How the halving schedule produces a hard cap
Each block produces a fixed subsidy that halves every 210,000 blocks: 50 BTC at genesis, 25 after the first halving, down to the current 3.125 BTC, continuing to halve roughly every four years. Because each halving is a geometric reduction (never quite reaching zero, only ever getting closer), the total sum of all issuance ever produced converges toward a finite limit — 21 million coins — without a single explicit "stop" instruction anywhere in the code. It's a mathematical consequence of repeated halving rather than any separately enforced cutoff.
Why the curve flattens for so long
Because each halving cuts new issuance in half rather than to zero, the final fractional satoshis take an extremely long time to be mined — current estimates put the last new bitcoin around the year 2140. Well before then, in practical terms, new issuance becomes negligible: the overwhelming majority of the 21 million cap will already be in circulation decades earlier.
Why this is different from every other page on this site
Every other chart on this site — price, on-chain ratios, statistical fits — involves some form of estimation, measurement of past behavior, or extrapolation. The supply schedule's future line is the one exception: it's not a forecast about what miners, holders, or the market will do. It's a deterministic mathematical consequence of a fixed, already-written protocol rule that would require a consensus-breaking change to alter.