The charts will tell you that Bitcoin's mining centralization just got worse—that the closure of SBI Crypto's pool in late July was another nail in the coffin of Nakamoto's vision. The three largest pools now command more than 60% of all attributed blocks, and the reflexive conclusion writes itself: the network is consolidating, the giants are swallowing the small, and decentralization is a fading memory.
The charts are not wrong about the number. They are wrong about the causality.
The 60% threshold was crossed on July 20, while SBI was still mining. It was crossed again on July 27, two days before the pool produced its final block. The concentration did not arrive with SBI's departure. It was already seated at the table, waiting for the empty chair to be filled. Tracing the silent currents beneath the market, this is the current that matters: not the exit itself, but the structural reality that made the exit a statistical non-event.
This is not a comforting conclusion. It is, perhaps, a more uncomfortable one.
SBI Crypto was never a giant. At its operational peak, the pool commanded roughly 2% of global hash rate—a modest but meaningful presence that offered geographic diversity. Japan is not Texas, and it is certainly not China. A mining pool operated by a listed Japanese financial conglomerate carried symbolic weight beyond its raw hash power. It was evidence, however thin, that mainstream institutional capital could participate in the messy business of Proof-of-Work.
The pool's death was not a sudden collapse. It was a month-long exodus that the telemetry data captures with clinical precision.
On June 30, SBI's 7-day average hash rate stood at 16.222 EH/s. By July 30, that figure had fallen to 5.817 EH/s—a 64% monthly decline. On July 31, the 24-hour average registered 0.452 EH/s, a shadow of its former self. The pool stopped producing blocks after July 29. When SBI formally disconnected its Stratum services, the pool was already a corpse; the official shutdown merely pronounced the time of death.
For the miners still connected, the transition was almost trivial. Stratum is the communication protocol that connects miners to pools, and reconfiguring it means changing a connection URL and a worker credential. I have watched this migration happen dozens of times in my career. What looks like an infrastructure crisis from the outside is, for the individual miner, an afternoon of editing configuration files. The technical switching cost is near zero; the economic switching cost is what actually matters.
Let me be precise about what changed and what did not. The Bitcoin protocol—its Proof-of-Work consensus, its difficulty adjustment algorithm, its UTXO accounting model—was untouched by this event. SBI's exit did not alter a single line of code in the base layer. What changed was the operator of a Stratum aggregation service, a piece of infrastructure that exists between the miner and the chain. That distinction matters because it separates the network's actual security properties from the commercial landscape of its service providers.
The 0.452 EH/s that remained at the end represented roughly 0.07% of the network's total hash rate. This is the first number worth holding onto: the entire SBI shutdown, from first bleed to final block, removed less than 2.5% of the network's global hash rate at any given moment. The network absorbed it without a ripple. Difficulty adjustments continued on schedule. Block times held. The consensus layer never noticed.
This is the structural truth that gets lost in the centralization panic: mining pool consolidation is a phenomenon of the service layer, not the protocol layer. Bitcoin's security model does not depend on who operates Stratum servers. It depends on the total energy committed to solving the SHA-256 puzzle, and that energy barely moved.
The most-cited data point in the aftermath is the claim that the top three pools—Foundry USA at 26.67% of attributed blocks, AntPool at 17.13%, and F2Pool at 16.21%—now control just over 60% of the network. The number induces vertigo in anyone who has spent years defending Bitcoin against the accusation that it is not truly decentralized.
But the number deserves forensic scrutiny. The audit reveals what the algorithm omits.
Pool share statistics on platforms like Hashrate Index are not measurements of actual hash rate. They are measurements of attributed blocks—blocks found and publicly credited to a pool's public identifier. Attribution is not measurement. A miner who connects to Foundry through a private relay may contribute hash power that gets credited to an entity that is not their actual operator. Conversely, a pool that temporarily routes its miners' traffic to another pool for stability reasons can see its attributed share collapse even while its real hash power remains constant.
This is not a trivial methodological caveat. It is the difference between a snapshot and a photograph. The 60.01% reading that generated the headlines captures a single instant in a moving system. The week prior, the same metric recorded 64.8%. A week of variance of nearly four percentage points suggests that attribution data is noisy—influenced by block-finding luck as much as by genuine hash power migration.
That said, I do not want to overcorrect. Even accounting for statistical noise, the directional signal is real. The top three pools have consistently commanded a majority of attributed blocks for years. This is a genuine feature of the current mining landscape, and it deserves serious analysis rather than dismissal.
The uncomfortable truth is that the concentration was already baked in before SBI's exit—which means SBI's exit explains nothing about the concentration. The narrative that a Japanese pool's departure caused Bitcoin's centralization is chronological fiction.
This is where the macro lens comes into focus. The reflexive explanation—'SBI left because mining is becoming too concentrated'—inverts the causal chain. SBI did not leave because the big pools were winning. The big pools were winning because the economics of mining had already turned against operators like SBI.
Consider the timing. The 2024 halving cut the block subsidy from 6.25 BTC to 3.125 BTC. Mining pools derive their revenue from a cut of block subsidies and transaction fees—typically between 1% and 4% of a miner's output. When the subsidy halves, a pool's gross revenue per unit of hash power halves with it, unless transaction fees rise to compensate. They have not risen sufficiently. The result is a brutal squeeze on pool margins, and that squeeze is felt most acutely by mid-sized operators with higher operating costs.
This is not a Ponzi structure; it is important to state this plainly. Bitcoin miner revenue comes entirely from real network payments—block subsidies and transaction fees. The incentive model remains intact. But the cost side of the equation is not uniform across geographies.
SBI's parent company is Japanese. Japan's industrial electricity rates are among the highest in the developed world, and the yen's weakness has compounded the problem for any operation with international cost inputs. What we witnessed in July was not a strategic retreat. It was the delayed response to an economic equation that stopped balancing.
Liquidity is a mirage; reality is in the reserve. In mining, the reserve is the margin between mining cost and revenue. When that margin compresses toward zero, operators do not need a dramatic event to leave. They just stop renewing their power contracts, stop upgrading their ASIC fleets, and quietly let their hash rate decline. That is exactly what the SBI telemetry shows: a slow bleed that became an avalanche, consistent with an operator letting existing contracts expire rather than actively liquidating infrastructure.
There is a second-order effect worth naming here. During the month-long decline, SBI's remaining miners were not sitting idle. They were reconfiguring their Stratum connections and pointing their hashers at other pools. The official telemetry only reflects hash power attributed to SBI's service; it cannot show where that hash power went. Based on patterns I have observed in similar exits—including the post-2021 migration out of Chinese pools—the majority likely migrated to the top three, which offer institutional-grade stability and faster payment cycles. If that reading is correct, SBI's exit did not merely fail to cause concentration. It accelerated it.
The pool rankings tell a story that the headlines missed. Foundry's 26.67% share makes it the clear leader, built on a foundation of American regulatory compliance and institutional-grade custody and financial services. AntPool's 17.13% reflects deep ties to Asian hardware channels and diversified financial products. F2Pool, at 16.21%, is the veteran—the old guard that has survived every cycle since 2013.
But look at the tail. Luxor is rising, having built its brand on data services and hashrate derivatives rather than raw pool infrastructure. Braiins, the open-source pioneer that gave the world its first custom mining firmware, is losing share. NeoPool has disappeared from recent rankings entirely. This is not a story of simple centralization. It is a story of the middle being crushed.
Patterns emerge when we stop watching the price. The pattern here is bifurcation. The mining industry is splitting into two layers: a handful of large, compliant, institutionally-backed pools at the top, and a set of agile, data-driven specialists at the margins. The middle layer—pools that are too big to be nimble and too small to compete on institutional trust—is the layer being eliminated. That middle layer was SBI's home.
The economic consequence of this bifurcation is already visible in the fee market. Pools are competing for shrinking margins, offering fee rates between 1% and 4% of a miner's output, and differentiating on payment speed, transparency, and ancillary services like hashrate derivatives. For miners, the near-term effect is ambiguous. More competition among fewer, larger pools could push fees down and benefit the miner. But concentrated pools also gain bargaining power over time, and that power tends to express itself in policy choices rather than fee schedules.
Let me now make an argument that will annoy both the Bitcoin maximalists and the critics.
The maximalists insist that pool concentration is irrelevant because miners can switch pools at any moment. This is true in theory and steadily less true in practice. Pool operators exercise meaningful control over transaction selection and block template policy. A pool that chooses to exclude certain transaction types—whether Ordinals inscriptions, BRC-20 transfers, or addresses subject to sanctions—can effectively censor those transactions from the blocks it produces. As the top three pools accumulate more hash power, the practical cost to an individual miner of disagreeing with pool policy rises. The threat of 'vote with your hashrate' becomes less a free choice and more a coerced one.
The critics, meanwhile, insist that the 60% figure represents a fatal compromise of Bitcoin's security. This is mathematically sloppy. The 60% threshold only matters if the pools collude on an attack, and collusion is not the same as concentration. Foundry, AntPool, and F2Pool are commercial competitors with different jurisdictional exposures, different business models, and different relationships to the broader crypto economy. They have no history of coordinated behavior. A theoretical 51% attack requires unified intent, not just statistical proximity.
The deeper truth is that pool concentration, while not fatal, is a genuine degradation of the system's trust-minimization properties. It is a slow-moving problem, not a sudden crisis. And it is a problem that SBI's exit did not create and does not even meaningfully worsen.
What SBI's exit does reveal is the real vulnerability of the mining industry: the asymmetry between hash power and operational diversity. The network's resilience is not measured by the number of pools. It is measured by the number of independent economic actors with the capital, energy access, and technical competence to run mining operations. That number is declining. SBI's departure is one data point in that decline. NeoPool's absence is another. Braiins' shrinking share is a third. This is the current that runs beneath the market, and it has been running for years.
I came to this conclusion through a painful lesson. In 2020, when I was analyzing the fragility of algorithmic stablecoins, I calculated a fragility index of 0.85 and warned that the system was on the verge of collapse. The market ignored me, drunk on 300% APYs, until the crash validated the model. What I learned from that experience is the distance between technical reality and market sentiment. The same distance is visible here: the market treats SBI's exit as a centralization event, when the technical reality is that it is a cost-structure event with modest concentration side effects.
So where does this leave the careful observer? Not in panic, and not in complacency.
The market impact of SBI's exit was predictable and minimal. Mining infrastructure news rarely moves BTC price, and this was no exception. But for those of us who position across cycles, the signal is elsewhere.
The real risk to watch is not the 60% share of today. It is the block-template policy of the top pools tomorrow. The number that will matter in the next cycle is not hash rate concentration but policy concentration—whether the entities that dominate block production converge on a single standard for what transactions are valid. If that happens, the mining pools become the regulatory enforcement layer of last resort, and Bitcoin's censorship resistance weakens in a way that no amount of difficulty adjustment can fix.
The second signal to watch is fee-market dynamics. The post-halving squeeze that killed SBI is still grinding. The winners will be those who offer miners the best combination of stability, transparency, and value-added services. The miners, for their part, should treat this moment as an opportunity to demand better terms. Hash power migration remains the most powerful governance tool miners possess, and they have just demonstrated—unwittingly—how quickly they can use it.
SBI Crypto's shutdown was not a shock, a cause, or a turning point. It was a symptom—a mirror held up to an industry whose economics had already shifted. The charts will tell you that decentralization is dying. The data tells you that it changed shape. The question for the next cycle is not whether the giants rule. It is whether the network can survive the silence of the middle.