Renting hashrate into Lithos

This page covers renting Autolykos hashrate and pointing it at your own Lithos client: whether it pays, how to set up the client and the rental order, how large to make it, which difficulty commitment to make, and how to run and stop it. It does not explain how Lithos pays miners in general, which is in Mining on Lithos, or how forwards work, which is in Difficulty forwards.

Deciding whether renting pays#

Renting hashrate into Lithos pays at these prices. The following figures are for 25 GH/s rented at 0.0153 BTC/TH/day, with 5 GH/s already mining on Lithos:

That nets $11.62 a day before LIT, or $25.35 with LIT. With 35 GH/s already in Lithos, the same rental earns $37.96 a day gross before LIT ($51.75 with LIT) and nets $7.83 ($21.62 with LIT).

Four facts decide the strategy:

• Rented hashrate costs more than it earns on a normal pool. NiceHash bids run at about 98% of pool revenue before the 3% fee, so renting into a pool at best breaks even.
• The profit is the Lithos premium. Each Lithos block carries storage rent, about 1.9 ERG ($0.54) per rented GH/s per day, and 480 LIT, about 2.1 ERG ($0.57) per GH/s per day at spot.
• The premium is shared. Storage rent is about 214 ERG a day for the whole network. Every GH/s added to Lithos, yours or anyone else's, shrinks everyone's share of it.
• A difficulty commitment fixes your size for about 28 hours, never your duration. Stopping costs only the income you skip.

The strategy takes five steps:

1. Price the rental against ERG-only revenue, and treat LIT as upside.
2. Size the rental to the hashrate already in Lithos (25 to 30 GH/s today), not to what NiceHash can supply.
3. Commit for 14 to 21 super shares per window, depending on your share of Lithos's committed score, and hold delivered speed at that target.
4. Stop the order when the margin turns negative, and resize once per commitment cycle.
5. Hedge difficulty with forwards only for fixed-term rentals or large positions.

Understanding where the profit comes from#

A Lithos block pays miners about the same ERG as a pool block, plus storage rent and LIT. Rental prices track pool revenue, so rent and LIT are the margin. The following table gives gross revenue per rented GH/s per day:

Against that, a NiceHash rental at 0.0150 BTC/TH/day plus the 3% buyer fee costs 4.25 ERG ($1.19) per GH/s per day. The rent row assumes 25 GH/s rented on top of 25 GH/s already in Lithos. The other rows hold at any size while network hashrate stays the same.

Every row comes from one rate: each GH/s finds 720 ÷ network GH/s blocks a day. At the 7-day mean difficulty of 65.3T the network runs 544 GH/s, which gives 1.322 blocks a day. The three Lithos rows are:

base  = 1.322 × 2.93
LIT   = 1.322 × 480 × P_LIT
rent  = R ÷ (H_L + h + H_o)

The symbols mean the following:

Three notes on the table:

• Everything in a Lithos block's rewards is split by committed score. That covers collateral, rent, batching fees and LIT, so the commitment cushion in Choosing your difficulty commitment scales all three Lithos rows.
• Large pools earn little beyond the block reward. The two largest Ergo pools collect no storage rent, so pool revenue stays close to the 3 ERG reward.
• Renting into a normal pool loses money. It returns $1.10 against $1.19 of rental per GH/s per day.

Choosing where to rent#

Rent on NiceHash↗. It is the only liquid Autolykos market, with about 98 GH/s delivered on 2 November 2026, about 18% of Ergo's hashrate. The following table compares the two markets that list Autolykos:

MiningRigRentals' last 10 rentals averaged 0.000055 BTC/GH/day. Check it again if asks fall toward 0.0000143 BTC/GH/day, which is about pool revenue.

To convert a NiceHash price to ERG per GH/s per day:

1. Divide the BTC/TH/day price by 1,000.
2. Divide the result by the BTC price of one ERG.
3. Add 3% for the buyer fee.

At 3.64 × 10⁻⁶ BTC per ERG, 0.0150 BTC/TH/day is 4.25 ERG ($1.19).

NiceHash rules that shape the strategy#

The following NiceHash rules decide how you bid, resize and stop:

Setting up the client for a rental#

A rental needs a Lithos client that NiceHash can reach, a registered difficulty commitment, and a one-day test order before you scale up. The stratum settings change from their shipped defaults as follows:

stratum {
  bindAddress = "0.0.0.0"
  stratumPort = 4444
  reduceShareMessages = true
  reductionMultiplier = 100
  niceHashMode = true
  diff = "947.0M"
}

If Lithos Launcher writes your lithos.conf, set the difficulty and the stratum port in the launcher. Add only bindAddress, reduceShareMessages, reductionMultiplier and niceHashMode below its block, because a value written there overrides the launcher's and later launcher changes to it stop taking effect.

To set up the client and the order:

1. Run the latest Lithos client against your own synced, mining-enabled Ergo node. Leave storage rent collection on: stratum.candidate.sources.rent.enabled = true, the default. Rent is about half of the Lithos premium.
2. Open the stratum to the internet. Use a static IP address or a DNS name, and open stratum.stratumPort in your firewall and router.
3. Set the share difficulty. Apply the reductionMultiplier, niceHashMode and diff values from the table above. Connect no other mining software to this stratum while niceHashMode is on.
4. Fund the miner wallet with at least 5 ERG. Each Non-Interactive Share Proof (NISP) posts a refundable bond of score ÷ 25 nanoERG, which is 0.038 ERG at a 950M commitment, held for about 24 hours until its rollup pays. Each rollup also costs 0.002 ERG in fees.
5. Register and make your difficulty commitment. Use the web panel at Mining > Difficulty > Commitment, or send POST /mining/commitment with your API key. Registration locks 0.001 ERG and pays a 0.001 ERG fee. Commitments from the API need state.autoCommit = false, the default.
6. Add the pool on NiceHash. Choose the Autolykos algorithm, and use stratum+tcp://YOUR_HOST:4444 with any worker name and password. Lithos pays the Lithos client's wallet, not the worker. Check the pool with NiceHash's pool verifier, which needs a NiceHash account.
7. Run a test order for one commitment cycle. Commit for 5 GH/s at 17 super shares per window (212M) and order 5 GH/s with the 0.001 BTC minimum.

Over a day, the test should show the following:

• About 8.5 super shares an hour recorded by the Lithos client.
• A NISP submitted to nearly every rollup.
• A client hashrate reading that matches NiceHash's accepted speed.

A stratum difficulty mismatch can leave super shares unrecorded without any error. This test catches it cheaply. The test commitment then holds for about 28 hours before you can raise it to full size.

Sizing the rental#

Size the rental to the hashrate already in Lithos. With 5 GH/s in Lithos, profit before LIT rises to about 45 ERG ($12) a day at 30 GH/s and barely moves after that. With 35 GH/s in Lithos, profit peaks at about 29 ERG ($8) a day at 30 GH/s.

The following table gives daily rental cost against daily gross revenue, your share of each block's rewards plus LIT where shown, at the best cushion for each size:

At every size and both Lithos levels, gross revenue beats the rental cost, even before LIT. The following table gives what you keep, gross revenue minus rental cost, per day:

LIT at spot adds about 2.0 ERG ($0.55) per rented GH/s per day. Both tables assume the rental price rises with size, at 0.0140 + 0.00005 × GH/s BTC/TH/day. Profit per day at your inputs is:

profit(h) = hit × S ÷ (S + O) × [ 1.322 × (H_L + h) × (2.93 + 480 × P_LIT)
                                 + R × (H_L + h) ÷ (H_L + h + H_o) ]
            − c(h) × h

Here hit is your NISP hit rate, S your committed score, O the other miners' committed score and c(h) your rental cost per GH/s per day. Set P_LIT to zero for profit before LIT. Four things follow from the curve:

• Past about 30 GH/s the curve flattens. Each extra GH/s outbids more of the order book and earns a smaller share of a fixed rent supply.
• The low end of the flat range is the better choice. 30 GH/s nets $12.49 a day against $13.10 at 40 GH/s, costs $14 a day less to rent, and is less exposed to a price spike or a new renter.
• Break-even rental price at 25 GH/s is 0.0211 BTC/TH/day before LIT with 5 GH/s in Lithos, or 0.0192 with 35 GH/s. With LIT at spot it is 0.0281 and 0.0262.
• NiceHash's whole Autolykos supply is about 98 GH/s, so 30 GH/s is about a third of it.

Choosing your difficulty commitment#

Commit for 14 to 21 super shares per NISP window, depending on how much of Lithos's committed score is yours. A missed NISP only hands one rollup's rewards to the other miners, so a smaller cushion pays more unless you hold most of Lithos's score.

A NISP needs 10 super shares from the 60 blocks before its rollup, the NISP window. To average n super shares per 120 minutes, commit the following score, with h in H/s:

S = 7200 × h ÷ (10,000 × n)
 
25 GH/s at n = 19  →  S = 947M

A change to your difficulty commitment takes effect on the following schedule, counted from the block you send it at:

To change size within that schedule:

• Raise: send the change, then have the order at the new speed by block +63. Shares mined before the stratum switches do not count at the higher score.
• Cut: send the change and hold speed until block +125. Shares mined at the higher threshold still count at the lower one.
• Stop: stop at any time, with no penalty. Rollups whose window you covered still pay. Restarting at the same size needs no new commitment.

Pick n from your share of committed Lithos score. You are paid hit rate × S ÷ (S + others' score). A bigger score helps only while others hold most of the score, and a miss hurts most when you hold most of it. The following table gives the best n for each share:

Hit rates come from the NISP windows of the last 20,000 mainnet blocks, with 10% noise added to delivered hashrate. The middle 90% of windows ran 94 to 148 minutes.

Staggering commitments across tranches#

One Lithos client holds one miner identity and one difficulty commitment. Two or three Lithos clients on separate ports, each with its own wallet and NiceHash order, let you stagger their commitment cycles and resize one tranche every 9 to 14 hours. Each tranche has the following costs and limits:

• Each identity pays 0.002 ERG to register, 0.002 ERG per rollup in fees, and its own bonds.
• Tranches spread delivery problems, not short windows. Every tranche misses the same fast windows.

Running the order#

Most of the running work is keeping delivered speed at your commitment's target and stopping when the margin goes. Commitment changes happen once a cycle. A controller script does the following every 5 minutes:

1. Read delivered speed from the NiceHash order and from the Lithos client's share rate. The target is S × 10,000 × n ÷ 7200 H/s.
2. Raise the price by 0.0001 BTC/TH/day when speed is more than 5% below target on two reads. Raises take effect at once, and hashrate arrives over 20 to 40 minutes.
3. Lower the speed limit, not the price, when speed is more than 10% above target, since price cuts are throttled. Extra hashrate finds more Lithos blocks but does not raise your score, so you earn only your share of the extra rewards.
4. Check the margin with the sizing formula and live inputs. If it stays negative for an hour, drop the limit to the 3 GH/s minimum, wait for the rigs to leave, then cancel.
5. Keep the order funded. Refills are free. An empty order stops delivery, and every NISP after that misses.

Once a cycle, about every 28 hours or once per tranche, rerun the sizing and send the next difficulty commitment, following the raise and cut timing above. The controller reads its inputs from the following places:

The following alarms are worth wiring:

• Super shares per hour below 90% of n ÷ 2 for two hours: a delivery or stratum problem.
• A missed NISP in a window over 110 minutes: a delivery problem, not a short window.
• A jump in other miners' committed score: someone else added hashrate. Resize before your next cycle.

A rollup pays 720 blocks (about 24 hours) after it opens, but NiceHash is prepaid in BTC. Keep about two days of rental cost as float.

Planning around storage rent#

Rent is the dependable part of the premium, and it is flat from day to day: 196 to 233 ERG comes due in each of the next 14 days. Time rentals by the month, not the day.

A box becomes collectable once it is 1,051,200 blocks (about 4 years) old, counted from its creation height. The next block that collects it takes the rent. In a Lithos block, the Lithos client folds the rent into the holding box, so it is split by committed score like the rest of the block's rewards.

Lithos's share of rent is s ÷ (s + o), where s is Lithos's share of blocks and o is the block share of other pools that collect rent. Fitted on mainnet blocks of 25 to 28 October 2026, o is about 0.11. Lithos found too few blocks after that for a newer fit. That gives the following:

• Lithos takes 8% of rent at 1% of blocks, 31% at 5%, 48% at 10% and 58% at 15%.
• More Lithos hashrate takes rent from other pools, but less per block. Adding 25 GH/s to 25 GH/s lifts Lithos's rent from 62.9 to 97.2 ERG a day. The new hashrate earns 48.6 ERG a day of it: 34.3 from other pools and 14.3 from existing Lithos miners.
• No hours are worth aiming for. Over the next 14 days, the busiest 5% of hours hold just 8% of the rent, 19.0 ERG in the largest hour against 8.8 ERG in the median. A 28-hour commitment cycle could not target hours anyway.

The months matter. Forecast from unspent boxes by creation height, rent coming due runs as follows:

• About 214 ERG a day now.
• About 57 to 94 ERG a day for boxes turning collectable from late 2027 to late 2028.
• About 160 to 397 ERG a day in 2029 and 2030.

These are upper bounds, since some boxes will be spent before they age in. The following table gives the best size and its profit at each rent level, using today's rental prices and the rising price curve from Sizing the rental:

Batching fees add under 1 ERG a day across the whole network, so they do not change the table.

Hedging difficulty with forwards#

Forwards matter least of a renter's tools, and the market cannot yet hold a renter-sized hedge. On 2 November 2026 it held 5,000 LIT of offers, against about 51,000 LIT needed to cover one cycle at 25 GH/s.

A difficulty forward is a LIT-margined bet on the average difficulty over 1 to 120 epochs of 128 blocks, about 4.3 hours each. The window starts 2 to 120 epochs after the fill. The long side receives notional × clamp(window mean ÷ strike − 1, ±20%). Each side posts 20% of notional as margin, plus 7.5 LIT in fees.

A renter needs a hedge less than a rig owner, for the following reasons:

• You can stop at any time, so difficulty risk lasts at most one commitment cycle, about 6.6 epochs.
• NiceHash prices follow pool revenue, which falls when difficulty rises.
• Rent, 25% to 40% of your Lithos revenue, does not depend on difficulty.
• Rented NiceHash hashrate is pulled from other pools, so your own rental does not raise difficulty.

A hedge helps when your margin before LIT is under about 15% and you are about to commit a cycle, or when the position is large. A 10% rise in difficulty cuts base and LIT revenue by about 9%. To size a hedge:

1. Go long, with notional equal to the base and LIT revenue the window covers, in LIT. At 25 GH/s with 5 GH/s in Lithos, that is about 142 ERG ($40) a day, or about 51,000 LIT per 28-hour cycle at 0.003237 ERG.
2. Post 20% of notional as margin, about 10,200 LIT (about 33 ERG, $9).
3. Cover the epochs of your next cycle. Windows start at least 2 epochs out, so post the offer when you send the commitment change.
4. Strike at the forecast window mean, from realised hashrate rather than recorded difficulty. Dips have reverted with a half-life of about 3 epochs.

A hedge leaves the following open:

• The ±20% clamp binds in 1.5% to 5.5% of windows.
• The payout is in LIT.
• A fair hedge has zero expected value. It reduces variance and adds no profit.
• On 2 November 2026 the book held one long and one short offer, 2,500 LIT each. Filling the short covers 2,500 LIT of a long hedge if its window fits. For the rest, a renter would have to post long offers and wait for takers.

Managing the risks#

Two risks end the strategy: other renters diluting the premium, and storage rent falling in late 2027. The controller, the cushion and the stop rule handle the rest. The following table lists each risk:

Working through an example#

With 5 GH/s in Lithos, renting 25 GH/s nets about 42 ERG ($12) a day before LIT, or about 91 ERG ($25) a day with LIT at spot. The position is the following:

The following table gives what you pay and what you earn per day:

What you keep is gross revenue minus rental cost:

• Before LIT: 42 ERG ($11.62) a day, about $350 a month.
• With LIT at spot: 91 ERG ($25.35) a day, about $760 a month.

Checking before you rent#

Run the following checklist before scaling up a rental:

• Latest Lithos client, rent collection on, Ergo node synced and mining-enabled
• Stratum reachable from the internet, reductionMultiplier = 100, niceHashMode = true, stratum.diff set to the planned commitment
• Wallet funded with at least 5 ERG and registered
• One-cycle test at 212M and 5 GH/s showing about 8.5 super shares an hour and a NISP in nearly every rollup
• Rental sized from Lithos's current hashrate and the rent coming due
• Difficulty commitment at the n your share of score calls for
• Controller running speed, price, funding and the stop rule
• Alarms on the super-share rate, misses in long windows, and other miners' committed score
• Sizing rerun every cycle, and the rent forecast checked monthly
• Forwards hedge only if the margin before LIT is under 15% or the position is large

How the figures were measured#

Every chain figure was read from an Ergo mainnet node with the extra index, at height 1,908,274 on 2 November 2026. The following inputs went into it:

• Lithos blocks and rewards: 762 Lithos blocks. A block's rewards are its holding box at genesis height, and the top-up is its value above the base.
• NISPs: every submission since height 1,889,000, decoded for score and super-share timestamps.
• Rent coming due: unspent boxes created at heights 857,075 to 867,154. Rent per box is the smaller of its value and 1,250,000 nanoERG per byte.
• Network difficulty: the mean of the 5,014 blocks in the 7 days to 2 November 2026, 65.3T. The difficulty at the tip was 69.3T.
• Window lengths: the last 20,000 headers.
• LIT price: 1,800 states of the LithosDex ERG/LIT pool box.

The model behind the numbers uses the following:

• Blocks per day per GH/s = 720 ÷ network GH/s, with network GH/s = 7-day mean difficulty ÷ 120 seconds.
• Lithos rent share = s ÷ (s + o), fitted per 3-day window.
• Your pay = hit rate × S ÷ (S + others' score), with other miners at 17 super shares and no misses.
• Hit rate = the probability of at least 10 super shares over the measured windows, with 10% log-normal noise in delivered speed.
• Rental price = 0.0140 + 0.00005 × GH/s BTC/TH/day.
• LIT at spot, with no slippage.
• Rental costs converted from BTC, and revenue from ERG and LIT, at the spot prices at the top of this page.

Market data, all read on 2 November 2026, came from the following sources:

• NiceHash Autolykos order book↗
• NiceHash algorithm limits↗
• NiceHash buy info↗
• MiningRigRentals algorithms↗
• CoinGecko ERG and BTC prices↗
• NiceHash fee table↗
• NiceHash: before placing an order↗
• NiceHash pool operators↗ and supported pools↗

Next steps#

The following pages cover what this one builds on:

• Stratum configuration: every stratum setting this page changes.
• Candidate configuration: the storage rent source and its limits.
• Difficulty forwards: offering, filling and settling a hedge.
• Failure modes: the log lines to check when super shares stop arriving.