--- title: "Hyperscale: 500K TPS Public Test" path: "/contents/tech/research/hyperscale-500k-tps" version: "1.0.0" author: "Hydrate" createdAt: "2026-03-19T11:12:51.740Z" updatedAt: "2026-03-19T11:12:51.740Z" --- # Hyperscale: 500K TPS Public Test | **Date** | January 2026 (public test); February 20, 2026 (blog post) | | **Sustained TPS** | 500,000+ | | **Peak TPS** | 800,000+ | | **Shards** | 128 shards (64-shard comparison: ~250k TPS) | | **Nodes** | 384 bootstrap, 40 validators, 6 load-generation | | **Hardware** | m6i.xlarge (4 CPU cores, 16GB RAM) per node | | **Platform** | Amazon Web Services (Terraform + Ansible) | | **Transaction type** | Real cross-shard swap transactions | | **Open source** | Planned (pending agreement with third parties) | | **Related** | [hyperscale-rs](/contents/tech/research/hyperscale-rs), [Cerberus](/contents/tech/core-protocols/cerberus) | ## Overview In January 2026, the Radix Foundation completed the final public Hyperscale test, demonstrating that the [Cerberus consensus protocol](/contents/tech/core-protocols/cerberus) can sustain over 500,000 transactions per second and peak above 800,000 TPS — on commodity hardware using Amazon Web Services m6i.xlarge instances (4 CPU cores, 16 GB RAM per validator node). The test was [documented in a February 20, 2026 blog post](https://www.radixdlt.com/blog/interim-hyperscale-closing-the-chapter) titled "Interim Hyperscale: Closing the Chapter." The word "interim" signals that these results represent a milestone, not the ceiling — Xi'an's architecture is designed for continued linear scaling as more shards are added. ## Linear Scaling Confirmed The defining result of the Hyperscale test series is the confirmation of linear scaling. Private tests compared 64-shard and 128-shard configurations directly: - **64 shards:** ~250,000 TPS - **128 shards:** ~500,000 TPS Doubling the shard count doubled the throughput. This is the theoretical prediction of [Cerberus consensus](/contents/tech/core-protocols/cerberus), which uses a 2256 shard space and braids multi-shard consensus on-the-fly per transaction. The public test confirmed this property holds in practice with real cross-shard atomic swap transactions — not simplified transfer workloads. The bottleneck during testing was transaction generation, not consensus. At peak load, the six m6i.12xlarge spam nodes (48 CPU cores, 192 GB RAM each) reached their generation limits before the validator network reached its processing limit. This suggests the actual protocol capacity exceeds measured results. ## Test Infrastructure The public test involved 384 bootstrap nodes, 40 validator nodes, and 6 load-generation nodes, all provisioned on AWS using Terraform and Ansible automation scripts. Over 590 community nodes also participated in the public phase. Community participants ran nodes on hardware ranging from datacenter servers to personal laptops, demonstrating that the network operates on diverse hardware — not just specialized infrastructure. Cross-shard atomic integrity was maintained throughout testing. Every swap transaction that crossed shard boundaries either fully succeeded or fully reverted — no partial state was ever committed. This preserves the [composability guarantees](/contents/tech/core-concepts/radix-for-ai-agents) that make Radix suitable for complex DeFi and autonomous agent workflows. ## Open Sourcing Plans The Radix Foundation stated that code, tooling, and test logs are prepared for publication to enable independent reproduction of the results. Open sourcing requires agreement with parties outside the Foundation and did not have a confirmed timeline at the time of the blog post. The community implementation [hyperscale-rs](/contents/tech/research/hyperscale-rs) — a Rust implementation of Cerberus consensus developed independently — predates this announcement and provides a reference implementation for researchers. ## Context: Xi'an and Mainnet Path The Hyperscale tests validate the consensus layer that powers [Xi'an](/contents/tech/releases/xian), Radix's sharded mainnet upgrade. Xi'an is currently in development. The test results establish that the protocol can support the transaction volumes expected from a global DeFi and autonomous agent ecosystem before Xi'an launches. The Cerberus whitepaper, [published in 2020](https://arxiv.org/pdf/2008.04450) and peer-reviewed in the Journal of Systems Research in June 2023, predicted these scaling properties theoretically. The 2026 public test provides empirical validation at production scale. ## External Links - [Interim Hyperscale: Closing the Chapter — Radix Blog (Feb 20, 2026)](https://www.radixdlt.com/blog/interim-hyperscale-closing-the-chapter) - [Hyperscale Update: 500k+ Public Test Done — Radix Blog](https://www.radixdlt.com/blog/hyperscale-update-500k-public-test-done) - [Cerberus Whitepaper — arXiv 2020](https://arxiv.org/pdf/2008.04450) - [hyperscale-rs — Community Rust Implementation](/contents/tech/research/hyperscale-rs) - [Cerberus Consensus Protocol](/contents/tech/core-protocols/cerberus)