LSN โ€” Local Sovereign Network

๐Ÿ’ฐ TI-2026-025D โ€” The Economics

Monero, Mining Pools, and the Mathematics of Parasitic Revenue

Series: The SSH Parasite (TI-2026-025)  |  Part: 4 of 6
Classification: FINANCIAL INTELLIGENCE  |  Confidence: HIGH (computational analysis)
Sources: XMRig configuration analysis, Monero network data, botnet scale measurements, energy cost modeling
Date: June 2026  |  Analyst: LSN Threat Intelligence

Chapter 1: The Business Model

Strip away the folklore and the IRC channels and the eight-year persistence and what remains is an economic proposition: convert stolen electricity into cryptocurrency at zero marginal cost.

The mdrfckr botnet is not a ransomware operation demanding payment. It is not a data theft ring selling credentials. It is not a DDoS-for-hire service. It is something simpler and, in its own way, more elegant: a parasitic energy arbitrage. The operator pays nothing for the CPU cycles, nothing for the electricity, nothing for the bandwidth. Every satoshi of revenue is pure profit.

$0
Operator's infrastructure cost
$3.50
Monthly C2 server cost
81+
Active mining nodes (observed)
8 years
Continuous operation

The entire operational overhead is one IRC server at $3.50/month. The rest โ€” the compromised servers doing the actual mining โ€” belongs to someone else. Their electricity bills rise by $20-80/month per machine. Their CPU capacity degrades. Their network bandwidth is consumed. And the operator collects the Monero.

Chapter 2: Why Monero

The mdrfckr operation mines exclusively Monero (XMR). Not Bitcoin. Not Ethereum. Not any of the hundreds of alternatives. This choice is not arbitrary โ€” it is the only rational choice for a parasitic mining operation.

The Technical Argument

RandomX Algorithm (Monero's Proof-of-Work):
  • CPU-optimized by design โ€” deliberately makes GPUs and ASICs inefficient
  • Every general-purpose CPU is viable mining hardware โ€” no specialized equipment needed
  • 4-core server CPU yields 500-1,000 H/s โ€” useful hashrate from even modest machines
  • Memory-hard โ€” requires 2GB RAM per mining thread, limiting density but ensuring broad hardware compatibility

Bitcoin mining requires ASICs costing $5,000-15,000 each. Ethereum moved to proof-of-stake. Every other major cryptocurrency requires specialized hardware. Only Monero can be profitably mined on the kind of hardware a botnet controls: general-purpose server CPUs.

The Privacy Argument

Monero Wallet Address (from XMRig config):
483fmPjXKgi9ZzCTMftKcXR4Xyph6B9BGjAnoJGBNHPnbVFe45x1jVm2rs5RanTPGxRzBjwxfaSA1R2sUiHWLFfAG39Ewwg

Can we track this wallet? No. Monero uses stealth addresses (one-time addresses per transaction), RingCT (ring confidential transactions that hide amounts), and Bulletproofs (range proofs that hide values). Without the private view key, this wallet address reveals nothing โ€” not the balance, not the transaction history, not the mining pool payouts.

This is precisely why every major cryptojacking operation uses Monero. It is the only cryptocurrency where:

  • The wallet balance cannot be queried by law enforcement
  • Pool payouts cannot be traced back to specific mining sessions
  • Exchange deposits cannot be linked to mining origin
  • The total revenue of the operation remains permanently unknown

๐Ÿ” The Privacy Paradox

Monero's privacy features were designed for legitimate use โ€” protecting financial privacy in an era of surveillance capitalism. The same features that prevent a government from tracking a dissident's donations also prevent anyone from calculating how much a botnet operator has earned.

This is not a bug in Monero's design. It is a feature that reveals an uncomfortable truth: privacy is indivisible. You cannot build a system that protects political dissidents but not criminals. The math doesn't know the difference.

The mdrfckr operator benefits from the same cryptography that protects journalists in authoritarian regimes. The cost of that protection is that we will never know how much they have earned.

Chapter 3: The XMRig Configuration

The mining software is XMRig โ€” the most widely used open-source Monero miner. Versions 6.19.0 (Kaspersky observation) and 6.22.1 (Elastic observation) have been documented. Both are UPX-packed ELF binaries with embedded configuration.

Configuration Analysis

ParameterValueSignificance
AlgorithmRandomX (Monero)CPU-only, no GPU needed
Process namekswapd0Mimics Linux kernel thread
BackgroundtrueRuns as daemon, no terminal attachment
OpenCL/CUDADisabledCPU-only โ€” avoids GPU detection artifacts
Donation0%Removes XMRig's built-in 1% dev fee
Pool ports80, 4444, 442, TorMultiple fallbacks, one over Tor
TLSMixed (442 = SSL)Note: port 442, not 443 โ€” avoids competing with IRC C2
Max threadsAll availableMaximizes hashrate at cost of host performance
NOTABLE
Port 442 vs 443: The miner connects to pools on port 442 (one below standard HTTPS). The IRC C2 operates on port 443. This deliberate separation prevents port conflicts on machines where both the miner and the Shellbot run simultaneously. It's a small engineering detail that reveals operational maturity โ€” the operator has encountered and solved the port collision problem.

The kswapd0 Masquerade

The XMRig binary is renamed to kswapd0 before execution. In Linux, kswapd0 is a legitimate kernel thread responsible for memory page swapping. The real kernel thread appears in process listings as [kswapd0] (with brackets, indicating kernel space). The malware appears as kswapd0 (without brackets, indicating user space).

A system administrator running top or htop will see a process named "kswapd0" consuming 100% of all CPU cores. Unless they know to check for the bracket distinction, they may assume it's a kernel memory management issue rather than a cryptominer. This trick has worked for eight years.

Chapter 4: The Mathematics of Parasitic Mining

Let us calculate what the mdrfckr operation actually earns.

Current Scale (81 Observed Nodes)

Input Variables: Nodes observed: 81 Average CPU cores per node: 4 (mix of VPS and dedicated) Hashrate per core (RandomX): 150-250 H/s Average per node: 700 H/s (conservative โ€” some nodes have 8-16 cores) Total Botnet Hashrate: 81 nodes ร— 700 H/s = 56,700 H/s โ‰ˆ 56.7 KH/s Monero Network Context (June 2026): Total network hashrate: ~3.2 GH/s (3,200,000 KH/s) Block reward: ~0.6 XMR per block (post-tail emission) Block time: 2 minutes Blocks per day: 720 Daily Revenue Calculation: Botnet share of network: 56.7 / 3,200,000 = 0.00177% Daily XMR mined: 720 blocks ร— 0.6 XMR ร— 0.0000177 = 0.0077 XMR/day At ~$160/XMR (June 2026): โ‰ˆ $1.23/day โš ๏ธ This assumes ONLY our observed 81 nodes. The actual botnet is larger.

Wait. $1.23 per day? For an operation running since 2018? Something doesn't add up.

The Missing Nodes

CRITICAL
Our honeypot sees only a fraction of the botnet. The 81 nodes that scanned our single honeypot IP in a 45-day window represent a sample, not a census. Kaspersky (2025) documented infections across multiple continents. Elastic (2025) documented separate incidents. The NINGI researcher operates a different honeypot seeing different nodes. The actual botnet size is estimated at 500-2,000 active nodes at current operational tempo.

Realistic Revenue Estimates

ScenarioNodesHashrateDaily XMRDaily USDMonthly USD
Our observation (minimum)8156.7 KH/s0.008$1.23$37
Conservative estimate500350 KH/s0.047$7.56$227
Moderate estimate1,000700 KH/s0.095$15.12$454
High estimate2,0001.4 MH/s0.189$30.24$907
Peak (2018, documented)180,000126 MH/s17.01$2,721$81,648

At moderate estimates: $15/day, $450/month. For Romania, where the average monthly salary is approximately โ‚ฌ800-1,200 ($850-1,280), this represents a 35-50% income supplement requiring zero ongoing labor beyond maintaining the C2 server.

Chapter 5: The Energy Theft

The operator pays nothing for electricity. The victims pay everything.

Per-Node Energy Cost

CPU Power Consumption at 100%: Typical server CPU (Xeon E5 / EPYC): 120-180W at full load Budget VPS (shared CPU, 4 cores allocated): ~60-80W allocated Average across botnet (mix): ~100W per node Monthly Electricity Cost (per infected node): 100W ร— 24h ร— 30 days = 72 kWh/month Average electricity rate (global): $0.12-0.25/kWh Cost to victim: $8.64 - $18.00/month per node Total Energy Theft (1,000 nodes): 72,000 kWh/month = 72 MWh/month $8,640 - $18,000/month in stolen electricity Equivalent COโ‚‚: ~30 tons/month (at global average 0.42 kg COโ‚‚/kWh)
KEY INSIGHT
The botnet steals more in electricity than it earns in Monero. At 1,000 nodes, it generates ~$450/month in XMR while stealing $8,640-$18,000/month in electricity. The operator captures less than 5% of the economic damage they inflict. The remaining 95% is pure deadweight loss โ€” value destroyed without benefit to anyone.

Chapter 6: The Pool Architecture

Mining solo (directly to the Monero blockchain) requires significant hashrate to find blocks consistently. At 700 KH/s, the botnet would wait weeks between block discoveries. Instead, miners connect to mining pools โ€” cooperatives that combine hashrate and distribute rewards proportionally.

Pool Connection Strategy

Documented pool connection ports in XMRig config:
  • Port 80 โ€” HTTP port, bypasses most firewalls, looks like web traffic
  • Port 4444 โ€” Standard mining pool port (Stratum protocol)
  • Port 442 โ€” TLS mining (one below HTTPS to avoid IRC C2 conflict)
  • Tor hidden service โ€” Fallback via .onion address, maximum privacy

The multi-port strategy ensures connectivity from almost any network environment. Corporate firewalls that block port 4444 still allow port 80. Networks that block Tor still allow port 442. The miner will find a path to the pool.

Pool Selection Intelligence

The specific pool addresses are embedded in the UPX-packed binary and require decompression to extract. Published analyses document connections to:

  • Major pools (MoneroOcean, Nanopool, SupportXMR) โ€” high reliability, higher fees (1-2%)
  • Tor-enabled pools โ€” sacrifice speed for privacy
  • Private/semi-private pools โ€” operator-controlled, zero fees, full privacy

๐Ÿ” The Pool Operator Question

The donation level is set to 0% โ€” meaning the operator removed XMRig's built-in 1% developer fee. This saves approximately $0.15/day at moderate scale. A trivial amount.

But it raises a question: does the operator run their own mining pool?

A private pool provides maximum privacy (no third-party sees your hashrate), eliminates fees entirely, and prevents pool operators from correlating wallet addresses with IP addresses. The technical skill required to run a basic Monero pool is moderate โ€” within the capability range of the mdrfckr operator profile.

If the operator runs their own pool, the privacy improvement is significant: no pool operator can be subpoenaed for payout records. The entire chain from CPU to wallet becomes invisible.

Chapter 7: The 2018 Peak โ€” 180,000 Nodes

In 2018, when Trend Micro first documented Outlaw, the botnet comprised an estimated 180,000 compromised hosts. Let us calculate what that was worth.

2018 Peak

  • Nodes: 180,000
  • Hashrate: ~126 MH/s
  • Network share: ~5-8%
  • XMR price (2018 avg): $120
  • Daily revenue: $3,000-8,000
  • Annual revenue: $1M-3M

2026 Present

  • Nodes: ~500-2,000 (estimated)
  • Hashrate: ~350 KH/s - 1.4 MH/s
  • Network share: ~0.01-0.04%
  • XMR price (2026): $160
  • Daily revenue: $7-30
  • Annual revenue: $2,700-11,000

What Happened Between 2018 and 2026?

The Decline:
  • RandomX adoption (2019) โ€” Changed mining algorithm, temporarily disrupted all miners including botnets
  • Increased network difficulty โ€” Legitimate mining operations grew, diluting botnet revenue share
  • Cloud provider automation โ€” AWS, Azure, GCP automated cryptominer detection (imperfect but reduced lifespan per node)
  • Competition โ€” TeamTNT, Kinsing, and other botnet families entered the space, fragmenting the victim pool
  • SSH key rotation โ€” Some enterprises finally implemented key management, reducing re-infection rates

The botnet shrunk 99% from peak โ€” 180,000 to ~1,000 nodes. Revenue dropped from $1M+/year to under $11K/year. And yet the operation continues. Why?

Chapter 8: The Zero-Cost Persistence Theorem

Traditional businesses fail when revenue drops below costs. The mdrfckr operation has costs approaching zero:

Fixed Monthly Costs: IRC C2 server (FranTech VPS): $3.50 Domain registration (amortized): $0.83 (~$10/year) Total fixed costs: $4.33/month Variable Costs: Per compromised host: $0.00 (victim pays electricity) Per mining session: $0.00 (victim pays bandwidth) Per XMR payout: $0.00 (pool handles distribution) Total variable costs: $0.00 Break-Even Calculation: Monthly revenue needed to sustain: $4.33 At $160/XMR, need: 0.027 XMR/month Hashrate needed: ~5 KH/s Break-even: approximately 7 infected nodes
CRITICAL INSIGHT
The mdrfckr botnet breaks even at 7 nodes. Any number of infected machines above seven generates pure profit. At 81 observed nodes, the operation is profitable. At 1,000 estimated nodes, it's very profitable relative to effort. At 180,000 peak nodes, it was a million-dollar operation.

This is why it never dies. There is no economic pressure to shut down. The operator can lose 99% of their botnet and still be profitable. The only thing that could make this operation unprofitable is losing ALL nodes โ€” and the SSH key ensures that never happens as long as even a few machines remain unpatched.

Chapter 9: The Victim's Bill

Who actually pays for the mdrfckr operation?

Direct Costs to Victims (Per Infected Host)

Cost CategoryMonthly ImpactWho Pays
Electricity (CPU at 100%)$8-18Host operator / datacenter
CPU capacity loss (performance degradation)$50-200 (opportunity cost)Host operator / end users
Bandwidth consumption (pool communication)$1-5ISP / host operator
Incident response (if detected)$500-5,000 (one-time)Organization security team
Potential data breach (SSH access persists)UnquantifiableOrganization / customers

Aggregate Damage (1,000 nodes, annual)

$103K-$216K
Electricity stolen annually
$600K-$2.4M
CPU opportunity cost
360 tons
COโ‚‚ emissions annually
$5K-$11K
Operator's actual revenue

The ratio is staggering. For every dollar the operator earns, victims collectively lose $50-200. This is perhaps the most economically wasteful form of crime in existence โ€” destruction-to-profit ratio worse than vandalism, worse than theft, worse than fraud.

Chapter 10: The Monero-to-Cash Pipeline

Mining Monero is one thing. Converting it to spendable currency is another.

Possible Off-Ramp Methods

Method 1: Direct Exchange (KYC)
  • Sell XMR on Kraken, Binance, or similar exchange
  • Requires identity verification (KYC/AML)
  • Leaves a paper trail connecting wallet to bank account
  • Risk: High โ€” regulatory compliance teams can flag mining-derived deposits
Method 2: P2P Trading (Pseudo-anonymous)
  • Sell XMR on LocalMonero (now defunct), Bisq, or Telegram channels
  • Receive fiat via cash deposit, gift cards, or mobile payment
  • No centralized KYC โ€” counterparty risk only
  • Risk: Low-Medium โ€” requires finding buyers at small scale
Method 3: Cross-chain Swap (No KYC)
  • Swap XMR โ†’ BTC via atomic swaps or services like TradeOgre
  • Then BTC โ†’ fiat through any of thousands of Bitcoin ATMs
  • Or BTC โ†’ goods (direct purchases)
  • Risk: Low โ€” atomic swaps leave no centralized record
Method 4: Never Cash Out
  • Hold XMR as savings (digital mattress money)
  • Spend only when needed, in small amounts
  • Value fluctuates but no conversion trail exists
  • Risk: Zero โ€” no off-ramp activity to detect

For a Romanian operator earning $15-30/day, the most likely approach is Method 2 (P2P trades for Romanian lei or euros) combined with Method 4 (holding as savings). The amounts are small enough to avoid triggering exchange compliance alerts and large enough to matter in the Romanian economy.

Chapter 11: The Long Game

If the operator has been mining since 2018, what is the cumulative total?

Estimated Cumulative Revenue (2018-2026): 2018 (peak, 180K nodes, XMR ~$120): ~126 MH/s, ~5% network share Est. 8-25 XMR/day ร— 365 = 2,920 - 9,125 XMR 2019-2020 (decline, RandomX transition, ~10K nodes): Est. 0.5-2 XMR/day ร— 730 = 365 - 1,460 XMR 2021-2023 (steady state, ~2-5K nodes): Est. 0.1-0.5 XMR/day ร— 1,095 = 110 - 548 XMR 2024-2026 (reduced, ~500-2K nodes): Est. 0.03-0.19 XMR/day ร— 900 = 27 - 171 XMR Total Estimated Lifetime Mining: 3,422 - 11,304 XMR Value at Current Price ($160/XMR): $547,520 - $1,808,640 Value if Held from 2018 Peak ($480/XMR, Jan 2018): Original mining value: $1,642,560 - $5,425,920 โš ๏ธ These are estimates. Actual amounts are unknowable due to Monero privacy.

๐Ÿ” The Half-Million Dollar Question

If the conservative estimate is correct โ€” somewhere between $500K and $1.8M in Monero mined over 8 years โ€” several questions arise:

  • Where is the money? In Romania, $500K buys multiple apartments outright. Has the operator accumulated visible wealth? Or are they holding XMR?
  • Tax implications: Romania's National Agency for Fiscal Administration (ANAF) taxes cryptocurrency gains at 10%. Has the operator ever declared any of this? Almost certainly not.
  • At what scale does law enforcement care? At $50/day, this is below enforcement thresholds. At $500K cumulative, it crosses into organized crime territory in any jurisdiction.
  • The holding question: If the operator mined 3,000+ XMR in 2018 and held it, they are sitting on a significant fortune that requires no conversion to have value (Monero is directly spendable). They could be wealthy without ever touching fiat currency.

Chapter 12: The Economic Immunity

The mdrfckr botnet has achieved something unusual in criminal economics: immunity from market forces.

What can't kill this operation:
  • Monero price crash โ€” At zero cost, any price above zero is profitable. Even at $1/XMR, the C2 server costs $4.33/month and 7 nodes cover it.
  • Competition โ€” The operator doesn't need exclusive access to hosts. Even if other botnets infect the same machines, the SSH key provides independent access.
  • Node loss โ€” Losing 99% of nodes still leaves enough for profitability. The self-propagating mechanism continuously acquires new hosts.
  • Mining difficulty increase โ€” Reduces revenue but never to zero at zero cost.
  • Individual takedowns โ€” Each node is independent. Removing one has zero effect on others.
What CAN kill this operation:
  • SSH key compromise โ€” If the private key leaks, anyone could access all nodes (and the operator loses control). But 8 years without key leak suggests good OPSEC on this specific artifact.
  • IRC C2 seizure โ€” Law enforcement seizing 198.98.56.205 would sever command capability. But the operator could rent a new VPS in hours and redeploy. The nodes would continue mining to the last known pool config.
  • Operator arrest โ€” The only permanent solution. Requires international cooperation between Romania, the US (C2 location), and victim countries.
  • Monero protocol change โ€” If Monero moved away from CPU mining (extremely unlikely given its design philosophy), the botnet would become useless.

The Perpetual Motion Machine

The mdrfckr botnet is the closest thing to a perpetual motion machine in criminal economics. It consumes no energy from the operator. It requires no ongoing investment. It repairs itself (self-propagation replaces lost nodes). It is resilient to partial destruction. It has survived 8 years of security industry attention.

The operator built this machine once, in 2018, and it has been running ever since. The daily effort required is close to zero โ€” check the IRC channel occasionally, maybe update a password scheme when researchers get too close. The machine does the rest.

FINAL ASSESSMENT
The mdrfckr operation will never stop voluntarily. There is no economic incentive to shut down. There is no technical forcing function. The operator cannot be outcompeted, cannot be bankrupted, and cannot be made unprofitable short of Monero ceasing to exist. The only termination scenarios are: arrest, death, or the operator's voluntary retirement with their accumulated XMR fortune.

Until one of those occurs, the parasite continues. The haiduc rides on.
Next in series: TI-2026-025E โ€” The Victims: Education Networks, Government Systems, and the Geography of Neglect
Previously: 025A โ€” The Anatomy | 025B โ€” The Infrastructure | 025C โ€” The Operator
โš  Personal capacity. Research published independently โ€” not reflecting employer views. Derived from passive observation of attacks against personal infrastructure. Full disclaimer โ†’
โ† Previous The SSH Parasite โ€” 9 / 14 Next โ†’