Four Pools Control 70% of Bitcoin Hashrate, Nakamoto Coefficient Hits 3
Four mining pools held more than 70% of Bitcoin's hashrate in a June 2026 snapshot. With a Nakamoto coefficient of 3, pool concentration has moved from theoretical concern to operational baseline.
Bitcoin's block production is now concentrated enough that three pool operators could coordinate to exceed 51% of the network.
Key takeaways
- Foundry Digital (~31%), AntPool (~18%), ViaBTC (~13%), and F2Pool (~10%) collectively held more than 70% of Bitcoin's hashrate in a June 23, 2026 snapshot per miningpoolstats.stream.
- D-Central's H1 2026 mining report put Bitcoin's Nakamoto coefficient at 3, meaning only three pools are needed to mine more than half of all blocks, with Foundry USA alone producing roughly 27% of them.
- The structural fix exists: Stratum V2 with job negotiation hands block-template construction back to individual miners. AntPool, F2Pool, and ViaBTC had not deployed it in production as of June 2026, despite joining the Stratum V2 Working Group on May 7.
Four Bitcoin mining pools controlled more than 70% of the network's hashrate as of a June 23, 2026 snapshot from miningpoolstats.stream, with D-Central's concurrent H1 2026 report placing the network's Nakamoto coefficient at 3. That number is not an abstraction: it means three pool operators, acting together, can control the majority of Bitcoin's block space today.
The June 23 split: Foundry Digital at 31%, AntPool at 18%, ViaBTC at 13%, F2Pool at 10%. A more recent 7-day window from Simple Mining (updated July 17, 2026) shows the positions have shifted but the concentration has not: Foundry at 27.6%, F2Pool at 17.8%, AntPool at 17.3%, ViaBTC at 9.5%, SpiderPool at 5.7%. Different order, same story.
The data was first aggregated in a CryptoSlate partner article published July 8, 2026.
This Is a Censorship-Resistance Problem, Not a Hashrate Metric
Hashrate is distributed across thousands of machines worldwide. Block templates are not.
Under Stratum V1, the protocol all four dominant pools still run, miners contribute computation but the pool operator constructs every block template. The pool decides which transactions get included. The miner has no say. Point your ASICs at Foundry, AntPool, F2Pool, or ViaBTC today and you are lending your hashrate to an entity that controls what gets confirmed.
Foundry is US-based, backed by Digital Currency Group, and built explicitly for institutional and publicly traded mining companies with strict KYC requirements. That description carries real weight for censorship resistance: a single KYC-gated, US-regulated entity producing roughly 27-31% of all Bitcoin blocks means one regulatory request, one OFAC pressure letter, one bad-faith coordination event across three operators is the entire attack surface.
The D-Central H1 2026 report puts Bitcoin's Nakamoto coefficient at 3. Foundry alone sits within a rounding error of 30% on any given week. Three pools clear 51%.
This same concentration is producing a practical market problem for independent miners. Pools optimized for institutional fleets treat smaller operators as edge cases. ViaBTC, holding roughly 9-13% of hashrate depending on the window, has reportedly faced regulatory scrutiny affecting miners in Russia and CIS countries, including reported account restrictions and fund freezes, per the CryptoSlate partner reporting.
When your payout account gets frozen by a pool responding to jurisdictional pressure, the decentralization argument becomes very concrete, very fast. The SBI Crypto pool shutdown earlier this summer is another data point in the same direction: pool dependency is a single point of failure.
There is also the state-level angle. When a government decides it wants mandatory pool control, the attack vector is exactly this concentration. Fewer pools means fewer entities to coerce.
The Fix Exists. The Deployment Does Not.
Stratum V2 with job negotiation inverts the template problem. Instead of the pool assembling the block and handing it to miners, the miner proposes the template and the pool validates and submits it. Even a pool that wanted to filter transactions could not do so against a miner running job negotiation, as long as the pool accepts miner-proposed templates.
On May 7, 2026, seven pools including Foundry, AntPool, F2Pool, SpiderPool, MARA Pool, Block Inc., and DMND joined the Stratum V2 Working Group, collectively representing approximately 75% of global hashrate at the time, per spark.money research. That is directionally correct.
Per D-Central's Stratum Support Matrix from June 2026, AntPool, F2Pool, and ViaBTC had not deployed Stratum V2 with job negotiation in production. Working Group membership is not a deployed protocol. The pools that have moved, OCEAN (running DATUM) and Braiins Pool, are the only options today where a miner retains template control.
If the three holdouts actually ship production V2 with job negotiation enabled, the censorship concern weakens materially. Until then, the commitment is just a commitment.
For miners operating outside institutional rails, the choice of pool is not a preference, it is a protocol decision about who constructs the block.
What to Watch
The Stratum V2 Working Group's production deployment timeline is the number that matters. If AntPool, F2Pool, and ViaBTC ship V2 with job negotiation enabled within the next 12 months and adoption reaches meaningful hashrate share, the pool concentration problem shifts from existential to manageable. If the Working Group commitment quietly stalls and the four-pool oligopoly holds through 2027, the network's censorship-resistance posture remains exactly what the June 2026 data shows: three phone calls away from a majority.
Sources
- miningpoolstats.stream Bitcoin pool data
- D-Central: State of Bitcoin Mining, H1 2026
- D-Central: Stratum Protocol Support Matrix, June 2026
- Simple Mining: Bitcoin mining pool rankings, updated July 17, 2026
- squaredtech.co: Bitcoin Mining Pools in 2026 (originating partner article)
- Stratum V2 specification repository
Frequently Asked Questions
Not on its own, but the math is uncomfortably close. No single pool currently holds 51% of hashrate. The Nakamoto coefficient of 3 means three pools collectively exceed that threshold. A coordinated 51% attack would require those three operators to collude and redirect their combined hashrate, an economically irrational move that would destroy the value of the asset they profit from.
The more live risk is transaction censorship: under Stratum V1, pool operators can exclude specific transactions from every block they produce without triggering a 51% alarm, silently and indefinitely.
Stratum V2's job negotiation feature moves block-template construction from the pool to the individual miner. Under V2 with job negotiation enabled, a miner proposes which transactions go in the block. The pool validates proof of work and submits the block, but it no longer controls the content. Even if a pool operator faced regulatory pressure to filter transactions, a miner using job negotiation could include those transactions anyway.
The critical caveat: the pool has to accept miner-proposed templates. If the pool silently rejects templates containing certain transactions, the censorship problem resurfaces at a different layer. Production deployment with genuine job negotiation is the threshold, not Working Group membership.
Point some hashrate at a pool that already has miner-side template construction in production. OCEAN (running the DATUM protocol) and Braiins Pool (Stratum V2) are the only major options as of mid-2026. Even a partial redirect reduces concentration at the dominant four pools and sends a market signal. The fee structures and payout models differ, so the economics need to check out per operation, but the decentralization benefit is direct and immediate.
