AEREDIUM Holdings Inc. (“The Trust Layer”) today announced benchmark results demonstrating ingestion throughput of 1,436,820 transactions per second (TPS) on its live TEE-BFT validator network under test conditions — placing it among the fastest blockchain networks ever measured.

The benchmark processed 2,000,000 transactions in 1.39 seconds with a 0% failure rate across a three-node validator network deployed in three separate geographic regions. All transactions were processed inside hardware-attested Trusted Execution Environments, meaning every result is cryptographically verifiable — not just fast, but provably correct.

Benchmark Results

How AEREDIUM Compares

Architecture

AEREDIUM’s throughput is achieved through a purpose-built architecture combining three innovations.

First, all validator nodes execute inside hardware-isolated Trusted Execution Environments where every block, every signature, and every state transition is bound to a cryptographic attestation of the code that produced it. This is not a software-level guarantee — it is enforced by the hardware itself. The operator cannot modify the running code, access the validator’s memory, or forge signatures. An external auditor can verify any block’s authenticity without trusting anyone in the chain of custody.

Second, the transport layer uses a micro-batching protocol that amortizes network overhead across thousands of transactions per message. Rather than transmitting each transaction individually — which caps throughput at the round-trip speed of the transport layer — AEREDIUM accumulates transactions into batches of up to 1,000 and delivers them as a single serialized message. The transport pipe sustains over 13,500 batch messages per second, and each message carries 1,000 transactions, yielding the 1.4 million TPS ingestion throughput measured in the benchmark.

Third, inside the enclave, a parallel processing engine fans each batch across a pool of 64 concurrent workers. Transaction validation — signature recovery, nonce verification, gas estimation — runs in parallel, with order preservation enforced at the batch boundary. This design ensures that the enclave’s multi-core allocation is fully utilized, rather than bottlenecked on sequential processing.

Trust Model

The result is a blockchain that is not merely faster than its predecessors but fundamentally different in its trust model. Traditional high-throughput chains achieve speed by relaxing decentralization or deferring finality. AEREDIUM achieves speed while strengthening trust — every block carries hardware-attested proof that it was produced by verified, unmodified code running in an isolated execution environment. The integrity guarantee is not probabilistic; it is cryptographic and immediate.

Scaling

AEREDIUM’s Group Model architecture enables linear horizontal scaling without the coordination overhead that limits monolithic blockchain designs. Each group of three validators operates as an independent consensus unit with its own block production, state management, and attestation chain. Groups can be deployed in any data centre, any region, and any jurisdiction. At 10 groups, the projected throughput is 14.3 million TPS. At 100 groups, over 143 million TPS. There is no theoretical ceiling — throughput scales with the number of groups deployed.

Institutional Infrastructure

This positions AEREDIUM as institutional-grade infrastructure for the stablecoin economy and the broader real-world asset ecosystem. Stablecoin settlement demands throughput that matches traditional payment rails — not the 15–30 TPS of first-generation blockchains, but the millions of transactions per second required for global commerce. AEREDIUM delivers that throughput today, with the hardware-attested integrity that regulated institutions require.

Methodology

The benchmark was conducted on production hardware using real TEE attestation across intercontinental distances. No simulation. No synthetic conditions. No local-only testing. The 1,436,820 TPS figure represents actual measured ingestion performance on a live, geographically distributed validator network under test conditions.