ASICs vs GPUs Payback Period

The Capital Allocation Decision

You have $5,000 to deploy. You can buy Bitcoin directly at current price, or you can buy an Antminer S19 and wait for it to mine that amount back. At $0.10 per kilowatt-hour electricity and October 2026 difficulty of 132.7 trillion, the miner costs you $7.79 per day in power. It generates roughly $3 to $5 per day in Bitcoin at current hashprice. You are bleeding capital from day one.

This is not a corner case. This is the default outcome for residential mining in 2026.

Mining hardware represents a capital deployment competing with staking, lending, or direct asset purchase. The decision is not whether mining is profitable in absolute terms. The decision is whether mining delivers better risk-adjusted returns than the alternatives. At current network difficulty and residential power rates, it does not. This article ranks the hardware by real payback period and identifies the narrow edge cases where mining still makes economic sense.

ASIC Miners: Bitcoin SHA-256 Hardware

The Antminer S19 remains the most widely discussed ASIC in 2026, though it is no longer flagship. Hashrate is 95 terahashes per second. Power draw is 3,250 watts. Efficiency is 34.2 joules per terahash. At Canadian average residential rate of $0.10 per kilowatt-hour, electricity costs are $7.79 per day, which compounds to $2,843 per year.

The S19 is available on the secondary market at a fraction of new-generation pricing. You can find units for $1,500 to $2,500. But discounted hardware does not fix the underlying math. At current Bitcoin difficulty near 133 trillion and hashprice around $24 per petahash per day, the S19 generates approximately $2.28 per day in gross revenue. Subtract $7.79 in power. You lose $5.51 per day, every day, until difficulty drops or Bitcoin price rises enough to flip the equation.

The Antminer S19 XP improves the numbers but does not save the model. Hashrate climbs to 140 terahashes per second. Power draw falls slightly to 3,010 watts. Efficiency improves to 21.5 joules per terahash. Monthly power cost at $0.10 per kilowatt-hour is $217. Gross revenue at current difficulty and hashprice is roughly $3.36 per day. Net loss is still $3.88 per day.

The Whatsminer M50 delivers 114 terahashes per second at 3,306 watts and 29 joules per terahash. At $0.10 per kilowatt-hour, it generates a daily loss of $7.93. At $0.07 per kilowatt-hour, the loss compresses to $0.79 per day, which is $289 per year. Still negative.

If you lower electricity cost to $0.05 per kilowatt-hour, the game changes. Top-tier ASIC miners like the Antminer S21 XP can break even within 18 to 27 months at rates below $0.05 per kilowatt-hour. But $0.05 per kilowatt-hour is not a residential rate. It is an industrial rate, or it is stranded hydro, or it is curtailed wind in West Texas. If you are reading this article to decide whether to plug an ASIC into your home circuit, you do not have $0.05 per kilowatt-hour access.

A roughly $4,800 Antminer S21 XP running at $0.08 per kilowatt-hour all-in power rate might produce on the order of $100 per month in estimated net profit. Payback period is five to six years. In equities, a six-year breakeven on hardware that becomes obsolete in three years would not pass initial due diligence.

GPU Mining: Kaspa, Ergo, and the ASIC Problem

GPU mining profitability collapsed in 2026. Estimated payback periods now stretch to 1,500 days or more for most configurations. The best-case scenario involves mining Kaspa with a high-end GPU at electricity rates below $0.07 per kilowatt-hour. Even that scenario delivers marginal returns.

An RTX 4090 achieves roughly 0.9 to 1 gigahash per second on kHeavyHash, the algorithm Kaspa uses. The Antminer KS5 Pro delivers 21,500 gigahashes per second. The ASIC outperforms the GPU by 21,500 times in raw hashrate. Power consumption per hash is not 21,500 times higher. The ASIC wins on every dimension.

At $0.10 per kilowatt-hour, the RTX 4090 generates approximately $0.80 to $1.20 per day in profit mining Kaspa. The card costs $1,600 or more. Payback period stretches to four to five years, assuming difficulty and coin price remain static. They will not remain static. Difficulty has been climbing steadily as more KAS-dedicated ASICs come online. Your payback period extends every time another miner joins the network.

Ergo mining with an RTX 3070 produces similar results. The card delivers 160 megahashes per second at 125 watts. Used price is roughly $280. Current profitability sits around $0.80 to $2.00 per day after electricity costs. At the high end of that range, payback is 140 days. At the low end, payback is 350 days. But those numbers assume you can sell Ergo immediately at current market price with no slippage, no volatility, and no difficulty adjustment. In practice, small-cap altcoin mining introduces liquidity risk that Bitcoin mining does not.

If you own a GPU already and your electricity is included in rent or otherwise unmetered, running the card to mine Kaspa or Ergo can make sense as marginal income. If you are buying hardware specifically to mine, the ROI does not justify the capital outlay. You would generate better returns buying the underlying asset directly and staking it where applicable.

For a deeper explanation of how mining rewards, fees, and difficulty adjustments interact to determine real profitability, see What Crypto Mining Produces: Rewards, Costs, And ROI At Current Difficulty.

Electricity Cost as the Dominant Variable

Electricity cost is the single most important variable in crypto mining profitability, outweighing even hardware efficiency in determining break-even timelines. This is not a talking point. This is the result of running the numbers across every major ASIC and GPU configuration at every residential and industrial power rate.

At $0.10 per kilowatt-hour, most hardware operates at a loss. At $0.07 per kilowatt-hour, operations struggle to justify continued mining at current Bitcoin prices. At $0.05 per kilowatt-hour or below, mining becomes viable if you have access to efficient hardware at 12 joules per terahash or better.

The threshold is sharp. Moving from $0.08 to $0.06 per kilowatt-hour can compress payback period by 40 percent or more. The difference between $0.10 and $0.05 per kilowatt-hour is the difference between hemorrhaging capital and slow, steady accumulation.

If you have access to sub-$0.05 power, mining competes with other yield strategies. Compare it directly to staking returns. Ethereum staking pays roughly 4 percent annual percentage yield with 0.5 percent inflation. Cosmos pays 18.5 percent with 12 percent inflation. If your mining setup delivers net annualized return below 4 percent after accounting for hardware depreciation, electricity, and difficulty increases, you are better off staking ETH. For a full breakdown of staking yields across chains, see Crypto Staking Returns: Where Yield Comes From And What It Costs.

Most residential miners do not have sub-$0.05 power. Most do not have $0.07 power. The national average in the United States is roughly $0.14 per kilowatt-hour. Canada averages closer to $0.10. Europe is higher. If you are paying residential rates, you are not competing with industrial farms. You are subsidizing them by driving difficulty higher while losing money per unit of hashrate.

Edge Cases Where Mining Economics Work

There are four scenarios where mining hardware delivers competitive risk-adjusted returns compared to direct asset purchase. All four require conditions that most retail participants cannot access.

First: sub-$0.05 per kilowatt-hour industrial power. This requires either a commercial contract with a utility, access to stranded energy from hydro or wind, or participation in curtailment programs where miners absorb excess grid capacity during low-demand periods. Cryptocurrency mining functions as a flexible economic battery in these arrangements, storing energy as digital asset with global value. Solar project payback periods can drop from 8.1 years to 3.5 years when mining is coupled with renewable resources. Mining profitability increases by over 20 percent through participation in ancillary service markets.

Second: heat reuse in cold climates. The Antminer S19 produces 3,250 watts of heat, which converts to roughly 11,089 BTU per hour. If you would otherwise run a space heater in winter, the ASIC replaces that heater and mines Bitcoin as a byproduct. The electricity cost is sunk either way. This does not eliminate the cost, but it does change the opportunity cost calculus. You are paying for heat, not for mining. The mining revenue becomes a rebate on your heating bill.

Third: hardware depreciation tax treatment for institutional buyers. Bitcoin mining hardware qualifies for 100 percent bonus depreciation under current U.S. tax code. If you operate a mining business and have sufficient taxable income to shield, the depreciation benefit can compress effective payback period by 20 to 30 percent depending on your marginal rate. This edge is unavailable to individual retail miners unless they structure as a business and have other income to offset.

Fourth: access to zero-marginal-cost power through direct ownership of generation assets. If you own solar panels, wind turbines, or micro-hydro and produce more power than you consume, mining converts surplus power into an exportable asset. Grid buyback rates for excess solar in most jurisdictions are lower than retail rates. Mining captures more value per kilowatt-hour than selling back to the grid.

These are the edge cases. They are real. But they are edges, not defaults. If you are evaluating mining hardware as a passive income strategy without access to one of these four conditions, the ROI does not work.

Direct Asset Purchase vs. Mining Hardware: The Apples-to-Apples Comparison

Here is the test. You have $5,000. You can buy Bitcoin today at current market price, or you can buy an Antminer S19 XP on the secondary market for roughly $3,000, pay $2,000 for electrical setup and cooling, and begin mining.

If you buy Bitcoin directly, you own $5,000 of Bitcoin on day one. Your exposure is immediate. Your downside is Bitcoin price risk. Your upside is Bitcoin price appreciation. No ongoing costs.

If you buy the miner, you own zero Bitcoin on day one. You have $5,000 of sunk cost in hardware and infrastructure. You pay $7.20 per day in electricity at $0.10 per kilowatt-hour. You mine approximately $3.50 per day in Bitcoin at current difficulty and hashprice. Net loss is $3.70 per day, which is $1,350 per year. Your breakeven requires either Bitcoin price to rise significantly or difficulty to drop. Neither is under your control.

The scenarios where mining wins are scenarios where Bitcoin price rises sharply after you buy the hardware. But in those scenarios, buying Bitcoin directly would have delivered the same or better returns with less operational complexity, no ongoing electricity cost, and no hardware obsolescence risk.

The only scenario where mining outperforms direct purchase is the scenario where you have access to one of the four edge cases described above. If your electricity cost is $0.04 per kilowatt-hour, the math flips. If you are replacing a space heater, the math flips. If you have depreciation to capture, the math improves. Absent those conditions, direct purchase wins.

For investors seeking yield on crypto holdings, staking and lending strategies typically deliver better risk-adjusted returns than mining. Mining introduces operational risk, obsolescence risk, and ongoing costs that staking does not. If you are deploying capital for passive income, compare mining returns against staking returns after accounting for all costs. Most of the time, staking wins.

Payback Period Summary: The Realistic Numbers

Antminer S19 at $0.10/kWh: Never. Negative daily cash flow.

Antminer S19 at $0.07/kWh: Marginal. Payback exceeds five years if difficulty remains flat, which it will not.

Antminer S19 XP at $0.10/kWh: Never. Negative daily cash flow.

Antminer S19 XP at $0.05/kWh: 18 to 27 months, assuming stable difficulty and Bitcoin price.

Whatsminer M50 at $0.10/kWh: Never. Loss of $7.93 per day.

RTX 4090 mining Kaspa at $0.10/kWh: Four to five years, assuming no difficulty increase and stable coin price. Both assumptions are false.

RTX 3070 mining Ergo at $0.10/kWh: Five months to one year, depending on Ergo price and liquidity. Liquidity risk is high.

These are the real numbers at current network difficulty, current coin prices, and current hashrate. If Bitcoin price doubles, all payback periods compress proportionally. If difficulty doubles, all payback periods extend or turn negative. You are betting on the ratio of price to difficulty. That ratio has been compressing for most of 2026 as institutional hashrate has come online and Bitcoin price has remained range-bound.

If you are considering mining as a capital allocation decision, model your breakeven under three scenarios: base case (current price and difficulty), bear case (20 percent price drop and 15 percent difficulty increase), and bull case (50 percent price increase and 10 percent difficulty increase). If your payback period exceeds three years in the base case, the investment does not compete with alternatives.

For tools to model your own scenarios and track mining profitability in real time, see What is Cloud Mining? for an alternative model that removes hardware risk but introduces counterparty risk.

Tax and Reporting Implications

Mining income is taxable as ordinary income at the fair market value of the coin on the day you receive it. If you mine 0.01 Bitcoin on a day when Bitcoin trades at $60,000, you have $600 of taxable income. When you later sell that Bitcoin, you recognize capital gain or loss based on the difference between your $600 cost basis and the sale price.

This creates a reporting burden. Every block reward is a taxable event. Every payout from a mining pool is a taxable event. If you mine daily for a year, you have 365 taxable events to track. Most tax software handles this, but you need to feed it accurate data. If you use a mining pool, download your payout history. If you mine solo, track your block rewards manually.

Hardware depreciation is deductible if you operate as a business. If you file Schedule C as a sole proprietor or operate through an LLC, you can depreciate mining hardware over five years using MACRS, or you can take 100 percent bonus depreciation in year one under current rules. This benefit is large. A $5,000 ASIC that you depreciate fully in year one reduces your taxable income by $5,000, which saves you $1,850 in tax if your marginal rate is 37 percent.

Electricity is deductible as an operating expense if you operate as a business. If you pay $2,800 per year in electricity to run a mining operation, that $2,800 reduces your taxable income. Combined with depreciation, the tax benefit can turn a marginal mining operation into a positive after-tax return, even if pre-tax cash flow is negative.

None of this applies if you mine as a hobby. If you do not operate as a business, you cannot deduct expenses under current U.S. tax law. You still owe tax on the income. You just lose the deductions. This is why most serious miners structure as businesses.

For a full breakdown of crypto tax reporting at different transaction volumes and how to track mining income efficiently, see What Crypto Tax Software Really Costs At Your Transaction Volume.

The Takeaway

Mining hardware is a leveraged bet on the ratio of coin price to network difficulty, paid for with daily electricity cost. At $0.10 per kilowatt-hour, that bet loses money on every major ASIC and GPU in 2026. The breakeven threshold is $0.05 per kilowatt-hour or lower, which eliminates most residential participants. If you cannot access industrial power rates, or reuse the heat, or depreciate the hardware against business income, buying the asset directly delivers better risk-adjusted returns than mining it. The edge cases are real, but they are not the default. Model your payback under realistic assumptions before you deploy capital into hardware that might never return it.

Frequently Asked Questions

What is the realistic payback period for an Antminer S19 at home electricity rates?

At $0.10 per kilowatt-hour, the Antminer S19 operates at a net loss of $5.51 per day. It never reaches payback. Power costs $7.79 daily while mining revenue is roughly $2.28 per day at October 2026 difficulty of 132.7 trillion. Even at $0.07 per kilowatt-hour, payback exceeds five years, assuming difficulty and Bitcoin price remain constant, which they will not.

Is GPU mining still profitable in 2026?

GPU mining profitability collapsed in 2026, with payback periods exceeding 1,500 days for most configurations. An RTX 4090 mining Kaspa at $0.10 per kilowatt-hour generates $0.80 to $1.20 daily profit. With a $1,600 purchase price, payback stretches to four to five years. Dedicated Kaspa ASICs outperform GPUs by 21,500 times in hashrate, eliminating any GPU efficiency advantage.

At what electricity cost does Bitcoin ASIC mining become profitable?

Top-tier ASIC miners break even within 18 to 27 months at electricity rates below $0.05 per kilowatt-hour. At $0.07 per kilowatt-hour, operations struggle to stay profitable at current Bitcoin prices. At $0.10 per kilowatt-hour, even efficient hardware operates at a loss. Industrial power contracts, stranded renewable energy, or curtailment programs are required to access sub-$0.05 rates.

When does mining hardware deliver better returns than buying Bitcoin directly?

Mining outperforms direct purchase only when you have sub-$0.05 per kilowatt-hour power, heat reuse value in cold climates, 100 percent bonus depreciation from business tax treatment, or zero-marginal-cost power from owned generation assets. Without one of these four edge cases, buying Bitcoin directly delivers better risk-adjusted returns with no operational costs, hardware obsolescence, or ongoing electricity expense.

Can I deduct mining electricity and hardware costs on my taxes?

You can deduct electricity as an operating expense and depreciate hardware if you operate mining as a business filing Schedule C or through an LLC. Bitcoin mining hardware qualifies for 100 percent bonus depreciation under current U.S. tax code. If you mine as a hobby, you owe tax on income but cannot deduct expenses. Most serious miners structure as businesses to capture depreciation and expense deductions that can turn marginal operations into positive after-tax returns.