New Study Reveals Blockchain Throughput is Overestimated by 75%

A new report from Taraxa claims that many leading blockchain projects have dramatically overestimated their throughput. The study shows that major blockchain networks like Sonic, Solana, and Aptos have a significant gap between theoretical TPS (transaction per second) and the actual max TPS on the mainnet.

The findings suggest a massive overestimation of network efficiency and speed for these networks.

Most Blockchains Overestimate Efficiency

Taraxa, a Layer-1 blockchain, conducted an extensive analysis of several leading blockchains. It’s evident that most networks publicize new advances in their blockchain’s throughput, but many of these tests are conducted in ideal conditions. This study wished to observe how the most ‘bullish’ claims compare to regular operating conditions.

“Investors, developers, and users deserve transparency. The blockchain industry has long been obsessed with theoretical performance figures, but numbers generated in a lab mean little if they can’t be replicated in real-world conditions,” Taraxa co-founder Steven Pu said in an exclusive press release shared with BeInCrypto.

This investigation sought to assess these real-world conditions through a metric called “TPS per dollar.” Taraxa compared a blockchain’s transactions per second to the actual cost of running a validator node and used that to determine actual throughput.

This would be a more accurate way to determine how well these firms can live up to expectations.

Case in point, the study looked at the highest-ever recorded throughput on several blockchain projects, with a few important caveats. Permissioned and sharded networks were excluded, and some specific transactions (like voting transactions) were discarded to avoid number inflation.

Then, these figures were compared to developer-provided TPS claims:

The results of this test revealed extremely high exaggeration. Sonic (formerly Fantom) reported blockchain throughput over 100x its actual capabilities, but the industry average was 20x. The L1 blockchain space is full of fierce competition, providing a clear incentive for this systematic inflation.

“Our research also shows that many networks require expensive hardware just to achieve modest transaction rates, which is neither technically impressive nor decentralized. By focusing on verifiable data from live networks, we can shift the conversation toward meaningful performance metrics,” Pu added.

Comparing TPS to dollar costs also provided interesting data. Solana had the highest costs by far, but it used these resources efficiently to maintain a high blockchain throughput. Taraxa also claimed that it had the best ratio in the entire industry by wide margins, which may impact its reasons for conducting the study and using this metric.

Regardless of the firm’s desire to market its own capabilities, blockchain throughput estimations seem heavily inflated across the whole industry. Taraxa has been analyzing several crucial Web3 sectors, such as the AI industry, and its results seem valuable.

Hopefully, some hard data here will encourage more realistic reporting from these projects.