Hedging the Block: A Hashrate Derivatives Primer

A walk through what a hashrate derivative is, why miners need one, why earlier attempts stalled, and how the broader “blockspace as an asset class” thesis fits together. Drawn primarily from Alkimiya’s published research.

1. The miner’s problem, in one sentence

A miner who runs an industrial-scale operation has access to a deep, liquid futures market for the price of the coin she earns. CME Bitcoin futures, BitMEX perps, ETH options on Deribit — pick your venue. What she does not have is a market for the quantity she’ll earn.

Her share of network rewards is (her hashrate ÷ global hashrate) × (subsidy + fees), and three of those four numbers move on her without warning. By the time the BTC she’s hedging actually shows up in her pool wallet, the amount may be 30% smaller than she modeled — because difficulty went up, or because fees went down, or because a new ASIC generation pushed network hashrate higher.

Hedging the price of a quantity you can’t lock in is half a hedge. A hashrate derivative is the instrument that closes the other half.

2. The producer’s P&L, decomposed

The canonical identity for a PoW miner’s profit:

Mining Profit = (Block Reward + Fees) × Price × (miner_hashrate / global_hashrate)
                − (Electricity Expense + Hardware Depreciation)

Five inputs. Three of them — Block Reward, Price, global_hashrate — are macro variables the miner doesn’t control. Two of them — Electricity and Hardware Depreciation — are roughly fixed once she’s signed her power contract and bought her rigs. The only true control variable left is miner_hashrate, and that’s bounded by the rigs she already owns.

Alkimiya frames the macro variables as a “three-body problem” of ungrounded forces:¹

• The emission schedule — pre-programmed halvings, predictable on the calendar but devastating on the cash-flow model.
• The climate cycle — Sichuan rainy seasons and Texas heat domes set the industry-average electricity cost.
• The hardware iteration — every two years, the next-gen ASIC redraws the hashrate-vs-watt frontier, and yesterday’s rigs become tomorrow’s depreciation.

Crucially, hashpower is not fungible across time: every unit of hashpower is unique, and one TH/s today is produced at a different cost basis than a TH/s yesterday.¹ You can’t carry a TH/s forward at a constant cost the way you can a barrel of oil. That non-fungibility across time is exactly what a forward market is supposed to handle, and exactly why the absence of one is so painful for miners running 36-month financing on hardware whose useful life is 24 months and whose revenue is reset every difficulty epoch.

3. How a hashrate forward actually works

Strip the design back to its bones and a hashrate derivative is an energy swap, transposed onto blockspace. Alkimiya describes the construction directly:

“It mirrors the energy swaps found in traditional commodity markets: a buyer pays a producer (in this case, miners/network validators) a fixed payment of stable coins that vests over the duration of the contract while the producer pays all rewards from validation according to a certain index over the same time period.”²

Two legs:

• Fixed leg. Buyer pays the producer a known stablecoin amount upfront (or vesting over the tenor).
• Floating leg. Producer streams the buyer all rewards attributable to a contracted slice of hashrate over the same window — block subsidy plus transaction fees, in proportion to (contracted_hashrate / global_hashrate).

A worked example: an ETH miner sells 1 TH/s of rewards on a 15-day contract. If global hashrate sits around 1,000 TH/s over the period, the miner is on the hook to deliver roughly 0.1% of network block rewards over those 15 days — before pool fees, randomness, and contract costs.²

A well-designed hashrate hedge has to do two things simultaneously:²

1. Isolate network activity risk — fees, congestion, demand for inclusion.
2. Isolate network asset price risk — the BTC/ETH spot price.

This is why staked-asset derivatives — say, a stETH option — make a poor hashrate hedge: their parity to spot is too tight to decouple revenue from coin price. A swap on the index (hashes-per-second over time, or fees-per-byte over time) is the right abstraction because it lets you pick exactly which exposures you want to keep and which you want to shed.

The general form of the index is “validation power over a unit of indexed time.”² PoW: hashes per second. PoS: stake per epoch. Storage networks: GB-seconds. ZK proving networks: proofs per epoch. The construction is consensus-mechanism-agnostic — and that generalization is the seed of the broader thesis we’ll come back to.

4. The graveyard of earlier attempts

The market has tried to build this instrument before. Most of those tries are dead. Alkimiya’s framing of why is clinical:

“Hashrate indices/futures and gas tokens failed to attain significant liquidity. These markets were often opaque and onerous to both the seller and buyer. In other cases, the demand was never realized as difficulties in pricing assets like gas tokens made it confusing to liquidity provision on-chain and on centralized exchanges.”²

The recurring failure pattern: most issuers of cloud-mining contracts and “hashrate tokens” ended up purchasing their own machines and absorbing the damage when the market turned. The intermediary became the de-facto counterparty, warehoused the basis risk, and blew up.

A quick history:

• 2014–2018: Cloud mining contracts. Genesis Mining, Hashflare, and the rest. Forward-sold hashrate to retail, often outright frauds, and at best opaque discretionary contracts where the issuer ate every adverse move.
• January 2020: BitOoda. Brokered the first regulated hashrate swap in the US, between Coinmint and an undisclosed buyer.³ OTC, not exchange-traded.
• May 2020: FTX hashrate futures. Settled against the average network difficulty over a fixed period. Had a real spec. Liquidity was thin. Died with FTX in November 2022.
• Late 2022 onward: Luxor Hashrate Forwards. OTC, settled against Luxor’s Hashprice Index, gated on Luxor pool participation. Per Luxor’s own materials, this has now done $500mm+ in cumulative OTC volume.⁴
• May 2024: Luxor × Bitnomial. First CFTC-regulated Bitcoin mining derivatives in the US. 1 PH contract size, monthly tenor, daily mark-to-market, USD-quoted.⁵

The recurring failure modes:

1. Centralized issuer warehouses risk and blows up (cloud mining).
2. Liquidity never bootstraps because pricing the index is hard and order books are thin (FTX hashrate futures, gas tokens).
3. Physical settlement is operationally painful — you have to deliver actual hashrate, which means you have to be a pool, which means the only viable issuer is a vertically integrated mining outfit (Luxor’s structural moat, also its bottleneck).
4. Venue collapse (FTX, again).

The interesting question is which of these are fundamental to the instrument and which are accidents of who built the first version. The on-chain bet — Alkimiya’s bet — is that they’re mostly accidents: that a well-designed protocol can sidestep all four. The competing bet, from regulated venues, is that the bottleneck is regulatory and settlement infrastructure, not protocol design. Both can be right for different customer segments.

5. The shape of the on-chain instrument

Without prescribing a specific implementation, the on-chain version of this instrument has a few invariants worth naming:

• The underlying is a network index, not a coin. What gets settled is a quantitative reading of the network — hashes per second, fees per byte, difficulty, baseFee, stake per epoch — averaged or sampled over a defined tenor. The instrument is consensus-mechanism-agnostic by construction.
• The oracle is the protocol’s hardest single problem. Whoever publishes the index reading at settlement time controls the contract outcome. Designs vary — federated oracle networks pulling from pools, on-chain commitments to chain state, transparent indices computed deterministically from chain data — but every on-chain hashrate derivative lives or dies on whether its oracle is credibly neutral.
• Settlement need not move physical hashrate. A miner doesn’t need to “deliver” hashrate. She only needs to deliver, or accept, the cash flow that hashrate produces — and that cash flow is observable from chain data or pool data without anyone moving hardware.
• Settlement style is a customer-segmentation choice. Continuous mark-to-market (the perp-DEX shape) suits speculators who enter and exit. Period-averaged settlement (the utility-bill shape) suits service providers who want a predictable monthly cost on inputs they consume. They are complementary instruments, not competitors.

The customer surface is also broader than is sometimes assumed. Miners hedging revenue is the obvious case. Less obvious, and increasingly important, is the demand side: paymasters, wallets, bridges, L2 batchers, exchanges, and market-makers who eat variable on-chain fees and would prefer to quote fixed costs to their users.⁶ As Alkimiya frames it, on-chain services that frequently settle on a base layer have the same hedging need as a utility company budgeting against energy prices: they want their monthly fee bill to be predictable, regardless of what happens intra-period to the spot fee market.

The post-Ordinals fee regime made this real. BTC transaction fees have spiked 20× to 500× during mint events.⁶ Anyone whose business model assumed “BTC fees are roughly free” — and many bridges and L2 batchers were built on exactly that assumption — has discovered they have a fee-volatility problem they didn’t know about, and a market for hedging it didn’t exist.

6. The thesis: blockspace as an asset class

Hashrate derivatives are interesting on their own. The bigger bet — the one Alkimiya has been making since its earliest writing — is that they’re the first instrument in a much larger asset class.

The argument runs through every commodity that’s ever industrialized:

“Using rice tickets as a basis, the merchants developed many derivative contracts that are now commonly traded today, such as short sales, forwards, and options… the volume of these financial contracts is much larger than the volume of the physical commodities.”²

The 1697 Dōjima Rice Exchange begat futures. The Chicago Board of Trade (1848) begat the modern grain complex. In every case the paper market eventually overtook the physical market in volume by an order of magnitude or more. Alkimiya tightens this into a slogan: “what gets priced gets managed.”⁶

Blockspace, in this telling, is the foundational commodity of the on-chain economy. Every transaction demands it; every miner and validator supplies it. It has a clearing price (gas, in lots of small auctions). It has time value: blockspace tomorrow is worth less than blockspace today, because tomorrow’s NFT might be sold out and tomorrow’s interest rate might be lower.² All the conditions for a paper market are present. None of the paper market exists yet at scale.

And the construction generalizes. The same swap structure works across every consensus mechanism, with the index swapped out:

• PoW: hashes per second → hashrate forwards.
• PoS: stake per epoch → staking-yield swaps.
• Storage (Filecoin, Arweave): GB-seconds → storage forwards.
• ZK proving networks: proofs per epoch → proving-power forwards.

Alkimiya’s umbrella term for this is “consensus capital markets” — every consensus mechanism produces a quantitative index, and every quantitative index can be the underlying of a forward.²

The closing argument: two trends will shape blockspace production over the next decade — industrialization and financialization.¹ Industrialization mostly happened between 2017 and 2022, as mining went from hobbyist to vertically integrated firms with power contracts, ASIC supply lines, and balance sheets to defend. The next leg is financialization. The question is whether the on-chain primitive or the regulated-venue primitive captures the liquidity flywheel first — and whether the demand side (services hedging fee bills) ends up driving more volume than the supply side (miners hedging revenue), which would invert the conventional sequencing.

7. What to actually take away

A few sharper claims that fall out of the above:

1. A hashrate derivative is a swap on validation_power × time. Not on coin price, not on difficulty alone, not on miner equity. Get the index right and the rest of the design follows.
2. The two real failure modes are issuer-warehouses-risk and oracle-is-fragile. The first kills cloud mining. The second is what dragged FTX hashrate futures into thinness. On-chain bilateral settlement with a credibly neutral oracle addresses both — but the oracle problem is genuinely hard, and “credibly neutral” is doing a lot of work in that sentence.
3. Settlement style is a customer-segmentation choice. Period-averaged pools serve utilities. Continuous mark-to-market serves speculators. They’re complementary instruments, not competitors, and a mature market needs both.
4. The demand side may surprise on the upside. The live customers may not be miners hedging revenue but paymasters, L2 operators, and bridges hedging fee bills. That inverts the conventional “build the producer hedge first” sequencing.
5. The thesis to disagree with, if you want to disagree: that blockspace is enough like a physical commodity to grow a paper market 10× its size. The case for: it’s a metered, exhaustible, time-stamped resource with a clearing price. The case against: subsidies expire, demand is bursty and event-driven, and the underlying chains keep changing in ways no commodity ever did.

The instrument is real, the math is straightforward, and the historical pattern (Dōjima → CBOT → every commodity since) suggests this asset class will exist eventually. The implementation that wins is still wide open. The interesting design space is in the oracle, the settlement cadence, and which side of the market — supply or demand — ends up writing the marginal contract.

Footnotes