Bitcoin Halving Cycle Math: The Great Hashrate Migration

A Difficulty Cut Nobody Was Watching

On July 11, 2026, Bitcoin’s network difficulty fell 5%, from 133.87 trillion to 127.17 trillion, the third-lowest reading of the year. The cause was mechanical: hashrate had dropped 7.9% in ten days, from roughly 986 exahashes per second (EH/s) on July 1 to 908 EH/s by July 11, pushing average block times out to 10 minutes and 32 seconds, about 5.1% slower than the protocol’s target. Hashprice, the standard measure of miner revenue per unit of computing power, slumped to a floor of $27.6 per petahash per second per day before recovering 12.5% to $31.1 once the difficulty cut took effect, according to Cointribune’s tracking of the adjustment. It was the network’s 14th retarget of 2026, and the eighth negative one.

None of that made front-page news elsewhere. Bitcoin’s price was busy climbing back toward $65,000 the same week on a broader risk-on rally, and headlines went to that instead. But the swing is a useful reminder of something the halving cycle math conversation tends to skip: the supply schedule is fixed and public, a new block roughly every ten minutes, subsidy halving every 210,000 blocks, but the hashrate securing that schedule is not evenly or permanently distributed. It moves, sometimes by tens of exahashes in a matter of days, as miners chase electricity prices, dodge regulators, or get caught in wars and blackouts thousands of miles from any crypto exchange.

With Halving Five estimated for April 17, 2028, at block 1,050,000, roughly 91,862 blocks and 21 months away as of this writing, the more interesting question is not only what the block subsidy will be (1.5625 BTC, down from 3.125 BTC), but who will still be mining it, and from where. This piece maps where Bitcoin’s hashrate physically sits today, how it got there, and why that migration matters more for Halving Five predictions than most price charts do.

The Halving Math, Briefly

Bitcoin’s issuance is coded, not decided. Bitcoin Core’s consensus rules cut the block subsidy in half every 210,000 blocks, a schedule enforced by every full node on the network. The first four halvings landed in November 2012, July 2016, May 2020 and April 2024, cutting the reward from 50 BTC to 25, then 12.5, then 6.25 and finally 3.125 BTC per block. Halving Five, expected around April 2028, drops it to 1.5625 BTC. The schedule runs until roughly the year 2140, when the 33rd and final halving brings the subsidy to zero and the last fraction of the 21 million BTC cap is mined.

As of mid-July 2026, circulating supply sits at about 20.057 million BTC, roughly 95.5% of the eventual total, with Bitcoin trading near $64,700 and a market capitalization around $1.3 trillion. We have covered the forecasting side of this math elsewhere, including how stock-to-flow, power-law and macro-asset models disagree about what April 2028 means for price, and how the post-halving squeeze is already reshaping miner balance sheets. What gets less attention is the other half of the security equation: not how many bitcoin are issued, but which machines, in which countries, are competing to issue them.

Why Geography Became a Predictions Variable

For Bitcoin’s first decade, hashrate geography was mostly a curiosity. Then, in mid-2021, China banned domestic mining outright, and a majority of global hashrate had to relocate within months, one of the best-documented shocks in the network’s history. Kazakhstan absorbed a large share of that diaspora almost overnight, briefly climbing toward the high teens as a percentage of global hashrate before its power grid buckled under the load. Texas absorbed another wave and never gave it back.

That single event taught the industry, and by extension anyone trying to predict Halving Five outcomes, three lessons:

• Hashrate follows the cheapest available electricity with almost no loyalty to any one jurisdiction, because ASICs are mobile, containerized, and can be re-plugged in a new country within weeks.
• A government can shift a double-digit share of a $1.3 trillion network’s security budget with a single policy announcement, as China proved in both directions: banning mining in 2021, then tacitly tolerating its return in the years since.
• The network’s difficulty adjustment absorbs these shocks, as the July 2026 retarget above shows, but does not eliminate the underlying concentration risk. It just reprices it.

Every halving narrows the margin for error a little further. Miners running on thin electricity spreads today will run on thinner ones once the subsidy halves again in 2028, which means the jurisdictions willing to subsidize, tolerate or simply ignore Bitcoin mining before Halving Five are a reasonable preview of who is still mining after it. That is the migration this piece tracks.

The United States: Still the Center of Gravity

The US has led global hashrate share since the China exodus and, if anything, has extended its lead. Hashrate Index’s most recent global hashrate heatmap update puts the country at 37.4% of network hashrate, roughly 375 EH/s, with Texas and Wyoming the dominant hubs thanks to deregulated grids, demand-response programs that pay miners to curtail during peak load, and some of the cheapest industrial power in the developed world.

Policy has reinforced the economics. In March 2025, President Trump signed an executive order establishing a Strategic Bitcoin Reserve, capitalized with forfeited BTC and explicitly framed as a store of reserve assets the government would not sell; independent estimates put existing US government BTC holdings, mostly from criminal and civil forfeitures, at roughly 200,000 BTC, though no full audit has ever been published. A year later, the SEC’s Division of Corporation Finance went further, stating in a formal statement that proof-of-work mining, solo or pooled, does not implicate US securities law, removing a legal question mark that had shadowed the industry since the multi-year enforcement era that preceded the current SEC’s rulemaking-focused approach. For more on how that shift fits together, see our explainer on how the SEC decides what counts as a security in 2026.

The clearest sign of how politically normalized US mining has become is American Bitcoin Corp., the venture launched by Hut 8 alongside Eric Trump and Donald Trump Jr., with Hut 8 contributing its ASIC fleet for an 80% stake and Eric Trump serving as chief strategy officer; the company has since agreed to go public through a merger with Nasdaq-listed Gryphon Digital Mining. None of this fully insulates the US position. A parallel pivot toward AI and HPC hosting, documented in our coverage of the 2026 mining margin squeeze, means some of the same public miners expanding their US footprint are doing so with data centers that may never run a single ASIC. Still, CoinShares’ James Butterfill found the US gained roughly two percentage points of global hashrate share quarter over quarter in the firm’s Q1 2026 mining report, even as the AI pivot accelerated, suggesting the two trends are additive rather than competing for now.

Where the Rest of the Hashrate Sits

Zoom out and the US is not competing against one rival but a long tail of jurisdictions, each betting on a different combination of cheap power, political tolerance and, in a few cases, outright state backing. Hashrate Index’s April 2026 heatmap update put the global top ten at the following shares, with the top three alone accounting for roughly two-thirds of the entire network:

The ranking has barely moved since the previous quarterly snapshot, which is itself informative: despite constant headlines about miners fleeing to new frontiers, the top of the table is sticky. Where the real movement shows up is further down the list, among the countries hovering around 2% to 4%, several of which did not exist as mining hubs five years ago.

Russia’s Legal Paradox

Russia is the network’s second-largest mining jurisdiction by Hashrate Index’s count, at roughly 16.9% of global hashrate, built on Siberian hydropower and stranded natural gas that would otherwise be flared. It is also host to Bitcoin mining’s strangest regulatory story. Moscow legalized crypto mining nationally in a law that took effect in November 2024, then spent the following year layering regional bans on top of that legalization: a temporary, seasonal restriction in several energy-strapped regions hardened into a permanent, multi-year prohibition in Buryatia and Zabaikalsky Krai starting in early 2026, part of a wider ban across ten regions that Russian state media attributed to strained winter power grids.

BitRiver, the country’s largest industrial miner and the anchor tenant of Irkutsk’s power infrastructure, illustrates the squeeze from both directions. Western capital has been closed to it since April 2022, when the US Treasury’s Office of Foreign Assets Control sanctioned BitRiver AG and ten subsidiaries, the first time OFAC had ever sanctioned a crypto mining company, for allegedly helping the Russian state offset the impact of sanctions. Now the company faces pressure from home too: Russia’s Federal Tax Service filed a bankruptcy petition against a BitRiver subsidiary in January 2026, around the same time Buryatia’s mining ban took hold. A jurisdiction can be simultaneously fast-growing and legally isolated at once, which is close to Russia’s actual position heading into Halving Five, and a reminder that legal and investable are not the same thing.

China’s Shadow Return

No jurisdiction captures the limits of banning Bitcoin mining better than China. Beijing outlawed the activity outright in mid-2021, and for a while the ban looked total, with China’s share of global hashrate collapsing toward zero within months. Five years later, it is back to being the network’s third-largest jurisdiction. Hashrate Index puts China at 12.0% of global hashrate as of its most recent update; a Reuters investigation reported by CoinDesk in November 2025 put the figure closer to 14%, and CryptoQuant has floated estimates as high as 15% to 20%, a spread that says as much about how hard illegal mining is to measure as it does about China’s actual footprint. Operations cluster in Xinjiang and Sichuan, where local electricity is cheap, abundant, and, per Reuters’ sourcing, tacitly tolerated by regional authorities even though the national ban remains on the books.

The irony sharpens once you look at hardware rather than hashrate. Bitmain, MicroBT and Canaan, the three firms that manufacture the overwhelming majority of the world’s mining ASICs, are all Chinese companies, meaning China continues to profit from Bitcoin mining as an exporter of machines even where it forbids running them domestically. That dependency is itself a Halving Five variable: any US tariff escalation on Chinese-made mining hardware would raise costs fastest for the American miners currently expanding the most, a policy risk sitting on the opposite side of the ledger from the Strategic Reserve and SEC tailwinds described above.

The Gulf’s Sovereign Bet

If China’s resurgence is a story about tolerance, the United Arab Emirates and Oman are a story about deliberate state participation. Combined, the two account for roughly 6% of global hashrate, according to Hashrate Index, built almost entirely in the past three years. Abu Dhabi-listed Phoenix Group anchors much of it: the company operates more than 550 megawatts of mining and data center capacity across the UAE, Oman, Ethiopia, Canada and the US, including a Masdar City facility contributing over two exahashes that is majority owned by UAE entities alongside a minority stake held by Nasdaq-listed Marathon Digital. A separate venture, Citadel Mining, has built a reported reserve of several thousand BTC through industrial-scale mining on Abu Dhabi’s Al Reem Island, explicitly framed by its backers as a transparent, state-linked alternative to opaque offshore mining capital.

Oman’s entry runs through a similar template, with Green Data City signing a 150-megawatt mining agreement with Phoenix Group that became operational the following year. What both countries share is a willingness to treat mining as sovereign industrial policy rather than a private arbitrage play, backed by state-linked capital and some of the world’s cheapest natural gas. That gives the Gulf model more balance-sheet durability than the frontier plays described next, though it concentrates a growing share of hashrate in the hands of a small number of politically connected operators, a different kind of centralization risk than the one China or Russia represent.

Frontier Power Arbitrage: Ethiopia, Paraguay and Pakistan

Below the Gulf sits a cluster of countries betting stranded or underused power generation on Bitcoin mining, with far less certainty about how long the window stays open. Ethiopia is the most developed example. The Grand Ethiopian Renaissance Dam, inaugurated in September 2025 after 14 years of construction and roughly $5 billion in spending, gave the country a 5,150-megawatt hydropower asset and, alongside it, a mining sector that barely existed five years ago. More than twenty firms, most of them Chinese-backed, have signed power purchase agreements clustered around Addis Ababa, and Ethiopia now sits inside Hashrate Index’s global top ten at roughly 2.5% of network hashrate. The arrangement looks fragile on closer inspection: state utility Ethiopian Electric Power has signaled it wants to phase mining out altogether, and proposed rate increases were projected to make roughly half of existing operations unprofitable by mid-2026 and the large majority by 2027, according to industry reporting on the utility’s own numbers.

Paraguay runs a calmer version of the same trade, monetizing surplus hydropower from the Itaipu Dam it shares with Brazil; CoinShares’ Q1 2026 mining report credits Nasdaq-listed HIVE Digital’s 300-megawatt build there as a driver of the country’s rise into the global top five. Pakistan is the newest and least proven entrant. Islamabad’s Pakistan Crypto Council announced in May 2025 that it would redirect 2,000 megawatts of surplus power, mostly from underused coal plants, toward Bitcoin mining and AI data centers, a plan the IMF promptly complicated by rejecting a request to subsidize the electricity rate. None of Pakistan’s capacity shows up in the hashrate tables yet; whether it ever does is as much a test of the IMF relationship as of the mining economics.

The Kazakhstan Warning

Every frontier-market pitch eventually runs into Kazakhstan’s example. When China’s 2021 ban sent miners scrambling for new homes, Kazakhstan absorbed one of the largest single waves, briefly pushing its share of global hashrate toward the high teens as containers of ASICs arrived faster than the grid could plan for them. The result was rolling blackouts, public backlash, and a government crackdown that included mandatory registration and, eventually, direct rationing. Hashrate Index now puts Kazakhstan’s share at just 1.8%, a fall of roughly ten percentage points from its peak, achieved through a state-run electricity marketplace that caps miners at one megawatt-hour per transaction and gives regulators real-time visibility into who is drawing power and when.

The lesson generalizes: cheap, underused power is often underused for a structural reason, whether that is grid fragility, seasonal hydro variance, or simple lack of transmission capacity, and mining at scale can expose that reason faster than any other industrial consumer. Ethiopia’s utility is already gesturing at the same trade-off Kazakhstan made in 2022. Kazakh authorities have periodically broken up illegal mining networks siphoning power through corrupted utility connections, a reminder that rationing regimes create their own black markets rather than eliminating the underlying demand.

When Sovereigns Become Sellers: Bhutan’s Retreat

Bhutan does not show up in the top-ten hashrate tables, but its Druk Holding and Investments sovereign fund is one of the more closely watched miners precisely because it mines with the state’s own balance sheet, using hydropower the kingdom already had in surplus, in partnership with Nasdaq-listed Bitdeer. Through 2026, on-chain trackers have shown Druk Holding steadily drawing down its Bitcoin position, with CoinDesk reporting in April that the fund had sold roughly 70% of its holdings over eighteen months, alongside signs that active mining itself may have slowed or stopped. Bhutanese officials have pushed back on the scale of the reported sales, in one case telling CoinDesk they could not recall the specific transactions being tracked on-chain, leaving the precise current holding genuinely disputed rather than settled.

What is not disputed is the direction of travel: a government that spent years quietly accumulating Bitcoin through mining is now converting a meaningful share of it into fiat for domestic infrastructure spending. That matters for Halving Five predictions for a specific reason. Sovereign miners were supposed to be the more patient, less rate-sensitive class of hashrate, insulated from quarterly hashprice swings in a way public miners are not. Bhutan’s drawdown suggests even patient capital reassesses when the reward for mining keeps shrinking, which is exactly what happens, mechanically, at every halving.

Pool Concentration: The Layer Migration Doesn’t Fix

Spreading physical hashrate across a dozen countries solves one kind of centralization problem and leaves another almost untouched. Bitcoin miners do not build blocks individually; they join pools that aggregate hashrate and smooth out payouts, and those pools, not the underlying countries, are what actually assemble block templates under the still-dominant Stratum V1 protocol. A June 2026 review by mining-hardware firm D-Central found the Nakamoto coefficient for Bitcoin mining pools sitting at just 3, meaning only three pools were needed to exceed half of all blocks produced, with Foundry USA alone controlling roughly 27% of network hashrate. A US-based miner and a Chinese-based miner can both be pointing their machines at the same American pool operator, so the country-level diversification described throughout this piece coexists with real concentration at the pool layer.

That concentration has already produced a censorship flashpoint. F2Pool, one of the larger pools by share, was flagged again in early 2026 by a Bitcoin developer for filtering transactions tied to OFAC-sanctioned addresses, echoing a similar finding first reported in 2023. Marathon briefly ran an explicitly OFAC-compliant pool in 2021 and abandoned it within a month once it proved commercially unworkable. The industry’s actual fix is technical rather than geographic: Stratum V2’s Job Declaration feature lets individual miners build their own block templates and simply have the pool validate and smooth payouts, removing the operator’s ability to unilaterally filter transactions. Seven pools representing roughly three-quarters of hashrate, including Foundry, AntPool and F2Pool, had joined a Stratum V2 working group by May 2026, and the first fully production Job Declaration block was mined in June. The contrast with proof-of-stake is instructive: as we have written in our explainer on how stakers actually get paid, Ethereum’s validators do not need a single physical location at all, which removes the geography question that dominates every Bitcoin mining forecast but reintroduces a different one, around staking-service and custody concentration instead.

That reshuffling is not hypothetical. SBI Crypto, the mining arm of Japan’s SBI Group, announced it will permanently shut down its public mining pool on July 31, 2026, a deadline just weeks away as of this writing, forcing roughly 20.9 EH/s, about 2.2% of global hashrate, to find a new pool operator almost overnight. SBI offered no public explanation beyond pointing miners toward alternative pools without formally endorsing any of them, though the move follows SBI Holdings’ roughly $289 million deal to acquire crypto exchange Bitbank. It is a small-scale preview of what happens whenever a mid-sized pool exits: hashrate does not disappear, it reshuffles toward whichever remaining operator offers the best mix of low fees, reliable payouts and, increasingly, Stratum V2 support, further concentrating the pool layer even as country-level mining keeps diversifying.

The Efficiency Counter-Argument

Not everyone reads this migration as a permanent feature of Bitcoin mining. Every hardware generation narrows the gap that makes chasing marginal power worthwhile in the first place. Bitmain’s current-generation Antminer S21 XP rates around 13.5 joules per terahash, a fraction of the power draw per unit of hashrate that machines from the pre-2021, China-dominated era needed. Each efficiency jump like this reduces the incentive to relocate to the cheapest, least stable grid available and increases the relative value of political stability, reliable uptime and proximity to capital markets, all of which favor the US and, to a lesser extent, Canada and the Gulf over frontier plays like Ethiopia or Pakistan.

That does not make geographic migration disappear, it changes what migration is for. Increasingly, relocation decisions are driven less by where power is cheapest and more by where a facility can also host AI or HPC workloads when Bitcoin economics are weak, which is precisely the calculation behind the overlapping mining-and-AI strategies documented in our miner shakeout coverage above. If that trend holds through Halving Five, the country-level hashrate table earlier in this piece may prove more stable over the next two years than the 2021 to 2023 period was, even as the reasons individual facilities exist keep changing underneath it.

What This Means for Halving Five

Put the pieces together and a few predictions become easier to state with some confidence, while a few stay genuinely open. The subsidy cut itself is not in question: barring a consensus change with essentially no developer support, block rewards fall from 3.125 to 1.5625 BTC at block 1,050,000, and fees still cover under 5% of miner revenue, so whoever is mining in 2028 will be defending the network on roughly half of today’s issuance-based budget. What is in question is whether that budget gets defended by a more geographically resilient network than the one that endured the 2021 China ban, or a more fragile one.

The case for resilience: no single country holds a majority any longer, the top three (the US, Russia and China) sit at a combined 66%, well below the roughly 75% China alone once represented, and genuine new entrants such as the Gulf states, Ethiopia and Paraguay have proven the model of monetizing stranded power is repeatable outside any one region. The case for fragility: two of the top three jurisdictions carry serious legal overhangs, Russia’s sanctions exposure and shifting regional bans, China’s outright ban, pool concentration remains far tighter than country concentration with a Nakamoto coefficient of just 3, and several of the fastest-growing frontier markets, Ethiopia most acutely, are already showing signs that the cheap-power window closes faster than miners can amortize their hardware.

Standard Chartered’s Geoffrey Kendrick, who called the cycle low in June 2026 (“Winter is over. Welcome back to crypto Spring,” he told CoinDesk), and Morgan Stanley’s Denny Galindo, who has framed the current period as the four-year cycle’s “fall” phase in the firm’s research, are both describing the same demand-side story from different angles. Neither addresses hashrate directly, but the mining data fills the gap: margin compression during a fall phase is exactly when mining consolidates toward the cheapest, most efficient operators, which is what the country and pool data above already show happening. Halving Five will not resolve that tension on its own. It will simply lower everyone’s margin at the same time, the way it did in 2024, which historically accelerates whatever consolidation was already underway rather than reversing it. That is the same dynamic driving the miner shakeout among public US operators, and the mirror image of the demand-side question we have examined separately in looking at what ETF flows actually change: issuance is supply, hashrate is the cost of defending that supply, and both are now shaped as much by policy and geography as by code.

Frequently Asked Questions

When will the next Bitcoin halving (Halving Five) happen?

Halving Five is currently estimated for around April 17, 2028, at block height 1,050,000, based on Bitcoin’s average 10-minute block time. The exact date shifts slightly as actual block production runs faster or slower than that average; as of mid-July 2026 the network was roughly 91,862 blocks away. When it arrives, the block subsidy drops from 3.125 BTC to 1.5625 BTC.

Which country has the most Bitcoin mining hashrate?

The United States, with roughly 37% of global Bitcoin hashrate as of mid-2026, according to Hashrate Index. Russia and China follow at a distant second and third, at around 17% and 12% respectively, meaning the top three jurisdictions together account for roughly two-thirds of the entire network.

Is Bitcoin mining still happening in China even though it’s banned?

Yes. China banned domestic Bitcoin mining in 2021, but underground and locally tolerated operations, concentrated in Xinjiang and Sichuan, have brought the country back to being the network’s third-largest mining jurisdiction, with estimates ranging from about 12% to as high as 20% of global hashrate depending on the source. China also remains the manufacturing base for most of the world’s mining hardware.

Does mining decentralization affect Bitcoin’s security?

Geographic spread of hashrate reduces the risk that any single government can disrupt the network the way China’s 2021 ban briefly threatened to. But geography is only one layer: block templates are still built mostly by mining pools, and just three pools currently control enough hashrate to produce more than half of all blocks, a tighter concentration than the country-level data alone suggests.

What happens to Bitcoin mining profitability after the next halving?

Mining revenue from the block subsidy is cut in half overnight, from 3.125 BTC to 1.5625 BTC per block, while transaction fees currently cover under 5% of miner income, so most of the lost revenue is not replaced by fees. Historically this squeezes out higher-cost operators first, accelerating consolidation toward the cheapest power and most efficient hardware, a pattern already visible in this cycle’s hashprice declines.

By the HOGE Wire markets desk, July 15, 2026.