Aptos Labs wants the blockchain to disappear

The peculiar thing about a cryptocurrency wallet is how quickly it can become a test. Before you have done anything useful, you may be asked to record a secret phrase, install an extension, acquire a token, and understand a fee. Imagine a supermarket asking you to study the checkout machinery before buying an apple. Aptos Labs has built a business around taking that machinery seriously, including the parts a customer should never need to see.

Avery Ching and Mo Shaikh came from Meta’s blockchain effort. Aptos Labs, legally Matonee Inc., dates to 2021; the Aptos mainnet arrived in October 2022. Ching, formerly the CTO, became CEO in December 2024. The inheritance was experience building distributed systems. The assignment was finding people with a reason to use them.

The queue is only half the problem

Aptos is a public Layer-1 blockchain: applications share a network that records and verifies transactions. Aptos Labs contributes engineering and builds products around it. The company is distinct from both the network and Aptos Foundation, the organization supporting ecosystem development. That distinction matters when someone mistakes network activity for a company’s sales.

Its technical signature is Block-STM, a parallel execution engine. Instead of making every transaction wait while the previous one completes, it attempts work concurrently, checks for conflicting dependencies, and repeats work when necessary. Developers do not have to declare those dependencies in advance. Consensus, the process of agreeing on transaction order, remains a separate job.

Move, the programming language inherited from the Diem research lineage, brings asset-oriented safeguards and formal-verification tooling. These are meaningful choices for financial software. They also place Aptos in competition with Ethereum, Solana, and the fellow Move-based network Sui. A team choosing Aptos is choosing an engineering environment and ecosystem, along with its performance characteristics.

A wallet should not feel like homework

The product lineup follows the customer’s difficulties surprisingly closely. Petra is a self-custodial wallet for holding and using assets. Transaction simulation helps explain an action before signing; gas controls let users adjust fees. Aptos Names replaces an intimidating address with a readable identity ending in .apt. An explorer lets someone inspect what actually happened.

Then comes Aptos Connect, introduced in July 2024. It lets an application offer account creation through a familiar Google login, without asking the newcomer to begin with a recovery phrase or wallet installation. There are still security and identity dependencies underneath. The useful change is where the complexity sits: more of it belongs to the system.

Geomi approaches the same problem from the developer’s side. Previously Aptos Build, it expanded beyond node APIs into indexing, historical data, live synchronization, and sponsored transactions. Its 2025 rebrand reflected a broader product. A developer building a rewards app needs usable records and an onboarding flow, as well as somewhere to execute code.

The bill moves, too

There is a concrete commercial model here. Geomi meters infrastructure consumption. Its published pricing offers $10 in monthly usage credits, increasing to $100 when an organization adds a payment method. Above the allowance, charges depend on resources consumed, including compute, storage, and data streaming. These figures describe the reviewed pricing, rather than a permanent promise.

Gas sponsorship is especially revealing. An app can pay transaction fees for its users, sparing them a preliminary token purchase. The expense lands with the sponsor. An easier customer experience can therefore require a more deliberate operating budget. “Gasless” describes the interaction; someone still receives the bill.

Aptos Labs announced $200 million in March 2022 and $150 million that July. The latter round was led by FTX Ventures and Jump Crypto. That bought room to build, but the useful business question remains whether developers keep using and paying for the products after experimentation.

Five hours the benchmark could not explain

On October 18, 2023, the Aptos network delayed transactions for roughly five hours. The incident report identified non-deterministic gas accounting: validators could agree that a transaction had insufficient gas, yet disagree about the gas consumed. A performance change and a later reporting feature exposed a case testing had missed.

“No transactions were lost”Aptos mainnet incident report · October 2023

Transaction load was not the cause. The response included a fix and further testing measures. For other builders, the practical lesson is specific: test whether machines reach identical answers on awkward, incomplete operations. Capacity testing alone will not answer that question.

A wallet goes to the world’s fair

The Expo 2025 wallet offers a different kind of evidence. After its January migration to Aptos, the company reported more than 133,000 new Aptos accounts and 558,000 transactions by April 21. Those are dated application figures, not millions of paying customers. They do show the network being used inside a recognizable consumer experience.

The institutional route requires different accommodations. Aptos Ascend, announced with Microsoft, Brevan Howard, and SK Telecom, with BCG consulting, addressed permissioning and digital-asset controls. A financial institution needs authority over who can do what, alongside settlement speed. Labs’ expertise spans cryptography, distributed infrastructure, and the application tooling that connects those capabilities to customers.

Make the machinery earn its place

In March 2026, Shelby, developed with Jump Crypto, opened testnet early access for verifiable object storage. In October, Aptos announced an encrypted mempool: users can hide pending transaction payloads until finalization. Addresses and fees remain visible. These projects extend the engineering agenda into data access and trading privacy.

The approach has conditions. Shared verification must be worth the added complexity. Move requires specialist development, and highly contended transactions limit parallel gains. A conventional database may suit a simpler application. The part readers can copy is less exotic: remove the unnecessary step, account for who pays, and test the failure your headline metric misses. Give the customer a useful next click.