What is Aptos (APT)?

Aptos and Sui came out of the same wreckage. When Meta abandoned Diem in 2022, two groups of its engineers left and each built a Layer 1 around Move, the language Diem had produced. Aptos took the more conservative path: it kept Move's original account-based model rather than redesigning around objects.

That choice runs through everything. Aptos looks more like a conventional blockchain — accounts hold balances, transactions update them — and its parallelism, Block-STM, works by executing transactions optimistically and re-running the ones that turn out to conflict. Developers do not have to declare what their transaction touches, which makes it easier to write and less predictable to schedule. Sui gets cleaner parallelism at the cost of a harder mental model; Aptos gets a familiar mental model at the cost of some throughput determinism.

Move, and what it buys you

Move makes assets first-class types the language protects. A coin cannot be copied or silently dropped, because the type system will not compile it. Reentrancy in the Solidity sense is not expressible. This is a genuine safety improvement, and it is also why neither Aptos nor Sui has attracted anything close to the developer base of an EVM chain — the safety comes from making the developer learn a new discipline.

The 2026 performance work

Raptr, using what Aptos calls Prefix Consensus, reached 250,000 transactions per second at 750 millisecond latency. Shardines goes further, decoupling execution, storage and consensus to target over a million TPS.

Treat those numbers as what they are: benchmark figures produced under controlled conditions, which is how every chain's throughput claims are produced. They demonstrate the architecture scales, they say very little about what the network does under real mixed load, and no chain in production has ever needed a million TPS. The relevant question for Aptos is not the ceiling but whether demand ever approaches it.

• Gas for transactions, at fees low enough to be negligible for a user.
• Staking to validators, now at roughly 2.6% after the rewards cut.
• Collateral and liquidity across the Aptos DeFi ecosystem, which has grown past 330 projects.
• Governance, where APT holders vote on protocol changes — including the one that cut their own rewards.

Aptos has positioned itself toward payments and institutional settlement rather than toward retail speculation, with a stated ambition of being a high-throughput settlement layer for regulated flows. That is a slower, quieter market than the one Solana won, and it is a market where the throughput work actually matters.

APT launched in October 2022 with 1 billion tokens, no hard cap, and staking issuance on a declining schedule. A large share went to core contributors, investors and the Aptos Foundation on multi-year vesting — the standard venture structure, and the standard consequence of persistent unlock supply through the early years.

Proposal 183 and the shift to lower issuance

Governance cut staking rewards from a 5.19% annual rate to roughly 2.6%, roughly halving issuance, as part of a broader move toward a more deflationary structure. It is the same decision Polkadot made in March 2026 and the one Solana's SIMD-0228 failed to make: reduce what the protocol pays to reduce dilution.

It is a credible signal that governance can act against stakers' immediate interest, and it is worth being clear about what it does and does not do. Lower issuance reduces sell pressure. It does not create buyers. A chain with strong demand and high issuance is a better asset than a chain with weak demand and low issuance, and Aptos's central problem was never the issuance rate.

Transaction fees are partially burned, which at current volumes is a small offset.

Staking APT means delegating to a validator. The reward rate is now roughly 2.6%, down from 5.19% before Proposal 183, and any figure you find quoting the old rate predates the change.

The mechanics are moderately restrictive. There is a minimum to delegate, rewards and unlocks move at epoch boundaries, and withdrawing requires waiting for the lockup cycle to complete rather than exiting on demand. Check the current epoch schedule before committing rather than assuming it behaves like Solana's daily epochs.

Slashing exists in the protocol design but has not been an active practical risk for delegators to date; the realistic downside is an underperforming validator costing you rewards.

Competing for a market that already has a winner

Aptos is fast, cheap and well engineered, and so is Solana, which arrived first, has deeper liquidity, more applications and a genuine consumer culture. Aptos's differentiators — Move's safety, the throughput ceiling, institutional positioning — are real and have not yet been enough to move developers or users at scale.

Developer supply

Move is a smaller talent pool than Solidity by orders of magnitude, and Aptos's dialect is not interchangeable with Sui's, so the two chains split an already small pool rather than sharing it.

Unlock pressure

Large allocations to investors, contributors and the foundation continue to vest. Proposal 183 reduced issuance; it did not touch the vesting schedules, and those remain the larger supply flow.

Benchmarks are not adoption

A quarter of a million TPS in testing and a million-TPS target are engineering achievements. No production chain has ever been constrained by a ceiling that high. Capability that demand does not reach is cost without return, and the risk here is building a road nobody drives on.

Lower rewards, unchanged demand

Cutting staking to 2.6% reduces dilution and also reduces the incentive to stake, which in proof of stake is the same as reducing the incentive to secure the chain. It is a reasonable trade at Aptos's current staking ratio and it is a trade, not a free improvement.