Aptos has just made a very interesting change to its tokenomics strategy, and this could significantly alter how we think about the project.

The Foundation announced a hard cap of 2.1 billion APT, shifting from an inflationary model to a deflationary one. Basically, we moved from uncontrolled growth to controlled scarcity. With the circulating supply around 1.196 billion tokens, this limit creates a real barrier.

What stands out most is the multi-layered approach. The Foundation permanently locked 210 million APT (18% of the current supply), functioning as a continuous burn while still generating rewards. Subsidies now follow progress milestones and do not come out automatically. This is quite different from what we saw before.

On the rewards side, the annual APY dropped from 5.19% to 2.6%. It seems drastic at first glance, but it significantly reduces inflationary pressure. Validators and delegators still earn, but total issuance is controlled. Gas fees increased 10x, but stablecoin transfers remain ridiculously cheap, like $0.00014 per transaction. And here’s the important detail: all fees are permanently burned.

Community governance almost unanimously approved this. The proposal received 335.2 million APT voting in favor, only 1,500 against. Participation was 39%, above the quorum of 35%. This shows real confidence in the new model.

Decibel, Aptos’s on-chain perpetual DEX, is the kind of application that can make this strategy work. Every order, match, and cancellation happens on-chain, generating massive volume. Projections suggest over 32 million APT burned annually at scale.

The four-year unlock cycle for early investors ends in October 2026, which should significantly ease supply pressure. The adjusted circulating supply is already around 795-805 million.

In the end, what Aptos is doing is quite clear: linking the token’s value directly to the actual use of the network. Less inflation, more burns, reduced supply with growing adoption. It’s a bet that sustainability beats rampant issuance. At least governance seems aligned with this.