The Diem Inheritance
Aptos Labs was founded in late 2021 by Mo Shaikh and Avery Ching, former engineers on Meta's Diem (originally Libra) project. Mysten Labs, the team behind Sui, was founded the same year by Evan Cheng, Sam Blackshear, and other Diem alumni. Both teams forked the Move language and a chunk of the Diem codebase, then raised aggressively. Aptos pulled in roughly $350 million across seed and Series A rounds led by a16z, FTX Ventures, and Multicoin. Mysten raised more than $300 million, with a Series B led by FTX Ventures at a $2 billion valuation in September 2022, weeks before FTX collapsed.
That collapse erased FTX's stake in both projects, but the engineering teams and the Move-based architecture survived. Aptos mainnet launched in October 2022, Sui in May 2023. Both shipped novel consensus mechanisms, parallel execution engines, and aggressive throughput claims, alongside foundation token holdings and unlock cliffs that became a recurring source of community frustration.
Algorand's lineage is different. Founded by Silvio Micali in 2017 with mainnet live since June 2019, it skipped the venture-capital arms race. The Algorand Foundation and Algorand Inc. were funded by an early ALGO sale and have run lean compared to the Aptos and Sui war chests. Less marketing budget, more uptime.
Move vs Algorand Python: Two Theories of Safety
Move is a resource-oriented language designed at Meta to prevent common smart-contract bugs by construction. Resources cannot be copied or implicitly discarded, types enforce ownership semantics, and the bytecode verifier catches a meaningful class of errors before deployment. Aptos and Sui both use Move, though Sui's variant has diverged into "Sui Move" with object-centric primitives that enable parallel execution.
This is a real engineering advantage. Move's resource semantics map cleanly to assets, and the language is harder to misuse than Solidity. The catch is ecosystem maturity. Three years in, Move tooling, libraries, audit firms, and documentation are still catching up to what Solidity has had since 2015. Hiring Move developers is harder, and porting Solidity contracts requires a rewrite.
Algorand took a different bet. Smart contracts compile to TEAL, a stack-based bytecode designed for fast verification and predictable cost. Above TEAL sit PyTeal and Algorand Python (announced 2024), a near-full Python subset that compiles directly to AVM bytecode. AlgoKit ties the toolchain together: project scaffolding, a localnet, typed clients, and an opinionated developer experience.
Honest read: Move is a more elegant language than Python on a whiteboard. In practice, the population of developers who can ship a working contract on Algorand in an afternoon is several orders of magnitude larger than the population that can do the same on Aptos or Sui. Algorand's protocol-level features (atomic transfers, native ASAs, rekeying, multisig) also mean developers write fewer custom contracts in the first place, which is its own form of safety.
Consensus: Three Different Theories
Aptos: AptosBFT (Jolteon)
Aptos uses Jolteon, a HotStuff-derived BFT consensus that combines two-chain and three-chain commit rules to reduce communication rounds. Block production is leader-based with rotating proposers, finality is sub-second under good conditions, and the protocol handles up to one-third Byzantine validators. The tradeoff is the usual leader-based weakness: a slow or malicious leader degrades throughput until the next view change.
Sui: Mysticeti
Sui replaced its earlier Narwhal-and-Bullshark stack with Mysticeti in 2024. Mysticeti is a DAG-based consensus where validators produce blocks in parallel and a commit rule extracts a total order from the DAG. Mysten cites roughly 390 millisecond commit latency in production. For "owned object" transactions, Sui can skip full consensus via the Fastpath, which is genuinely fast. Shared-object transactions take the full consensus path.
Algorand: Pure Proof of Stake
Algorand's consensus uses a verifiable random function (VRF) to secretly select a small committee of stakeholders for each block. The committee proposes and votes, Byzantine Agreement completes in a single round, and the block is final. No fork choice rule, no probabilistic settlement, no slashing. Block times sit at roughly 2.8 seconds in 2026 following the throughput upgrades of 2024 and 2025. Because the committee is selected privately and rotated every block, an adversary cannot identify or bribe validators in advance.
The Finality Reality
All three chains have meaningfully faster finality than Bitcoin or Ethereum L1. Sui's Fastpath can settle owned-object transfers in under half a second. Aptos finalizes in roughly one second. Algorand finalizes in roughly 2.8 seconds with cryptographic certainty and no leader assumption. For payment and settlement use cases, the difference between 0.4 and 2.8 seconds is rarely the deciding factor. The difference between a block with a single proposer and a block with a randomly selected committee is.
Throughput: Theoretical, Peak, and Real
Aptos has cited theoretical throughput north of 160,000 TPS in lab conditions. Sui has cited 297,000 TPS in similar tests. Algorand's measured benchmark sits at roughly 10,000 TPS following the 2024 upgrades.
Real-world utilization tells a different story. Per Artemis and Chainspect dashboards through Q1 2026, Aptos averages a few hundred TPS in sustained production with peaks in the low thousands during NFT mints. Sui averages similar numbers, with the parallel object model giving it a real edge during DEX spikes and on-chain games. Algorand averages low hundreds of TPS in normal load and has handled mass-mint events and stablecoin settlement waves without backing up.
Fair takeaway: theoretical TPS comparisons are nearly meaningless. None of these chains is throughput-constrained at L1 in 2026. Algorand wins on consistency and instant finality, Sui on parallel-execution headroom for object-heavy workloads, Aptos sits in the middle. Pick based on workload, not benchmark posters.
Decentralization: Validators and Token Distribution
Validator counts as of early 2026, drawn from each project's official explorer and validator dashboards:
- Aptos: roughly 150 active validators
- Sui: roughly 106 active validators
- Algorand: roughly 125 block-proposing nodes (relays plus participation nodes), with broader stake-weighted committee selection across all online ALGO holders
Raw validator count is an imperfect metric. Algorand's Pure PoS gives every online ALGO holder an equal probability of selection per block, weighted by stake. Aptos and Sui require minimum stakes (currently 1 million APT and 30 million SUI) that effectively gate participation to professional operators. None of the three is in Cardano's league of roughly 3,000 stake pools or Ethereum's hundreds of thousands of validators.
Token distribution is where Aptos and Sui have taken sustained criticism. At launch, the Aptos Foundation and Aptos Labs together held roughly 80 percent of the initial 1 billion APT supply, with a four-year unlock schedule that has produced regular sell-pressure events. Sui's Foundation and early backers similarly held a large majority of the 10 billion SUI supply, with monthly unlocks that the community has tracked obsessively. Algorand's distribution is not perfect, but the original allocation reserved a smaller share for the Foundation and core team, and the rewards program ended on schedule in 2024.
Tokenomics and Unlocks
ALGO has a fixed maximum supply of 10 billion tokens, with circulating supply around 8.6 billion in early 2026 and the bulk of insider unlocks behind it. Inflation has tapered as designed and is now governed by community decisions on rewards. APT has a starting supply of 1 billion with no hard cap, an inflationary staking-rewards schedule, and a vesting tail that runs through 2026 and beyond. SUI has a fixed maximum of 10 billion with circulating supply growing through scheduled unlocks; CoinGecko data shows monthly SUI unlocks averaging tens of millions of tokens through 2025 and 2026, a steady headwind for spot price. Neither is a death spiral, but both involve ongoing dilution that ALGO holders no longer face.
Adoption: TVL, Users, and Shipped Projects
DeFiLlama numbers in Q1 2026 show Sui's TVL in the low billions, Aptos in the high hundreds of millions, and Algorand in the low hundreds of millions. Sui leads cleanly, driven by Cetus, Navi, Suilend, and on-chain games. Aptos has Thala, Aries Markets, and a steady DeFi base. Daily active addresses on Artemis show Sui ahead through most of 2025 and early 2026, with Aptos second and Algorand third.
The picture flips on enterprise and public-sector adoption. Algorand's roster includes FIFA's collectibles platform, the Italian Securities and Exchange Commission's blockchain monitoring program, T-Mobile's loyalty pilot, the Marshall Islands' SOV national digital currency framework, and a growing tokenized RWA footprint via Lofty, Quantoz, and Midas. Aptos has its Microsoft AI partnership and a handful of fintech pilots. Sui has Alibaba Cloud and several gaming-studio integrations. Algorand's enterprise adoption is more diversified and mostly in production rather than announcement-stage.
Honest read: Sui has won the consumer-crypto attention war among the three. Aptos has held its ground. Algorand's on-chain consumer numbers are smaller, but its production deployments outside trading and gaming run deeper.
Where the Newer Chains Have a Real Edge
Sui's parallel object model is genuinely innovative and a better fit for some workloads (high-frequency on-chain trading, MMOs, anything with many independent state objects) than any account-model L1. Aptos has a strong engineering team and a conservative design. Both shipped Move tooling that, while still maturing, is high quality. Both have raised enough capital to fund developer ecosystems on a scale Algorand cannot match dollar for dollar. If your dApp requires parallel execution across thousands of independent objects, Sui is a credible technical choice. The newcomers are not vapor.
Where Algorand Has a Real Edge
Eight years of mainnet uptime is not a marketing claim, it is a track record. Algorand has not had a chain halt or rollback. Aptos and Sui have each had multi-hour outages, Sui in November 2024 and Aptos in October 2022 just days after launch. Instant cryptographic finality without a leader assumption matters for payments, RWA settlement, and any use case where reorgs or stalls are unacceptable. Native protocol features mean fewer custom contracts and less surface area for bugs. The community-aligned distribution means fewer overhanging unlocks. The developer onboarding through Algorand Python and AlgoKit is faster than anything in the Move ecosystem in 2026.
| Dimension | Algorand | Aptos | Sui |
|---|---|---|---|
| Mainnet launch | June 2019 | October 2022 | May 2023 |
| Consensus | Pure PoS + BA* | AptosBFT (Jolteon) | Mysticeti DAG |
| Finality | ~2.8s, single block | ~1s | ~0.4s (Fastpath) |
| Smart contracts | Algorand Python, PyTeal, TEAL | Move | Sui Move (object-centric) |
| Active validators | ~125 + stake-weighted committees | ~150 | ~106 |
| Max supply | 10B ALGO (capped) | 1B APT start, no cap | 10B SUI (capped) |
| Foundation/insider unlocks | Mostly complete | Ongoing through 2026+ | Ongoing through 2030 |
| Documented mainnet halts | None | Yes (2022) | Yes (2024) |
Where Each Wins
Sui: highest TVL of the three, parallel object model, fastest sub-second finality on owned-object transfers, strongest on-chain consumer and gaming traction.
Aptos: conservative Move stack with strong engineering, sub-second finality, large enterprise partnerships, deepest VC-funded developer programs.
Algorand: longest production track record without halts, instant cryptographic finality without leader assumption, mature enterprise and public-sector deployments, capped supply with most insider unlocks behind it, faster developer onboarding via Python tooling.
The Question That Actually Matters
The 2022 framing was Solana-killers. By 2026, the better question is which Layer 1 a serious application will actually trust with settlement. Aptos and Sui have demonstrated that the Diem engineering team produced real technology, and Sui in particular has shipped consumer applications at a pace Algorand has not matched. But the chains are still working through unlock overhang, validator concentration, and the operational maturity that Algorand quietly accumulated while no one was watching.
For a high-frequency on-chain game, Sui is a defensible pick. For a fintech building a payments rail, an RWA issuer launching tokenized treasuries, or a public-sector pilot that cannot afford a chain halt, the calculus runs the other way. Different tools, different jobs.
Most of the chains that called themselves Solana-killers in 2022 are gone or fading. Aptos, Sui, and Algorand are all still here, still shipping, and still publishing real numbers. Which one wins your build depends on what you are actually building.