a treasury nor a foundation: money is not accumulated, it is raised for a specific task.

The technical layer of use is atomic swaps between Bitcoin and Monero, that is direct exchange between two people with no intermediary and no need to trust each other. After the wave of delistings in 2024 this channel grew markedly in importance: it does not depend on the decision of an exchange and requires no checks to pass. For a miner that matters in practice, because it gives a route to sell what was mined that is not tied to a venue that may leave your jurisdiction.

The limits are worth naming too. Privacy demands discipline: it protects the transaction itself, but not against you linking an address to your own name at an exchange or a shop. Beyond that, mandatory privacy is precisely the cause of regulatory pressure: it is why the coin is pulled from venues, and from July 10, 2027 supervised service providers in the European Union are barred from handling it. Ownership and transfers between private individuals are not affected.

Network statistics

Parameter	Value
Price	Roughly 320 to 345 dollars, about 280 to 300 euros. Sources differ by as much as 15 percent, see the note
Market capitalisation	Roughly 6.0 to 6.4 billion dollars, in the third ten of the market
Network hashrate	5.658 GH/s per the chain explorer, about 5.71 GH/s per trackers
Difficulty	679,022,268,801
Chain height	3,730,862 blocks
Block reward	0.6 XMR, tail emission, unchanged since May 2022
Daily emission	432 XMR for the whole world fleet of hardware, that is 720 blocks of 0.6 XMR
In circulation	About 18,446,744 XMR, there is no emission cap
Annual inflation	About 0.85 percent and falling
Transactions per day	About 23,600
Average fee	About 0.00029 XMR, on the order of a tenth of a dollar
All time high price	797.73 dollars, January 14, 2026

All figures were taken on August 1, 2026 from the xmrchain.net chain explorer, difficulty trackers and market aggregators, and they change constantly: price and turnover within hours, hashrate and difficulty within a day. A separate warning about the price. On the date they were taken, major sources disagreed very sharply: one aggregator showed about 316.8 dollars, another about 343 dollars, while exchange quotes on the same page ran as high as 363 dollars. The spread reaches 15 percent, so we give a range rather than a single value and strongly advise taking the price from your own source at the moment of calculation. The euro to dollar rate on that date was 1.1536. We deliberately do not publish the solo mining share or the exact node count: neither could be confirmed from a trusted source on the build date of the page, and we will not invent figures. For income calculations use fresh calculator data rather than this snapshot.

Network security

The security of Monero rests on a base unusual for the industry. In networks with ASICs an attack requires building or renting a fleet of dedicated machines, which is expensive and conspicuous. Here it is enough to rent processor capacity, of which there is plenty in the world and which is available to anyone with a bank card. On the other hand the honest hashrate of the network is, for the same reason, very widely distributed geographically and across owners. This is a different threat model, not a weaker or a stronger one.

The August 2025 episode deserves a balanced account, because it is retold very inaccurately. The operator of the Qubic pool publicly claimed to have captured a majority of network power. The actual estimates are more modest: a share on the order of 28 to 39 percent, while at some moment a short lived edge over the last hundred blocks did indeed appear, which is not the same as a sustained majority. The method used was withholding found blocks, not rewriting the history of the chain.

What happened by the facts. A chain reorganisation 6 blocks deep was recorded and about sixty blocks were orphaned. Monero requires 10 confirmations before funds unlock, so transactions confirmed under network rules were not reversed. No double spends were publicly recorded. Some venues temporarily raised their confirmation requirements. It ended with a distributed denial of service attack against the infrastructure of the operator itself, after which coordination among participants fell apart and the claimed power collapsed, while the community moved part of the hashrate to other pools in response.

The real systemic risk here is a different one and it has not gone anywhere: pool concentration. As of early 2026 SupportXMR alone held on the order of 52 percent of network power, and the two largest pools together held more than 70 percent. That, not the cost of renting processors, is what makes such episodes possible. It is exactly why P2Pool is taken so seriously on Monero: it spreads coordination across participants instead of funnelling it into a single point. At the protocol level, over the whole life of the network since 2014, no major incident with loss of funds has been publicly recorded.

Compatibility and requirements

Parameter	Value
Hardware type	Central processors. No dedicated machine with a real advantage on RandomX exists
Models in the catalogue	One, the Antminer X5 at 212 kH/s and 1,350 W
What is inside the X5	By the description from Bitmain itself these are RISC-V architecture processors, that is general purpose cores rather than a custom circuit for the algorithm
Graphics cards	They work but lose to processors. The algorithm is built on branching and random memory access, not on masses of identical operations
Memory for mining	2,080 MiB per miner process, preferably with large memory pages enabled
Our firmware	None. The algorithm is not on the support list of our tooling
Third party firmware	No mature public builds for RandomX exist, and there is almost nothing to win by nudging clocks here
Power supply	The Antminer X5 ships with its own, up to 1,350 W on the line. For a processor build, an ordinary quality unit with headroom
Cooling and noise	The X5 is air cooled at industrial noise levels, a living space is out of the question. A processor build is incomparably quieter
Full node requirements	About 230 gigabytes for the chain on a solid state drive, 4 gigabytes of RAM, open port 18080

Comparison with similar coins

Option	Hashrate	Power draw	Efficiency	Electricity threshold	Status
Antminer X5	212 kH/s	1,350 W	6.368 J/kH	About 0.14 EUR per kWh	In the catalogue, secondary market
Top end desktop processor	About 25 kH/s	About 170 W	About 6.8 J/kH	About 0.13 EUR per kWh	Ordinary hardware, many already own it
A build of eight such processors	About 204 kH/s	About 1,360 W	About 6.8 J/kH	About 0.13 EUR per kWh	The same result as the X5
Antminer X9	1,000 kH/s as announced	2,472 W as announced	About 2.47 J/kH as announced	Not calculated, the product does not exist	Pulled in May 2026 before the first shipments

The electricity threshold is the rate at which an option breaks even on the power bill alone, without the cost of the hardware, the premises or servicing. The calculation was made on August 1, 2026 with network daily emission of 432 XMR, a hashrate of about 5.7 GH/s and a price on the order of 280 to 300 euros per coin. Look at the two middle rows, the table was built for them: an ordinary desktop processor lands at roughly the same joules per kilohash as the dedicated box, and a build of eight such processors gives roughly the same hashrate at roughly the same draw. The efficiency difference is less than 1.3 times. For comparison, a real SHA-256 ASIC beats general purpose hardware by thousands of times, and it is exactly that gap which justifies buying a narrowly specialised machine. Caveats: processor power draw is taken from rated heat output, the real bill including the rest of the system will be higher and the threshold correspondingly lower. The Antminer X9 figures are from the December 2025 announcement and cannot be verified: the model was pulled before the first shipments. The Antminer X9 and processor builds are not in the asic.es catalogue, they are here only for comparison.

Outdated data and common errors

The Antminer X5 is an ASIC for Monero. The most widespread formulation on product cards, in hardware aggregators and in reviews. It misleads on the very thing a dedicated machine is bought for. Bitmain itself describes the X5 as built on RISC-V architecture processors, that is on general purpose cores, and the measured efficiency lead over an ordinary desktop processor is less than 1.3 times. A real ASIC wins by hundreds and thousands of times. It is more accurate to describe the X5 as a dense build of general purpose processors in a miner chassis.

The Antminer X5 pays for itself in four to six months. Such periods appear in seller marketing materials and in showcase reviews. Current figures do not support them. On our August 1, 2026 calculation the breakeven threshold on electricity alone sits around 0.14 euros per kilowatt hour, while the final rate in Spain for most buyers is 0.15 to 0.20 euros. The machine is in the red on power alone, so there is no capital payback at all, not merely a slow one. We do not calculate payback periods where none exist.

The Antminer X9 can be ordered, shipping in July 2026. Preorders for the X9 still sit on several shops with wording about a July 2026 batch. Bitmain pulled the model from the market in mid May 2026 before the first shipments, and resellers were returning prepayments. Not a single machine reached a buyer, which means the claimed 1,000 kH/s at 2,472 W has never been independently verified. Paying for a preorder of a withdrawn model is not worth it at any discount.

RandomX version 2 was activated on the main network in January 2026. Incorrect, and the error appears in analytical reviews. RandomX library version 2.0.1 did indeed come out, but on May 10, 2026, and it is a library release rather than a network update. It is not activated on the Monero main network. As of August 1, 2026 the protocol version in force is still the sixteenth, activated in August 2022, and no new hard fork is scheduled.

FCMP plus plus went live on the main network on May 6, 2026. Also incorrect, and the confusion came from the same circle of sources. What went live on May 6, 2026 was the second test network, the FCMP plus plus beta stressnet, together with the new CARROT address scheme. The first stressnet started on October 3, 2025. Neither FCMP plus plus nor CARROT is active on the main network as of the build date of this page, and no protocol update date is set.

In August 2025 Monero survived a successful 51 percent attack. The wording spread widely, including through major industry outlets and statements by the pool operator itself. The balanced picture is different. The share was on the order of 28 to 39 percent, the edge over the last hundred blocks was short lived, and the method was withholding found blocks rather than rewriting the history of the chain. The reorganisation was 6 blocks deep against the 10 confirmations required, no double spends were publicly recorded, and it ended with the coordination of the operator falling apart after an attack on their infrastructure.

Ring size in Monero is 11. Four years out of date. Ring size was raised to 16, that is 15 decoys, in August 2022 by the sixteenth version of the protocol, the same update that brought Bulletproofs plus and view tags. The figure 11 still turns up in reference sites and in tutorials, including ones that look recent.

Monero has an emission cap. There is no cap and none is planned. The main curve ended in May 2022 at roughly 18,132,000 XMR, after which open ended tail emission of 0.6 XMR per block took over. Reference sites that plug in a fixed maximum here by analogy with Bitcoin are wrong conceptually: a permanent reward in Monero is a deliberate decision, not an oversight.

Monero is mined on graphics cards like everything else. On this algorithm graphics cards lose to processors, and that is a rare case in the industry. RandomX is built around complex control flow, branching and random memory access, while a graphics card is strong at masses of identical simple operations. Guides that offer GPUs as the main way to mine XMR are simply carrying habits over from other algorithms.

Monero privacy works the same way as in networks with shielded addresses. The difference is fundamental and matters in practice. In networks with optional privacy, shielded transactions are a minority, and the very fact that the shielded mode is being used is visible from outside. On Monero privacy is mandatory for all transactions since 2017, so there is nothing to stand out with: anonymity comes not only from the cryptography but from everyone else behaving in exactly the same way.

Monero hashrate reflects sales of dedicated hardware. It does not, and this trips people up when planning. Network power is made up of a very large number of ordinary processors, including rented cloud instances and infected machines belonging to other people. It can grow or fall noticeably in a week for reasons tied neither to the coin price nor to the hardware market, and forecasting it by the logic of networks with ASICs does not work.

Questions and answers

What is Monero and what is it for

It is a proof of work blockchain in which privacy is always on and cannot be switched off. The chain shows neither the sender, nor the recipient, nor the amount of a transfer. The network has been running since April 18, 2014.

Who created Monero

The network has no single creator. It was launched as a fork of Bytecoin by a forum member using the handle thankful_for_today, and the CryptoNote protocol was described in 2013 under the pseudonym Nicolas van Saberhagen. There was never a premine, an ICO or an owning company.

Who develops the network and with whose money

A loose community of developers and researchers. Funding runs through the public CCS crowdfunding system: a proposal is discussed openly, funds are raised in public, and payment goes out in XMR. There is no foundation and no protocol treasury.

Which algorithm does Monero use

RandomX, activated on November 30, 2019 at block height 1,978,433, replacing the CryptoNight family. Every hash generates its own random program that has to be executed on a virtual machine.

Why is hashrate measured in kilohashes rather than terahashes

Because one hash here is the execution of a whole program plus hundreds of random trips to RAM, not a few passes through a fixed pipeline. Tens of kH/s on a good processor is a normal result, not a sign of weak hardware.

Is it true that the Antminer X5 is not an ASIC

In substance yes. Bitmain itself describes the machine as built on RISC-V architecture processors, that is on general purpose cores, and the measured efficiency lead over an ordinary desktop processor is less than 1.3 times. A real ASIC beats general purpose hardware by hundreds and thousands of times.

What happened to the Antminer X9

The model was announced in December 2025 with specifications of 1,000 kH/s at 2,472 W and shipments in July 2026, and in mid May 2026 Bitmain pulled it from the market before the first shipments. Resellers returned prepayments and the company named no reason publicly.

Is the Antminer X5 worth buying today

Only with an industrial rate well below 0.14 euros per kilowatt hour and only if you buy it cheaply on the secondary market. On ordinary Spanish rates the machine is in the red on electricity alone, and there is no payback on the investment.

Can Monero be mined at home

Yes, and this is exactly the case where home mining makes sense: an ordinary processor here is the main tool, not the fallback. What you cannot do at home is run an Antminer X5 at 1,350 W and industrial noise, that one needs a technical space.

Which ASIC is best for Monero

The question is imprecise, because no machine with a real advantage on RandomX exists. The catalogue holds one position, the Antminer X5, but its lead over an ordinary processor is less than 1.3 times, and on European rates it runs at a loss.

Can Monero be mined on a graphics card

Technically yes, but graphics cards lose to processors here, and that is a rare case in the industry. The algorithm is built on branching and random memory access, not on the masses of identical simple operations a graphics card is strong at.

How much memory does mining need

RandomX fast mode needs 2,080 MiB of shared memory per miner process, preferably with large memory pages enabled. Verifying blocks from others only needs light mode on 256 MiB, which is why a full node can run on modest hardware.

Which pool should I choose

If independence and the stability of the network matter to you, choose P2Pool: no fee, the payout arrives straight in the block, minimum 0.00027 XMR, but you need your own full node. If you want simplicity and predictable credits, SupportXMR, NanoPool or HashVault will do.

How is P2Pool different from an ordinary pool

It has no operator and no custodial balance: the reward is paid straight into the coinbase transaction of a Monero block, so your coins never sit in anyone else account. For a network with power concentrated in a couple of large pools that matters a great deal.

What about halving in Monero

There are no halvings in Monero. The main emission curve was smooth and ran out in May 2022, after which open ended tail emission of 0.6 XMR per block took over. The income cliffs seen in Bitcoin do not happen here.

Does Monero have an emission cap

No. The main curve ended at roughly 18,132,000 XMR, after which the network issues 0.6 XMR per block with no end date, which is 432 XMR a day. Annual inflation is about 0.85 percent and falling.

Why does difficulty change

Difficulty adapts to the actual power of the network in order to hold a block every 2 minutes. The retarget runs every block over the last 720 blocks, with twenty percent of timestamp outliers discarded.

Can third party firmware be installed on the Antminer X5

There is no firmware for this algorithm, neither from us nor on the third party market. There would be little point: the bottleneck on RandomX is memory, not arithmetic, and nudging core clocks barely moves the result. An offer to sell you firmware for the X5 is a reason for caution.

How can power consumption be reduced

For the Antminer X5 only external levers are available: stable power quality, lower intake air temperature and a pool with a lower fee. A processor build gives noticeably more freedom: power limits and power profile settings work with standard tools.

How loud is the Antminer X5

As loud as any full size industrial machine at 1,350 W: a constant noise at the level of an industrial fan. Unacceptable in a living space with any amount of chassis sound proofing. A processor build is incomparably quieter.

What power supply is needed

For the Antminer X5 the stock one that comes in the box, up to 1,350 W on the line, with a dedicated breaker. For a processor build an ordinary quality unit with about thirty percent headroom over real draw is enough.

Will Monero mining pay off

On the Antminer X5 at Spanish rates, no. The breakeven threshold on electricity alone is about 0.14 euros per kilowatt hour against a final rate of 0.15 to 0.20. Mining on a processor you already own is a different conversation, there is no capital outlay to recover there.

What happens to income if the price falls

It falls in direct proportion, because the reward is fixed in the coin: 432 XMR a day for the whole network regardless of price. A dedicated machine has no margin to spare at current rates, so any drop in price immediately increases the daily loss.

Does Monero have merged mining

No. Combined mining with a second coin is not provided for at the protocol level. Some pools offer automatic switching of hardware between algorithms with payout in XMR, but that is redistribution of power, not merged mining.

How many confirmations are needed and what is the network fee

The protocol requires 10 confirmations before funds unlock, roughly 20 minutes with a two minute block. The average transaction fee holds around 0.00029 XMR, on the order of a tenth of a dollar.