Watt It Costs: The Brutal Economics of Post-Halving Bitcoin Mining

• Every halving cuts the block subsidy by 50%. Miners without low-cost power and modern hardware do not survive long.
• Electricity runs 60 to 80% of operating cost. A few cents per kWh decides whether a rig keeps running.
• ASIC efficiency has improved roughly seven times since 2017. Anything above 25 J/TH is a liability in most markets.
• Survivors own or hedge their power supply, run modern hardware, and keep balance sheets clean. Bankrupt miners over-leveraged and ran old rigs on expensive grids.
• Most cloud mining services are Ponzi schemes. Guaranteed returns in mining are not a real thing.

Bitcoin Mining After the Halving: Cheap Power Wins, Everything Else Folds

Every four years, Bitcoin cuts the block subsidy in half. It is written into the protocol. It does not care about hash rate, debt loads, electricity contracts, or what the price did yesterday. Block 210,000 cuts the reward, then 420,000, then 630,000. We have now had four of these. The next one lands sometime in 2028.

Miners hate the halving in private and pretend to love it in public. The reason is straightforward. The day before, every block pays 6.25 BTC. The day after, it pays 3.125. Costs stay the same. Hardware stays the same. Electricity bills stay the same. Revenue gets cut in half overnight unless the price does the heavy lifting.

The future of finance is here. Trade crypto and more

The rest of this is about how the survivors navigate that math, why electricity is the entire game, what hardware works post-halving, and which operators have come through the last few cycles versus the ones that got carried out.

The Halving Without the Mythology

Bitcoin’s supply schedule is a fixed function. Every 210,000 blocks, roughly every four years, the subsidy halves. We have gone from 50 BTC per block in 2009 to 25 in 2012, 12.5 in 2016, 6.25 in 2020, and 3.125 in April 2024. Eventually the subsidy reaches zero and miners live entirely on transaction fees. That endgame was baked in from day one.

Each halving is a forced revenue cut. The network smooths some of it through the difficulty adjustment. If enough miners turn off after a halving, blocks come in slower, difficulty drops, and the remaining miners earn more per unit of hash. After the 2020 halving, hash rate dipped briefly before recovering as inefficient rigs went dark and newer ones came online.

Fees become more important each cycle. They used to be a rounding error in miner revenue. By late 2023, fees were around 20% of miner income, the highest share since 2017. Post-2024 halving, Coinbase Research warned that revenue per terahash was heading toward all-time lows unless the price filled the gap. Inscriptions, Runes, and other on-chain activity have made fee revenue more meaningful, but also more volatile.

Most cycles see a delayed rally that pulls miner economics back into the green. The months between the halving and the rally are where leveraged miners go broke.

Why Electricity Is the Whole Story

Power runs 60 to 80% of a miner’s operating cost. Cambridge’s mining index assumes a global average around $0.05 per kWh, which sounds low until you remember that some operations pay a fraction of that and others pay four times as much. The reality runs from a couple of cents in hydro-heavy regions to over 20 cents on retail rates.

At $0.10 per kWh, a 3,000-watt miner costs around $7.20 per day in electricity, or roughly $2,628 per year. Cut that to $0.05 and you save $1,300 a year per machine. Multiply by a few thousand rigs and you understand why miners chase cheap power the way airlines chase fuel hedges. The same machine in two different power markets can be a profitable asset in one and a liability in the other.

Buy bitcoin and stocks. Signup to Coinbase today

A rough breakeven map looks like this:

Grid prices are not fixed. Demand charges, peak vs off-peak pricing, regulatory shifts, and seasonal spikes can flip a profitable site into a loss-making one within a week. The Texas freeze in February 2021 forced miners offline during a price spike. The summer of 2022 saw Riot Blockchain earn $9.5 million in utility rebates for shutting down voluntarily, which exceeded its actual mining revenue that month. The grid paying you more to not mine than mining pays you is not a glitch. It is the future of grid-balancing economics, and miners are positioned to benefit from it when they sign the right contracts.

Hardware Has Improved Faster Than People Realize

In 2017, the Antminer S9 was the workhorse. 14 TH/s at around 1,320 watts, which works out to roughly 98 joules per terahash. That was state of the art at the time.

By 2024 and 2025, top-tier ASICs from Bitmain and MicroBT had pushed efficiency closer to 15 joules per terahash on their most efficient units, with hashrate per machine running into the hundreds of TH/s. The exact efficiency figures depend on which generation and which manufacturer you are looking at, and vendor specifications sometimes do not match real-world performance, but the trajectory is unmistakable. Across roughly eight years, mining hardware has become around seven times more efficient.

After the 2024 halving, anything running above about 25 J/TH started losing money in most electricity markets. A miner that was profitable at 30 J/TH the week before the halving was upside down the week after. No machine broke. The math changed underneath.

ROI windows used to run 12 to 18 months for top-end gear before the halving. Post-halving, those windows stretched to two to three years or longer in most cases, absent a major price move. Miners with access to cheap capital can carry the depreciation. Miners running on credit cannot.

Physical limits are catching up to the chip designers as well. The latest process nodes deliver smaller efficiency gains than the previous generation, and the gap will keep narrowing. Future improvements will come slower while breakeven difficulty keeps creeping up. The runway for new hardware to outpace difficulty is shorter than it used to be.

Try the leading VASP data free for 1 month. No card required.

How the Survivors Actually Operate

The miners who have come through multiple halvings tend to do the same handful of things.

Own the power, or lock it in long-term. Marathon Digital has invested in flare-gas projects to capture stranded methane that would otherwise be vented, and has acquired or contracted significant wind capacity in Texas. The exact megawatt figures and project structures shift with each quarterly update, and Marathon’s own filings should be the source for current numbers, but the strategic point holds. By generating or securing power at the source, miners reduce their exposure to spot-market volatility and can run older hardware that would be unprofitable on grid rates.

Get paid to turn off. Demand-response contracts pay miners to curtail operations during grid stress. Riot’s $9.5 million in power credits in July 2022 is the cleanest public example. In some months, curtailment payments have exceeded mining revenue. Miners function as a grid-scale battery, soaking up power when it is cheap and disappearing when it is expensive. Utilities like the predictability. Miners like the cash.

Run only the newest gear. Cambridge’s mining model assumes a maximum five-year hardware lifetime. Real life is shorter for most operations. Sites retire older rigs aggressively and only run efficient silicon. Anything else burns cash on the operating side and tied-up capital on the balance sheet side.

Lock in power prices. Fixed-rate power purchase agreements, long-term contracts with utilities, or direct ownership of generation. Whatever insulates the miner from spot-market swings is worth more than it costs to negotiate.

Capture waste heat. Some operators recover heat from rigs and sell or use it for district heating, agriculture, or industrial processes. Marathon has pilot projects in this space. It does not fix the BTC math on its own, but it adds a revenue line and softens the carbon optics with regulators and investors.

Diversify into compute. Some miners are testing GPU workloads for AI inference on the same power infrastructure. The pitch is that they can shift between Bitcoin mining and AI compute depending on which one pays more on a given day. The economics of this are still being worked out, but the optionality has real value.

Skip the wait. Buy a licensed company.

Spread the geography. Large miners run sites across multiple jurisdictions. If one grid breaks, one regulator gets hostile, or one country’s policy shifts, the rest of the operation keeps producing.

The Financial Layer

Mining is capital-intensive in a way that does not show up until something goes wrong. Public miners issue equity, take on convertible debt, and refinance constantly to fund hardware purchases. A well-capitalized balance sheet lets you ride through bad months. A leveraged one gets eaten by them.

Hedging the BTC price is possible but complicated. Some miners sell forward part of their expected production using futures or structured products. It caps upside but stabilizes revenue. Locking in power prices through PPAs is more common and arguably more important, since power costs are the larger swing factor day to day.

The cleanest survival strategy combines cheap power, modern hardware, and a balance sheet that can absorb a six-month drought. Miss any of those three and the miner becomes vulnerable to whatever happens next.

Risks That Do Not Care About Your Strategy

China’s 2021 mining ban erased hundreds of EH/s overnight. The US has flirted with mining taxes and energy-use restrictions. New York paused new mining permits in 2023. Whatever rules exist where you operate today could change next year, often with little warning.

Texas asks miners to curtail voluntarily most of the year, and sometimes the curtailment becomes mandatory. A cold snap can shut a site down for days. A blackout can damage hardware.

Bitcoin price drawdowns of 60% or more have historically broken leveraged miners. The 2022 crash combined with rising power costs is what killed Core Scientific, and similar dynamics have eaten smaller operators in every prior bear cycle.

The future of finance is here. Trade crypto and more

Counterparty exposure has wiped out otherwise competent miners. Core Scientific’s collapse was worsened by a $7 million unpaid electricity bill left behind by Celsius. Hosting customers can vanish. Suppliers can delay deliveries. Banks can pull credit lines during the exact months you need them most.

ESG and political pressure is the underrated risk. Mining is a soft target. Even when activism does not change formal policy, it can change financing terms and public perception fast enough to affect operations.

Who Lived and Who Did Not

Marathon Digital is the clearest example of vertical integration paying off. The company invested in domestic ASIC manufacturer Auradine, built or contracted power-generation capacity, and pursued waste-heat projects. Its cost per BTC produced has tracked among the lowest in the industry. The stock has been volatile, as all mining equities are, but the company has avoided the kind of distress that took down its weaker peers.

Core Scientific filed Chapter 11 in December 2022. Its filings blamed slumping Bitcoin prices, rising energy costs, and the $7 million unpaid debt from Celsius. The company lost $434 million in Q3 2022 alone. Heavy debt plus expensive power plus a price crash plus a major customer going under is the worst-case stack, and Core had all four at once. The company eventually restructured, but the existing equity was wiped out.

Hut 8 survived 2022 to 2023 by getting smaller. It paused dividends, issued stock, secured a Canadian demand-response credit line, and later sold its 310 MW power portfolio to TransAlta for debt relief. The pattern across mid-tier survivors is similar: trim, refinance, and wait for the cycle to turn.

Argo Blockchain, with sites across Colorado, South America, and Canada, used geographic spread to ride out regional power crunches. The company curtailed during 2022 like Riot did, and sold off non-mining assets to focus on Bitcoin operations.

The cloud-mining side has a separate body of failures. HashFlare, BitClub, and Genesis Mining all collapsed or shut down at various points. BitClub’s operators were prosecuted for running a Ponzi scheme dressed as a mining operation. These were not mining companies that failed. They were marketing operations that never had the hardware they claimed to have.

Buy bitcoin and stocks. Signup to Coinbase today

Mining Scams Worth Recognizing

Mining attracts scams because it sounds technical and most retail investors cannot verify any of the claims being made. The recurring patterns:

Fake cloud mining is the largest category by victim count. Sites sell hashpower at guaranteed daily returns with no verifiable mining operation behind them. The structure is a classic Ponzi, with early investors paid out of later deposits until the operator vanishes or gets indicted.

Counterfeit hardware sales target the smaller buyers. Brand-new ASICs at impossible prices, often payment-in-crypto only. The hardware never arrives, or it arrives as used gear, or it arrives pre-configured to mine into the seller’s wallet for a few months before the buyer notices.

Hosted mining fraud sells you a fee structure for ASICs that allegedly sit in a faraway facility you will never visit. Serial numbers do not correspond to real machines. Monthly statements get fabricated until withdrawal day arrives and something always goes wrong with the payout.

Multi-level mining schemes use pyramid structures where you earn commissions on recruits. Real mining does not pay recruitment bonuses. If the comp plan looks like an MLM, that is what it is.

Cryptojacking sits in a different category. Malware that hijacks your computer or phone to mine for someone else. The symptoms are slowdown, overheating, and antivirus alerts. Keep your software patched and avoid unknown executables.

The red flags across all of them:

Try the leading VASP data free for 1 month. No card required.

• Guaranteed or fixed returns. Real mining revenue swings with price, fees, difficulty, and power. None of those variables are stable enough to support a fixed yield.
• No verifiable physical operation. No photos with serial numbers, no facility address, no third-party verification, no audit trail.
• Payment only in crypto. Untraceable, non-refundable, off the financial system. The scammer’s preferred rail.
• High pressure and short windows. “Last spots,” “today only,” “guaranteed if you act in the next hour.”
• Heavy recruitment focus. Earnings driven by signing up new members rather than by mining output.
• Prices far below market. A new S21 series machine selling for a fraction of the manufacturer’s price is a scam, every time.

If two or more of these line up, walk away. If three or more line up, the question is no longer whether you are looking at a scam but how big it is.

Frequently Asked Questions (FAQ)

What is a Bitcoin halving and why does it affect miners so much? 

A protocol-mandated 50% cut to the block subsidy every 210,000 blocks, roughly every four years. The most recent halving in April 2024 dropped the reward from 6.25 to 3.125 BTC per block. Revenue is cut in half overnight. Costs stay the same. Inefficient miners turn off until difficulty adjusts downward or the price recovers.

How much does electricity cost really affect mining? 

It runs 60 to 80% of operating cost. At $0.10/kWh, a typical rig costs around $2,600 a year to run. At $0.05, it is closer to $1,300. The difference between profitable and not, in most cases, comes down to a few cents per kWh.

How do I spot a mining scam? 

Guaranteed returns is the biggest single tell. Beyond that: no verifiable hardware, payment only in crypto, recruitment-heavy compensation, prices far below market, and high-pressure sales tactics.

What did successful miners do through the last halving? 

Owned or locked in their power, ran only the latest ASICs, used demand-response contracts to monetize curtailment, and kept their balance sheets in shape. Marathon’s vertical integration into power generation is one example. Failed miners over-leveraged and ran older hardware on expensive grids.

Which ASICs are worth running today? 

The latest generation, with efficiency in the mid-teens of joules per terahash. Anything above 25 J/TH is marginal in most electricity markets, and likely a liability in retail-rate ones.

How long is ROI on new hardware now? 

Before the 2024 halving, top-end gear could pay back in 12 to 18 months in favorable conditions. After the halving, that has stretched to two to three years for most operations, longer at higher electricity costs. The exact figure depends heavily on the BTC price between purchase and break-even.

Skip the wait. Buy a licensed company.

Is cloud mining ever safe? 

Most cloud mining is fraudulent or close to it. A few legitimate services exist, usually run by large exchanges, but the category is dominated by Ponzi structures. Default skepticism is the right starting point.

What is the single biggest red flag in mining offers? 

Guaranteed returns. Mining revenue is variable by definition. Anyone promising fixed daily yields is running a scheme, not a mining operation.

How do I protect myself when buying ASICs? 

Buy from authorized distributors. Verify sellers through manufacturer channels. Use escrow for large orders. Avoid deals run entirely through instant messaging or social media. If the price is far below market, it is not a deal.

I think I was scammed. What now? 

Stop sending money. Document everything you have, including emails, transaction hashes, and screenshots. Report to local law enforcement and to crypto-specific scam trackers. Recovery is rare, but reporting helps protect the next person. Never pay an upfront fee to a “recovery service.” Those are secondary scams targeting victims of the first one.