TL;DR Bitcoin just mined its 20 millionth coin. That's 95.24% of all BTC that will ever exist. The remaining 1 million will take 114 years to mine. Meanwhile, the miners who secure the network are pivoting to AI data centers to survive shrinking block rewards.
The Milestone
On March 9, 2026, at block height 939,999, the Foundry USA mining pool collected a 3.125 BTC block reward that pushed Bitcoin's circulating supply past 20 million. It took 17 years, 2 months, and one week since Satoshi mined the Genesis block in January 2009.
That means 95.24% of Bitcoin's hard-capped 21 million supply now exists. The other 4.76% will trickle out over the next 114 years, with each halving cutting the new issuance rate in half every four years. The next halving hits in April 2028, dropping the block reward from 3.125 BTC to 1.5625 BTC.
For context: it took 17 years to mine 20 million coins. It'll take until roughly 2140 to mine the last million. If you want to understand what the world looks like priced in an asset this scarce, that's worth thinking about.
Why Scarcity Isn't Just a Meme
The stock-to-flow crowd has been talking about Bitcoin scarcity for years. Most of it is noise. But the 20 million milestone makes the math concrete in a way that models don't.
Consider the effective supply. An estimated 3.7 million BTC are permanently lost (forgotten wallets, lost keys, Satoshi's untouched stash). That drops the real circulating supply to around 16.3 million. With whales accumulating 270,000 BTC in the last 30 days and Strategy buying another $1.3 billion in March, the liquid supply is shrinking faster than new coins are being mined.
New daily issuance is currently about 450 BTC (144 blocks x 3.125 BTC). That's roughly $32 million worth at current prices. Spot Bitcoin ETFs alone absorbed more than that on multiple days in March. The math doesn't balance unless price adjusts.
The Miner Pivot
Here's what makes this milestone different from the 19 millionth coin (mined in April 2022): miners can't survive on block rewards alone anymore. The April 2024 halving cut rewards from 6.25 to 3.125 BTC. Transaction fees are inconsistent. Hash rate just crossed 1,000 EH/s, pushing up electricity costs per coin.
The response has been a massive pivot to AI. Marathon Digital, Hut 8, and CleanSpark have all converted mining capacity to AI data center hosting. It makes sense. The GPU infrastructure overlaps, the energy contracts are already in place, and AI compute demand is insatiable. But it means the companies securing Bitcoin's network are increasingly dependent on revenue from a completely different industry.
That's not necessarily bad. Diversified revenue makes miners more resilient, which makes the network more stable. But it does mean the "pure Bitcoin miner" is becoming an endangered species. For a deeper look at mining as an investment, the economics have fundamentally shifted since the halving.
What Comes Next
The 2028 halving will cut new supply to 225 BTC per day. The 2032 halving drops it to 112.5. At some point, likely within the next two halvings, transaction fees will need to exceed block rewards for mining to remain profitable without AI subsidies. That's Bitcoin's real long-term test.
For now, the 20 million milestone is a reminder of something easy to forget: this asset has a hard cap. No central bank can print more. No emergency issuance, no quantitative easing, no "transitory" supply expansion. The 21 million limit isn't a policy choice. It's math. And 95% of it is already spoken for. If you're thinking about how to custody that scarcity properly, now's the time.
Sean Ristau | @SeanRistau | 21Rates / The Daily Stack
