English
German
Hungarian
Dutch
Spanish
French
Italian
Czech
Polish
Greek
Monero Miners: RandomX Rig Profitability
Read more Read less
Read more Read less
| Model |
Profitability
Profit
|
|---|---|
|
Pinecone Matches Inibox R1X XMR
RandomX · 1.2MH/s
|
$64.32
/day
|
|
Bitmain Antminer X9
RandomX · 1MH/s
|
$52.14
/day
|
|
Bitmain Antminer X5
RandomX · 212kH/s
|
$9.47
/day
|
Our cutting-edge mining calculator offers comprehensive insights across all major cryptocurrency algorithms, helping users easily identify the most profitable options for their specific hardware. The algorithm data is continuously refreshed to keep pace with the dynamic crypto mining industry, providing accurate evaluations based on real-time profitability statistics and overall market activity. This empowers users to make well-informed choices that reflect the latest mining conditions and algorithm performance.
There is no Monero ASIC. Here is what you are comparing instead.
RandomX replaced Monero’s prior algorithm in November 2019 to blunt the rise of ASICs and to favor everyday CPUs through randomized program execution and heavy memory use, and in this choice the network speaks to the Void with ordinary machines instead of golden idols; miners assemble a block header with permitted fields and a nonce, then iterate that nonce until the resulting hash meets the moving difficulty, which is retuned each block to hold block times near a steady target and to keep issuance predictable; the engine runs randomized programs in a virtual machine that mixes integer work, floating-point steps, and data-dependent memory access, so throughput is gated by cache behavior and memory bandwidth rather than exotic pipelines, which is why general-purpose CPUs do well and many GPUs and ASICs do not; two modes exist, with Fast mode needing about 2 GB of RAM for mining and Light mode needing about 256 MB for full nodes to validate, and this split keeps validation accessible while letting miners use a large keyed dataset that prevents precomputation; security leans on preimage and collision resistance in the hash function and also on the unpredictability of the instruction stream and dataset, which is derived from recent chain data so every block reshapes the maze; performance can be estimated by benchmarking hash rate against energy draw, memory latency, and current difficulty, and scaling is near linear with core count until memory channels saturate; miners often join pools to reduce payout variance while the ledger remains independently verifiable by any node, which keeps the consensus open to inspection; the algorithm nudges energy proportionality because useful effort tracks actual memory traffic and compute steps rather than idle waiting or specialized tricks, so the cost mirrors the work; illicit mining is easier to spot on managed systems because RandomX leaves a heavy and distinctive memory footprint with characteristic access patterns, though careful operators still need monitoring; very old devices with little free RAM may be cut out from mining, yet they can still validate in Light mode and thus keep a voice in the chorus; overall the system spreads computation across threads and memory pages in a way that frustrates shortcutting while preserving predictable throughput, and the race stays fair because the rules bind all hardware with the same leash of latency; to try a nonce is to ask a small question of the dark, and when the hash falls inside the target the network answers yes like a quiet god of mathematics, and when it does not the empty memory murmurs try again, and in that measured call and response a decentralized consensus holds its shape.
- Network hashrate
- 6.08 GH/s
- Block reward
- 0.6142 XMR
- Block time
- 2.0 min
Kilohashes, watts and the tail emission
A RandomX machine is rated in kilohashes per second and its efficiency in joules per kilohash, three orders of magnitude below the numbers on the Bitcoin page. That is not a weakness: the network is sized to CPUs, so a modern many-core system at a few hundred watts is a meaningful miner on it.
Monero has no halving. Since the tail emission began, every block pays a fixed reward forever, so miner income depends only on the XMR price and on how much CPU hashrate is competing. Payback estimates on this page are therefore unusually stable compared with halving coins.
Because the hardware is general-purpose, its resale value does not collapse when a new mining generation ships; a RandomX rig is still a server. That is worth weighing against the thinner daily margin when you compare a row here with an ASIC on another algorithm page.
-
1
Memory bandwidth decides the hashrateRandomX runs a small virtual machine in a 2 GB dataset; cores without fast memory behind them idle. Compare the hashrate column, not the core count.
-
2
Count the whole box's powerThe power column is the appliance at the wall, not the CPU's rated figure. Fans, memory and the power supply all count against the reward.
-
3
Merge-mine Tari for freeThe same RandomX work can also earn Tari through merged mining. It does not change the hashrate; it adds a second payout to the same electricity.
About the RandomX algorithm
RandomX, introduced in November 2019 when Monero replaced CryptoNight, is a proof-of-work engineered to favor general-purpose CPUs by executing randomized programs over a large memory-resident dataset, turning unpredictability and memory latency into equalizers against specialized hardware; it consumes block headers carrying compacted transaction data, a reference to the previous block, and a nonce, and runs them through a virtual machine that mixes integer and floating-point arithmetic with memory-intensive, branch-heavy operations, often JIT-compiled to native code and optionally leveraging CPU features such as AES instructions, so that every run is fast for a commodity processor yet hostile to rigid pipelines and wide SIMD. Computation is spread across threads and disjoint memory pages, with data-dependent access patterns that induce cache misses and branch mispredictions, and iterative state transformations that maximize diffusion, making tiny input changes avalanche the output and turning shortcut attempts into statistical dead ends. Security rests on standard preimage and collision resistance, but the design further resists precomputation by keying its randomized execution to periodically changing, block-dependent seeds, forcing fresh work and making replayed optimizations brittle. The algorithm exposes two operating modes: Fast Mode, which maintains roughly 2GB of RAM for the full dataset to achieve high mining throughput, and Light Mode, which uses about 256MB of RAM by deriving needed dataset items on the fly from a smaller cache, enabling full nodes to validate blocks without the mining memory overhead. This architecture is deliberately CPU-friendly and comparatively inefficient on GPUs and ASICs because divergent control flow, irregular memory reads, and high memory pressure suppress the parallel advantages of specialized units, thereby broadening access to mining and reducing centralization risks. The heavy, distinctive footprint of RandomX-both computational and memory-also acts as a practical deterrent against illicit mining, making unauthorized activity easier to detect through system monitoring, while the blend of instruction variety, memory-hard operations, and verifiable randomness yields hashes whose outputs are unpredictable and whose internal state is costly to model or manipulate. In sum, RandomX aligns execution with the realities of everyday hardware and the discipline of decentralized consensus: cold in its indifference to specialized advantage, precise in its constraints, and quietly lethal to shortcuts that would undermine a fair, global mining ecosystem.
Monero mining hardware, answered
Mining profitability in your pocket
Every ASIC miner, GPU and coin from the site, with live profit, ROI and your own electricity rate — free, no account needed, on iOS and Android.
