Mining Centralization

What Is Mining Centralization?

Mining centralization describes the concentration of Bitcoin's proof-of-work mining power among a small number of entities, whether those are mining pools, geographic regions, or hardware manufacturers. While Bitcoin was designed so that anyone with a computer could participate in mining, the industrialization of the process has created significant concentration across multiple layers.

This concentration matters because Bitcoin's security model relies on the assumption that no single party or colluding group controls a majority of hashrate. When mining power consolidates, risks emerge: transaction censorship becomes feasible, 51% attacks become cheaper to coordinate, and the network's promise of permissionless, censorship-resistant transactions weakens.

Mining centralization is not a binary state but a spectrum. The Bitcoin network has never been perfectly decentralized, and the degree of centralization fluctuates as economics, regulation, and technology shift the landscape.

The Three Dimensions of Centralization

Pool-Level Centralization

The most visible form of mining centralization occurs at the pool level. As of mid-2026, the Bitcoin mining landscape is dominated by a handful of large pools:

The top two pools alone control roughly 45-50% of total network hashrate, and the top four pools exceed 70%. The Nakamoto coefficient for Bitcoin mining stands at approximately 3: only three pools would need to collude to command a majority of hashrate. This concentration exists despite there being over 20 identifiable pools on the network.

Pool shares fluctuate week to week. Foundry USA's share, for example, has been reported as high as 37% in early 2026 and as low as 24% later in the year, reflecting both real fluctuation and different measurement windows.

Geographic Centralization

Before China's mining ban in mid-2021, Chinese miners controlled an estimated 65-75% of global hashrate. The ban triggered the largest hashrate migration in Bitcoin's history, redistributing mining power across new jurisdictions. The current geographic distribution (mid-2026):

The top three countries (US, Russia, China) host approximately 65% of global hashrate. Geographic concentration creates systemic risk: in January 2025, winter storm "Fern" caused Foundry USA to lose roughly 60% of its hashrate (approximately 200 EH/s went offline) as US miners curtailed operations to relieve grid pressure. Block intervals exceeded 12 minutes during the event, illustrating how geographic concentration in a single jurisdiction can temporarily degrade the network.

Emerging mining hubs include Ethiopia, Oman, and Kyrgyzstan (which saw over 300% annual growth in hashrate). These regions benefit from cheap energy and, in some cases, state-supported mining initiatives.

Hardware Manufacturer Centralization

The ASIC manufacturing market is an extreme oligopoly. According to the Cambridge Digital Mining Industry Report (2025), three manufacturers control over 99% of the Bitcoin ASIC market:

This concentration creates supply chain risk: firmware backdoors, production delays, or export restrictions affecting a single manufacturer could impact the vast majority of mining hardware. Intel attempted to enter the market with its Blockscale chip in 2022 but exited just one year later. Auradine, a US-based startup, began shipping competitive 3nm ASIC miners in late 2024, though its market share remains marginal compared to Bitmain.

Why Miners Join Large Pools

Pool centralization persists because of economics, not ignorance. At current difficulty levels (approximately 133 trillion as of September 2026), a single Antminer S21 XP running at 270 TH/s has roughly a 1-in-24,000 chance of finding a block on any given day. The expected time for that miner to find one block solo: approximately 19,000 years. Even a farm with 1,000 such miners would expect one block every ~19 years.

By joining a pool, miners contribute hashrate collectively and receive regular, proportional payouts rather than gambling on extremely rare solo block discoveries. The dominant payout method, FPPS (Full Pay-Per-Share), pays miners for every valid share based on the statistical expected value of their contribution, including estimated transaction fees. The pool absorbs all variance risk.

Larger pools offer additional advantages: lower variance in payouts (the statistical smoothing improves with pool size), professional infrastructure, and the ability to offer FPPS payouts at competitive fee rates (typically 2-4%). Smaller pools using PPLNS (Pay-Per-Last-N-Shares) charge lower fees but expose miners to block-finding variance, making them less attractive to miners who need predictable cash flow.

Pool Operators vs. Individual Miner Sovereignty

A critical distinction in understanding mining centralization: pool hashrate is not the same as a single entity's hashrate. A pool with 30% of the network consists of thousands of independent miners who could switch pools at any time. However, this theoretical freedom obscures a practical problem.

Under the Stratum V1 protocol (still used by the vast majority of miners), pool operators have exclusive control over block templates. They decide which transactions from the mempool to include in a block and in what order. Individual miners simply receive work units and compute hashes: they have zero visibility into or control over which transactions they are validating.

This means pool operators can censor transactions, and documented incidents confirm this has happened:

• In May 2021, Marathon Digital launched a pool using wallet risk-scoring software that filtered out transactions involving OFAC-sanctioned addresses. After significant community backlash, Marathon reversed the policy later that year.
• In November 2023, researcher 0xB10C detected multiple Bitcoin transactions from OFAC-sanctioned addresses missing from blocks, with the majority mined by F2Pool. F2Pool acknowledged the filtering and stated the filter would be removed after community pressure.

With a Nakamoto coefficient of 3, if three pool operators coordinated to refuse building on blocks containing certain transactions, they could orphan non-compliant blocks and make targeted transactions permanently unconfirmable. Individual miners contributing to those pools would be unknowing participants in that censorship. For a deeper analysis of these risks, see the research article on mining pool centralization risks.

Stratum V2 and the Path to Miner Sovereignty

Stratum V2 is the most significant technical effort to address pool-level centralization. Originally developed by Braiins (operators of the former Slush Pool, the first-ever Bitcoin mining pool), Stratum V2 introduces several improvements over the original protocol:

• Mandatory encryption via the Noise Protocol framework, preventing hashrate hijacking and eavesdropping that are possible with V1's plaintext JSON-RPC
• Binary framing that reduces bandwidth usage by roughly 70%, with block change latency dropping from approximately 325ms to under 2ms
• Job Declaration Protocol: the key decentralization feature, allowing miners running their own full node to construct block templates locally and submit them to the pool for validation

With Job Declaration, individual miners select which transactions to include from their own mempool. The pool validates the template's correctness but cannot dictate its contents. This preserves the economic benefits of pooled mining while returning transaction selection sovereignty to individual miners.

Adoption has accelerated through 2026. In May 2026, seven major pools representing approximately 75% of global hashrate joined the Stratum V2 Working Group, including AntPool, F2Pool, Foundry, and MARA Foundation. On June 25, 2026, GoMining and DMND pool mined block 955,318 using Job Declaration: the first known production block where a miner constructed its own template via Stratum V2.

An important caveat: most miners currently connecting via Stratum V2 use only the encrypted transport and faster job delivery features. The percentage of miners actually running Job Declaration (constructing their own templates) remains small. True template decentralization is technically possible but not yet widely adopted.

OCEAN and DATUM

Alongside Stratum V2, the OCEAN pool (founded by Bitcoin Core developer Luke Dashjr in November 2023 with a $6.2M seed round led by Jack Dorsey) offers an alternative decentralization approach via its DATUM protocol (Decentralized Alternative Templates for Universal Mining). DATUM requires miners to run their own full node and construct block templates locally. The pool receives cryptographic proof of work but never sees the actual transactions being mined.

OCEAN is also the only fully non-custodial major pool: mining rewards are paid directly from the coinbase transaction to miner addresses. While OCEAN's share of total hashrate remains small (approximately 2%), it represents an important proof of concept for decentralized mining pool architectures.

Why It Matters

Mining centralization is not merely a theoretical concern. It directly impacts Bitcoin's core value propositions:

• Censorship resistance: concentrated pool control over block templates means a small number of operators can filter transactions, undermining Bitcoin's permissionless nature
• Geographic resilience: when one country hosts a disproportionate share of hashrate, natural disasters, grid failures, or regulatory actions in that jurisdiction affect the entire network
• Supply chain security: dependence on a single ASIC manufacturer creates a chokepoint that could be exploited through firmware manipulation, export controls, or production disruptions
• Regulatory pressure: concentrated mining operations are easier for governments to regulate or coerce, potentially turning miners into agents of state-level transaction surveillance

For Bitcoin layer 2 protocols like Spark and the Lightning Network, mining centralization matters because these systems ultimately rely on the base layer's security and censorship resistance for their own guarantees. If a user needs to broadcast a force-close transaction and miners are censoring it, the layer 2's safety model breaks down. For a comprehensive look at how hashrate has shifted geographically, see the research article on Bitcoin hashrate geographic redistribution.

Risks and Mitigations

Current Risks

• Transaction censorship by pool operators is technically feasible today and has been documented in practice
• The Nakamoto coefficient of 3 means only three colluding pools could execute a 51% attack or sustained censorship campaign
• Geographic concentration in the US creates a single point of regulatory and environmental risk
• ASIC manufacturer dominance by Bitmain introduces supply chain fragility
• The ongoing migration of mining capacity to AI computing (hashrate declined roughly 19% from November 2025 to mid-2026) could accelerate concentration among larger, better-funded operators

Mitigations in Progress

• Stratum V2 Job Declaration and OCEAN DATUM return block template control to individual miners
• Geographic diversification is occurring naturally as emerging regions (Ethiopia, Oman, Paraguay) grow their mining industries
• New ASIC entrants like Auradine reduce single-manufacturer dependency
• Open-source mining firmware (such as Braiins OS) reduces reliance on manufacturer-controlled software
• Community monitoring tools (like 0xB10C's miningpool-observer) provide transparency into pool transaction selection behavior

This glossary entry is for informational purposes only and does not constitute financial or investment advice. Always do your own research before using any protocol or technology.