While the Bitshares team suggests a theoretical limit of 100,000 transactions a second, they are using technologies and lessons from the LMAX exchange, which claims to be able to process over 6 million transactions a second. Most blockchains do not achieve a performance that is anywhere near the theoretical performance, with most blockchains achieving well under 1,000 transactions a second in practice. For instance, Bitcoin achieves approximately seven transactions a second, and Ethereum around 14. A decent server running MySQL can process 10,000–20,000 transactions a second of similar complexity. Thus, traditional approaches are much easier to scale to larger transaction volumes than blockchain systems.

The cost per performance of traditional database systems, including distributed database systems, is vastly cheaper on every level than blockchain systems. For any business, hiring experienced personnel for these systems is cheaper, easier, and less risky because they are so much better known and documented. Businesses considering blockchain must ask if a blockchain-based system provides a must-have feature in some way, because if not, other approaches are cheaper, easier, faster, or all of the preceding.

A business that is evaluating blockchain should pay close attention to their scaling requirements and cost models when using blockchain. In time, costs and performance may improve, but there are no guarantees that they will improve fast enough if current approaches aren't sufficient.