Bitcoin Cash stress test pushed 24-hour transactions above 2 million

Bitcoin Cash’s community-led mainnet stress test pushed the network above 2 million confirmed transactions in a rolling 24-hour window on September 1, 2018. The exercise was designed to flood the chain with minimum-fee transactions and observe whether miners, nodes and surrounding services could carry a workload far beyond ordinary use.

A contemporaneous Bitcoin.com report, published late on September 1, cited Fork.lol at 2,060,041 transactions over the preceding 24 hours, equal to 85,835 an hour or 23.8 a second. A September 2 Cointelegraph report, using BitInfoCharts, put the September 1 total above 2.1 million. Those figures use different snapshots and day boundaries, so they should be read as consistent evidence of the same order of magnitude—not as one perfectly interchangeable count.

A live test of the larger-block strategy

The event mattered because it tested Bitcoin Cash’s central scaling claim on the production network. Bitcoin ABC’s May 15, 2018 protocol upgrade had raised the maximum block size to 32 megabytes. Supporters argued that greater base-layer capacity could keep transactions cheap during demand spikes; critics questioned the storage, bandwidth and operational burden that larger blocks could impose on independently run infrastructure.

September 1 supplied an observable, if deliberately manufactured, data point. The public chain record preserves unusually transaction-heavy blocks from the test. Blockchair dates block 545,969 to 12:58:41 UTC on September 1, while contemporaneous reporting identified multiple blocks larger than the former 8-megabyte ceiling. This demonstrated that miners could include substantially larger batches under the upgraded rules.

It did not demonstrate that every node, wallet, explorer or merchant service performed equally well. Cointelegraph recorded a contemporaneous claim from engineer Jameson Lopp that some transaction and mempool visualizers broke under the load. That distinction is important: consensus continuing to advance is not the same as every dependent service remaining available.

Fees stayed low, but demand was synthetic

BitInfoCharts data cited on September 2 placed Bitcoin Cash’s average transaction fee at about $0.0017 during the test, slightly below the preceding roughly $0.002 reading. The unit was U.S. dollars per transaction, and the comparison came from an aggregator’s daily series rather than a universal fee quote. Transaction size, exchange rates and the exact observation boundary can change the dollar result.

The low fee reading supported the narrow claim that the generated load did not produce a sharp fee auction during the measured interval. It did not establish organic payment demand, economic value transferred, decentralization or long-run capacity. Test participants intentionally created large numbers of microtransactions; one payment between businesses and one test transfer each count as one transaction.

The market moved, but causation was unproven

Kraken’s September 1 daily report recorded BCH at $622.12, up 14.2% for the report’s interval, with $4.74 million traded in BCH markets on that exchange. The same venue reported $101 million across all its markets. These are Kraken-specific price, return and volume observations, not a consolidated global close.

The timing made the stress test part of the day’s market narrative, but the record does not isolate it as the cause of the price move. Cryptocurrency traded continuously across venues, and weekend liquidity, positioning and broader risk appetite could also affect returns.

The defensible September 1 conclusion was therefore narrower than either camp’s rhetoric: Bitcoin Cash processed an exceptional burst of deliberately generated on-chain activity without a documented fee spike, while exposing limits in some monitoring infrastructure. The test produced useful evidence about throughput under the 2018 rules. It did not settle the larger argument over whether bigger blocks could support sustained, decentralized economic use.