What Happens When All 21 Million Bitcoins Are Mined?

Bitcoin's monetary policy is enforced entirely in code. Every 210,000 blocks — roughly every four years, given the network's ~10-minute average block time — the block subsidy paid to miners for finding a valid block is cut in half. It started at 50 BTC per block in 2009, dropped to 25 in 2012, 12.5 in 2016, 6.25 in 2020, and 3.125 after the April 2024 halving. This geometric decay means the total supply approaches 21 million asymptotically rather than stopping abruptly; the subsidy keeps halving roughly every four years until it reaches zero satoshis around block height 6,930,000, expected near the year 2140.

Because each halving cuts the new-supply issuance rate in half while (all else equal) demand and hashrate trends continue independently, the event has historically been a major inflection point for miner economics — sharply reducing revenue per unit of hashrate overnight and forcing less efficient hardware out of profitable operation.

As of the current 3.125 BTC subsidy era, the next halving (expected around 2028) will cut the reward to 1.5625 BTC, then 0.78125 BTC around 2032, and so on. Each cycle, the subsidy becomes a smaller share of total network revenue, assuming transaction fee volume continues to grow. This is a known, non-negotiable schedule baked into the Bitcoin protocol — there's no scenario in which it changes without a coordinated, contentious hard fork that the vast majority of the network would need to accept, which has essentially never happened for a change this fundamental.

For miner operators, this means profitability modeling should never assume a static subsidy. A machine bought today running the numbers against the current 3.125 BTC reward will see its revenue-per-hash from subsidy alone cut in half again within a few years, regardless of Bitcoin's price. Efficiency — the J/TH figure discussed in our piece on ASIC miner performance factors — becomes increasingly important with every halving cycle, since less efficient hardware is disproportionately squeezed out as revenue per block falls.

Bitcoin's design always anticipated that transaction fees would eventually become the dominant, and ultimately sole, source of miner revenue once the subsidy trends toward zero. Fees are paid by users to have their transactions included in a block, and they fluctuate based on network congestion and demand for block space — during periods of high on-chain activity, fee revenue has occasionally spiked to rival or exceed the block subsidy for short windows, demonstrating that a fee market can generate meaningful revenue when there's enough demand for scarce block space.

Whether fee revenue alone will be sufficient to secure the network at current hashrate levels a century from now is genuinely unresolved and depends on Bitcoin's future transaction volume, adoption as a settlement layer, and the development of layer-2 systems that batch many transactions into fewer on-chain settlements. It's an open economic question, not a solved one, and it's reasonable for long-horizon investors to treat it as a real source of uncertainty rather than a guaranteed outcome.

Bitcoin's difficulty adjustment algorithm recalculates roughly every two weeks to keep block times near ten minutes regardless of how much total hashrate is competing on the network. If miner revenue falls — whether from a halving or a price decline — and enough miners shut down unprofitable hardware, hashrate drops and difficulty adjusts downward to compensate, restoring profitability for the remaining, more efficient miners. This self-correcting mechanism is why Bitcoin mining has survived every halving to date without the network grinding to a halt: less competitive hardware retires, difficulty falls, and mining remains viable for whoever is left running efficient equipment.

This dynamic already plays out on every halving cycle and will continue to play out as the subsidy shrinks toward zero. It rewards operators running efficient, well-maintained fleets like the current-generation Antminer S21 series or Whatsminer lineups, and it progressively punishes anyone still running older, higher J/TH hardware that can no longer clear the electricity cost bar at reduced subsidy levels.

It's impossible to answer with certainty, and anyone claiming otherwise is speculating past the available evidence. What can be said is that mining profitability has never depended solely on the subsidy in absolute terms — it depends on revenue relative to the cost of hashrate, which is a moving target shaped by Bitcoin's price, network fee demand, hardware efficiency, and electricity costs, all of which will look very different over a 100+ year horizon than they do today.

What is far more certain is that mining hardware itself has a much shorter useful life than the halving schedule. No ASIC bought today will still be economically competitive in 2140; hardware generations turn over every few years as chip efficiency improves. The relevant planning horizon for a hardware purchase is years, not decades, and decisions should be based on near-term halving cycles and efficiency trends rather than the theoretical endpoint of Bitcoin's issuance schedule.

Given that the subsidy is guaranteed to keep shrinking every four years, buyers should prioritize efficiency and total cost of ownership over headline hashrate. A unit with a lower J/TH figure has more margin to survive the next halving than a less efficient one at the same price, because its operating cost per unit of revenue is structurally lower. This is one of the reasons newer-generation SHA-256 miners like the Canaan Avalon Q (90Th/s, 1674W) have found demand even at relatively modest hashrate, since their efficiency profile holds up better as subsidy revenue declines.

It's also worth factoring in resale value and realistic depreciation. Machines purchased near the top of a hashprice cycle often lose value quickly once difficulty rises or a halving hits, so buyers should model returns conservatively rather than extrapolating current network conditions four or eight years forward.

Because fee revenue is the intended long-run substitute for the block subsidy, the growth of layer-2 protocols is a double-edged consideration for miners. Systems that batch many transactions off-chain and settle periodically reduce the number of individual on-chain transactions, which could suppress per-block fee revenue even as overall Bitcoin usage grows. On the other hand, some proposed uses of Bitcoin's block space — including inscriptions and other data-carrying transaction types that gained popularity in recent years — have shown that demand for block space isn't limited to simple payments, and can meaningfully add to fee revenue during periods of high activity.

None of this is settled, and it's genuinely one of the more interesting long-term open questions in Bitcoin's design. For a mining operator, the practical takeaway isn't to try to predict fee markets a century out — it's to build a hardware and operating cost base that's resilient across a range of outcomes, which in practice means favoring efficient hardware, sound cooling infrastructure, and realistic payback period assumptions over speculative long-run fee projections.