Avery Ching on Aptos, The Machine Economy, and the Quantum Clock

Avery Ching is co-founder and CEO of Aptos Labs, and before that he spent a decade at Meta, finishing as the tech lead for the Libra and Diem projects, Facebook's effort to ship a payments network and stablecoins with Visa, Mastercard, Uber, Lyft, and others.

From Diem to Move

When asked why so many Diem alumni went off to start their own chains, Avery framed the departure plainly. "It was mostly a necessity issue," he said. The Aptos Labs founders had assembled talent and built a stack inside Meta, then hit walls with regulators and the industry. Moving the work outside Meta became, in his words, the next logical conclusion. He notes how much the ground has shifted since, with a market that cares less about crypto tokens as speculation and more about the technology as useful for real applications. This shift in sentiment is what he calls an inflection point for the industry.

That history shaped what Aptos became. Looking at Bitcoin and Ethereum around 2018, Avery and his team concluded neither could carry mass-scale payments or trading. They set out to build a technology stack that would be scalable, applying the same principles used for building at Facebook. The goal was to support millions of transactions per second with sub-second latency. They built Move, a new smart contract language, and paired it with a chain designed around high transaction volume and security strong enough to hold billions, and eventually trillions, of dollars. Eight years on, he points to 25 millisecond block times and low fees as evidence the bet held up, and argues that any stablecoin arriving on Aptos gets the cheapest, fastest version of that product.

Pay the Cost When You Need To

A few months ago, Coinbase named Aptos as one of the chains best prepared for quantum computing. Avery credits a cryptography team he describes as always ahead of the curve, the same group behind the confidential assets and encrypted mempool work. His case for why post-quantum readiness matters is about timing and cost. Quantum-resistant signatures are currently more expensive than conventional ones, and adding a new signature scheme to a live blockchain requires development and performance analysis long before a practical attack exists.

Quantum-resistant cryptography is extremely expensive compared to normal cryptography, so pay the cost when you need to, but we're ready and prepared for that.

The practical worry is how fast the curve flips once quantum capability arrives. A chain that is not ready risks users losing funds and accounts being shut down. Aptos researchers proposed SLH-DSA-SHA2-128s account signatures in late 2025, and the public code includes a feature-gated implementation. This supports a migration path for account signatures, though the entire consensus stack is not yet fully post-quantum secure. Aptos lets users opt into quantum protection whenever they want, so nobody pays the performance penalty today for a threat that is still years out. Avery also cautions against treating correctness as the only concern: cryptography that degrades system performance to the point of being unusable is its own availability threat.

The Silent D in DEX

Avery is blunt about decentralized exchanges that are not very decentralized. Many keep centralized matching components or run on a single-machine L2, and to him the label does not hold. "The 'D' is a little silent to me," he said of those systems. Decibel is Aptos Labs' answer: a spot and perpetuals exchange whose central-limit order book, matching, clearing, margin, and settlement logic run in Move on Aptos.

The claim he makes is that this can match centralized venues, which operate in the 10 to 100 millisecond range, using the same 25 millisecond blocks with high availability. Decibel recently added spot trading to its perpetuals markets, and in the five or six months since launch it has done roughly $6 billion in perpetuals volume. Avery points to a big week for crypto showing up directly in Decibel's workload as evidence that a general-purpose chain can carry high frequency trading. He calls Decibel the most transparent system he knows of when it comes to reporting volumes.

Receipts for Agent Actions

Aptos Labs built Shelby with Jump Crypto as a global object store aimed at AI workloads, from generative video to training data. Getting long-term leases on H200 and B200 hardware is difficult, and that compute is scattered worldwide. Compute is a stateless resource, but storage is stateful. It has to physically move bytes from place to place efficiently, with the APIs developers expect.

Robotics is an area where Avery thinks could benefit from blockchain technology. He owns one robot vacuum he likes, and he expects more capable machines to follow. An agent in the home is constantly looking around and hearing conversations, which raises questions about what happens to that data and what actions the agent is taking. His answer points back to blockchains as immutable ledgers for logging agent actions and tracking data provenance. He referenced attacks on Hugging Face as a reason agents will need to record what they do and why.

Something that you can never change, something that's available worldwide, something global, that product is a blockchain.

On-chain receipts could let a frontier lab confirm that a user's data was kept out of a specific version of a model, checking that it was not used to train a successor. "An immutable ledger is what's necessary here," Avery said. The same logic extends to payments. Before agents negotiate anything exotic, he expects them to start by using existing websites and APIs, making a restaurant reservation or an Amazon purchase with a user's credentials. Later, agents may move toward brokering idle GPU time on a home RTX setup and settling it over a stablecoin rail.