Bitcoin’s Difficulty and the 2028 Halving: A Security Test

The Number That Barely Moved

On October 3, 2026, Bitcoin did something that almost never makes headlines: nearly nothing. The network ran its scheduled recalculation of mining difficulty, the self-adjusting setting that keeps new blocks arriving roughly every ten minutes, and the number barely budged. Difficulty slipped by about 0.03 percent, essentially a flat print, leaving the network dialed in near 132.72 trillion, according to mempool.space and the difficulty tracker at CoinWarz.

A flat difficulty reading is not boring. It is the sound of a machine that has found its level. After a brutal summer that pulled difficulty more than ten percent below where it started the year, miners have switched gear back on now that Bitcoin trades comfortably above their cost of production. The price sat near $85,970 on October 6, according to CoinGecko, within roughly a third of its record high, as the market leaned into the seasonal optimism that traders like to call Uptober. The next adjustment, due around October 17, is estimated to nudge difficulty up by about one percent.

So the thermostat is steady, revenue is recovering, and the chain is humming. That makes this a good moment to ask an uncomfortable question about the future. Bitcoin’s difficulty is not merely a mining statistic. It is the single clearest readout of how much security the network is paying for, and in April 2028 the largest source of that security spending gets cut in half overnight. This is the story of what difficulty actually measures, why it moves the way it does, and why the next halving is a stress test for the number that has quietly held Bitcoin together for more than fifteen years.

Difficulty in Sixty Seconds

Bitcoin aims to produce one block about every ten minutes, no matter how much computing power is pointed at it. The problem is that computing power is never constant. Machines get faster, operators switch rigs on and off, and whole regions come and go from the map. If the network did nothing, more hardware would mean faster blocks and a faster-than-planned issuance of new coins.

Difficulty is the correction. Every 2,016 blocks, roughly every two weeks, the software looks at how long those blocks actually took and compares it to the target of 20,160 minutes. If they came in fast, difficulty rises to slow them back down; if they came in slow, difficulty falls. No committee votes on it, and no company sets it. Satoshi Nakamoto described the mechanism in the original Bitcoin white paper: the proof-of-work difficulty is “determined by a moving average targeting an average number of blocks per hour. If they’re generated too fast, the difficulty increases.”

There are guardrails. A single adjustment cannot move difficulty by more than a factor of four up or a factor of four down, a clamp that has never once been hit in either direction. The system is deliberately slow and deliberately dumb, and that is the point. It is a thermostat with exactly one job, and it does that job on autopilot.

Why Difficulty Is Really a Security Number

Here is where most explainers stop and where the interesting part begins. Difficulty is not just about block timing. It is a near-perfect proxy for the total computing power guarding the chain, and that computing power is the thing an attacker would have to out-muscle to rewrite history.

The link is arithmetic. Network hashrate is not measured directly; it is inferred from difficulty using a simple relationship, hashrate is approximately difficulty multiplied by 2 to the 32nd power, divided by the 600-second block target. Plug in today’s 132.72 trillion and you get roughly 950 exahashes per second of honest work, a figure that matches what trackers report. When difficulty rises, it is telling you that more energy and more hardware have committed to the network. When it falls, that commitment is walking away. For the full picture of how that raw power has scaled, see our breakdown of what Bitcoin’s hashrate growth really means.

Security follows directly. To reverse a confirmed transaction or censor the chain, an attacker needs to produce more proof-of-work than the rest of the network combined, the classic 51 percent attack. The more hashrate behind Bitcoin, the more hardware, electricity, and time such an assault would cost. Difficulty, in other words, is the dollar sign on the chain’s front door. A higher number means a more expensive lock. That is why a falling difficulty is not just a mining curiosity; it is a quiet reduction in how much it costs to attack Bitcoin.

The Dial Heals the Heartbeat by Shedding the Muscle

The self-correcting thermostat has a feature that doubles as a flaw, and it sits at the center of everything that follows. Difficulty guarantees that Bitcoin keeps producing blocks. It does not guarantee that it keeps producing expensive ones.

Watch the feedback loop when revenue drops. Suppose the Bitcoin price falls below what it costs higher-cost miners to run their machines. Those operators power down because every block they mine loses money. With less hardware online, blocks slow down. Two weeks later, difficulty drops to compensate, which hands the surviving miners more coins per unit of their hashrate and restores profitability. Equilibrium returns, blocks are back to ten minutes, and the chain never skips a beat. But the network is now secured by less total hashrate than before. The heartbeat is fine; the muscle is thinner.

Nikolaos Panigirtzoglou and the global markets strategy team at JPMorgan have put hard numbers on this reflex. In a widely cited note summarized by TFTC, the bank described the mechanism plainly: “when bitcoin trades below its production cost, higher-cost miners power down, the hashrate declines, and difficulty adjusts lower.” JPMorgan pegs the sensitivity of hashrate to price at a beta of about 0.62, meaning a sustained move in price pulls difficulty along with it, just more gently. The strategists put the network’s all-in production cost in the neighborhood of $78,000 per coin earlier in 2026, which is precisely why the summer was so painful and why the autumn recovery has been so visible in the difficulty chart.

Keep that loop in mind, because the 2028 halving is going to run it on purpose, at the protocol level, on a scheduled date, whether or not the price cooperates.

2026 Was a Live Stress Test in Miniature

You do not have to imagine the feedback loop. It played out in full this year. Bitcoin entered 2026 with difficulty near 148 trillion. Through the first half of the year, as the price sagged and miners diverted capital and power toward artificial intelligence and high-performance computing, difficulty ground lower. On June 14 it recorded a 10.09 percent drop, the largest single adjustment of the year and, as The Block noted, among the biggest downward moves in the network’s history. At the time, production cost was estimated around $84,300 while the spot price languished near $63,780, leaving much of the network roughly a quarter underwater.

Then the loop ran in reverse. As the price climbed back through the autumn, machines came back online, blocks sped up, and difficulty recovered. September delivered a 4.16 percent jump, its biggest single increase since June, and the October 3 retarget held the line with that near-flat print. The round trip from roughly 148 trillion down to about 125 trillion and back above 132 trillion is, in one chart, the entire thesis of this article: difficulty tracks miner revenue, and miner revenue tracks price. The table below walks through the back half of the year.

The point of the round trip is not that difficulty fell and rose. It is that the network never once stopped working while it happened. Blocks kept coming, roughly on schedule, through a decline that shed roughly fifteen percent of the chain’s security at its worst. The thermostat did its job flawlessly, and that is exactly why it is so easy to miss what the job actually is.

The Subsidy Clock: What Changes in April 2028

Bitcoin runs on two clocks, and they are easy to confuse. The difficulty clock ticks every two weeks and reacts to the present: how fast blocks are coming right now. The subsidy clock ticks once every four years and does not react to anything at all. It is written into the code, counts in blocks rather than calendar days, and cannot be lobbied.

At block height 1,050,000, the reward a miner collects for finding a block falls from 3.125 BTC to 1.5625 BTC. As of early October, fewer than 80,000 blocks remained, which points to an arrival around April 12, 2028. On that day, daily issuance drops from 450 BTC to 225 BTC, and annual issuance falls from roughly 164,000 BTC to about 82,000 BTC. Supply inflation sinks under half a percent a year, below gold.

For holders, that scarcity story is the familiar bullish case. For the security of the chain, it is something more double-edged. The block subsidy is not just new supply; it is the overwhelming majority of what miners get paid. Cut it in half, and unless something else fills the gap, you have cut miner revenue close to in half on a fixed date. The difficulty thermostat we just watched absorb a summer slump will be asked to absorb a scheduled revenue cliff.

What the Security Budget Actually Is

Strip away the jargon and the security budget is simple: it is the total amount Bitcoin pays miners, in dollars, over a given period. That payment is what funds the hardware and electricity standing guard, and it is therefore what sets difficulty. More revenue pulls more hashrate online and pushes difficulty up; less revenue does the reverse. The security budget is the cause, and difficulty is the visible effect.

Today the budget is almost entirely subsidy. At 450 BTC a day and a price near $85,970, miners collectively earn on the order of $39 million a day, or roughly $14 billion a year. Transaction fees, the other component, currently make up well under one percent of the reward on an average block. The chain is, for now, paying for its own security mostly by printing new coins.

Satoshi anticipated that this could not last forever. The same white paper that describes the difficulty thermostat also sketches the exit: “Once a predetermined number of coins have entered circulation, the incentive can transition entirely to transaction fees and be completely inflation free.” That sentence is the quiet assumption underpinning Bitcoin’s entire long-term security model. Every four years the subsidy half of the budget shrinks, and the plan is for fees to grow into the gap. The 2028 halving is the next scheduled test of whether that transition is on track, and difficulty is the gauge that will show the result.

Here is the arithmetic that makes 2028 concrete. If the subsidy halves to 225 BTC a day and the price simply stayed at today’s level, the daily budget would fall to roughly $19 million, and the annual figure to around $7 billion. Difficulty would drift down to find the hashrate that $7 billion can support, unless the price roughly doubled or fees stepped up to cover the difference. That is the whole debate in one line.

The Arithmetic That Worries the Skeptics

The pessimistic case is uncomfortable precisely because it is just multiplication. If fees stay a rounding error and the subsidy halves every four years, then the only way to hold the dollar security budget steady is for the Bitcoin price to double over each halving cycle, forever. Double every four years is a spectacular growth rate to demand in perpetuity from an asset already worth well over a trillion dollars. Miss it, and the budget shrinks in real terms, difficulty settles lower, and the chain gets cheaper to attack with each passing cycle.

This is not a fringe worry. The most rigorous early treatment, a widely circulated model of Bitcoin’s security and the declining block subsidy authored by the pseudonymous researcher Hasu with James Prestwich and Brandon Curtis, laid out the mechanics years ago and remains the reference point for the debate. The uncomfortable version of the conclusion is that a security budget funded purely by fees requires either enormous transaction demand or prohibitively expensive block space, and possibly both. Some analysts have gone further, arguing that on current trends the model starts to strain within a decade.

The skeptics’ real point is not that Bitcoin breaks in 2028. It is that difficulty, the number everyone treats as a sign of strength, is downstream of a subsidy that is engineered to disappear. A record difficulty today says nothing about whether the budget survives three or four more halvings. The gauge is honest about the present and silent about the future.

The Rebuttal: Security Is a Price, Not a Paycheck

The optimistic camp thinks the whole framing is wrong, and it has some of the industry’s sharpest minds behind it. Pierre Rochard, vice president of research at Riot Platforms and a longtime Bitcoin economist, argues that treating security as a fixed annual budget is a category error. In comments reported by Bitcoinist, Rochard framed it this way: Bitcoin’s budget for finality is “NOT a fixed paycheck; it’s a market price that rises when needed.”

The logic runs like this. An attacker does not have to out-spend Bitcoin’s average annual security spend. They have to out-spend what the network will pay to defend the specific blocks they are trying to rewrite, at the specific moment they try it. And at that moment, fees spike. Users racing to confirm transactions bid up block space, mining pools compete, and the reward for the next block can balloon far above the subsidy. Rochard’s blunt summary of the incentive: Bitcoin “does not assume miners are altruists. It assumes they are paid to end your attack.”

He points to live evidence. During fee spikes, a single block’s transaction fees have at times approached or exceeded the subsidy itself; at a busy 1,000 satoshis per virtual byte across a full block, fees alone can run on the order of 10 BTC. In this view, difficulty will always find the level the market is willing to pay for, and the market pays most exactly when security matters most. The budget is not a line item set in 2009; it is a dynamic auction that reprices in real time.

Both camps, notice, agree on the mechanism. Difficulty follows revenue. They disagree about whether the revenue will be there. That is not a question the protocol can answer. It is a question about human demand for Bitcoin block space, which is why the next few years matter more than any single retarget.

The Real Swing Variable Is Fee Demand

If difficulty after 2028 depends on revenue, and the subsidy half of revenue is scheduled to shrink, then everything hinges on the other half: fees. And fees are the most volatile, least predictable number in the entire system.

Fee revenue is a function of demand for block space, and that demand has been wildly uneven. It surged during the Ordinals and inscriptions manias, cratered when the hype faded, and spikes briefly whenever the network congests. For long stretches of 2026, fees have been a fraction of one percent of the block reward. The optimists need that figure to climb structurally and stay there, driven by genuine, recurring demand rather than one-off speculative frenzies. Candidates exist: Bitcoin as a settlement layer for second-layer systems, tokenized assets, and stablecoin flows and other on-chain money that settle against the base chain. Whether any of them generates the sustained fee pressure the budget needs is still an open question.

The honest answer is that nobody knows. What we can say is that difficulty is the scoreboard. If fee demand grows into the subsidy gap over the coming halvings, difficulty will hold or climb through them, and the security-budget worry quietly dissolves. If it does not, difficulty will step down after each halving, and the skeptics will have been right. The 2028 retarget series will be the first clean data point, and it is less than two years away.

The New Bidder: AI Wants the Same Megawatts

There is a 2026 twist the original security-budget debate never contemplated. Even if the Bitcoin price holds and fees grow, difficulty can still stall, because mining now has a deep-pocketed rival for its single most important input: cheap electricity.

Artificial intelligence and high-performance computing pay far more per megawatt than Bitcoin mining does, and they pay it on steadier terms. Through 2026, listed miners have signed tens of billions of dollars in AI and HPC hosting deals and converted data-center capacity away from hashing. The research team at CoinShares, led by James Butterfill, has described the stretch as “one of the most challenging periods” for mining economics and projected that a large share of listed-miner revenue could come from AI rather than Bitcoin by the end of the year. That capital and power leaving the hashrate pool is one reason the 2026 difficulty decline was voluntary rather than forced. The same dynamic is reshaping the broader compute market, as our look at the DePIN GPU race between Akash, io.net, and Render lays out.

Fred Thiel, chief executive of MARA, has been blunt about where this leads. Mining, he told CoinDesk, is a game where “the floor is your energy cost,” and he has warned that “by 2028, you’ll either be a power generator, be owned by one, or be partnered with one.” Read through the lens of difficulty, that is a warning that the thermostat’s future setting depends on who wins the auction for the next marginal megawatt. If AI keeps outbidding mining, hashrate growth slows, and difficulty may struggle to climb through the 2028 revenue cut even in a bull market.

What Difficulty Does Not Promise

It is worth being precise about what a high difficulty does and does not buy. It makes a 51 percent attack expensive, but it does not make one impossible, and a falling difficulty makes one cheaper in lockstep. The number is a measure of the lock, not a guarantee the door stays shut.

Campbell Harvey, the Duke University finance professor, has modeled what breaking that lock would cost. His latest estimate puts a majority-hashrate attack at roughly $8 billion, which he frames as “about 50 basis points of the value of bitcoin,” once the attacker pairs the hardware with a large short position in offshore derivatives so the scheme profits from the price crash it causes. As reported by News.Bitcoin.com, the point is not that the attack is likely; most of the industry considers it economically self-defeating. The point is that the price tag scales directly with hashrate, and hashrate scales with difficulty. Let difficulty fall far enough for long enough, and that $8 billion figure falls with it.

This is also where Bitcoin’s design philosophy diverges sharply from proof-of-stake chains. Ethereum and its ecosystem secure the network with staked capital that can be slashed, a model that has spawned an entire layer of restaking economics built on reusing that collateral. Bitcoin secures itself with burned energy, an external, ongoing cost that cannot be recycled. Difficulty is the running meter on that energy spend. The 2028 question is whether the meter can keep running high when the subsidy that pays the bill is cut.

Reading the Dial Into 2028

So what should a careful observer actually watch between now and the halving? A handful of signals turn the abstract debate into something you can track from a chart.

• Difficulty versus price. With a price-to-hashrate beta near 0.62, difficulty should broadly follow the price. Watch for difficulty holding up when price wobbles; that is the sign of a network with low-cost power and committed operators.
• The hashprice floor. Hashprice, miner revenue per unit of computing power, sat around $40 per PH per day in early October. Many operators start switching off below roughly $35. The gap between those two numbers is the network’s margin of safety heading into a halving that will, all else equal, roughly halve hashprice overnight.
• Fee share of the reward. The single most important long-term number. If fees climb from a fraction of a percent toward double digits as a share of the block reward, the transition Satoshi described is happening. If they stay flat, the skeptics’ math takes over.
• The two clocks. Difficulty retargets every two weeks; the subsidy halves every four years. The first reacts; the second does not. 2028 is when the slow clock forces the fast one to respond.

One more point of context matters for readers who think about regulation. Mining and the difficulty adjustment are pure protocol mechanics, not an investment product, and United States regulators have treated them that way. The SEC’s Division of Corporation Finance stated in March 2025 that certain proof-of-work mining activities are not securities transactions, a posture that has held through the quieter enforcement environment our coverage of SEC crypto enforcement in 2026 describes. Difficulty sits outside the securities perimeter; it is physics and game theory, not a contract.

The Bottom Line: Alive Is Not the Same as Safe

Difficulty is the most reassuring number in Bitcoin and one of the most misunderstood. It never fails. Through government bans, price crashes, winter storms, and a voluntary 2026 exodus of power toward AI, it has kept blocks arriving about every ten minutes without a single missed beat. That reliability is real, and it is why the chain has never stopped running.

But reliability is not the same as security. The thermostat guarantees that Bitcoin keeps producing blocks; it does not guarantee that those blocks stay expensive to forge. When revenue falls, difficulty restores the heartbeat precisely by letting the muscle thin. The flat October print is a snapshot of a network perfectly at rest, revenue and hashrate in balance at a price near $86,000.

April 2028 moves that balance. Half the subsidy disappears on a date no one can lobby or delay, and the thermostat will do exactly what it always does: find the level that the remaining revenue can support. Whether that level is higher or lower than today is not up to the protocol. It is up to price and it is up to fees. The dial will keep the chain alive either way. The open question, the one worth watching retarget by retarget between now and the halving, is how much security it will be keeping alive.

Frequently Asked Questions

How often does Bitcoin mining difficulty change?

Bitcoin recalculates its mining difficulty every 2,016 blocks, which works out to roughly every two weeks. The network compares how long those blocks actually took against a target of 20,160 minutes and rescales difficulty up or down to keep the average block time near ten minutes.

What is Bitcoin’s mining difficulty in October 2026?

As of October 6, 2026, Bitcoin’s mining difficulty sits near 132.72 trillion. The October 3 retarget was essentially flat at about -0.03 percent, and the next adjustment, expected around October 17, is estimated to rise by roughly one percent.

What happens to mining difficulty after the 2028 Bitcoin halving?

The 2028 halving cuts the block subsidy from 3.125 BTC to 1.5625 BTC, which roughly halves miner revenue unless price or fees make up the gap. If revenue falls, some miners switch off and difficulty adjusts lower until the survivors are profitable again, leaving the network secured by less total hashrate.

What is the Bitcoin security budget?

The security budget is the total amount Bitcoin pays miners over a given period, combining the block subsidy and transaction fees. That revenue funds the hardware and electricity that secure the chain, so it is what ultimately sets the hashrate and the difficulty level. Today the budget runs around $14 billion a year and is almost entirely subsidy.

Does lower difficulty make Bitcoin easier to attack?

Yes. Difficulty is a proxy for the total hashrate guarding the chain, and a 51 percent attack must out-muscle that hashrate. When difficulty falls, the cost of such an attack falls with it. Duke professor Campbell Harvey estimates a majority attack would currently cost roughly $8 billion, a figure that scales directly with difficulty.

By Marcus Okafor, Bitcoin and mining desk, HOGE Wire.