What it means
Any shared ledger (a record of transactions kept by many computers at once) needs a way to settle disputes about which transaction came first. Without that, someone could spend the same money twice.
A consensus mechanism is the agreed rulebook for settling the order, and hashgraph is one such rulebook. Hashgraph was invented by Leemon Baird and is built on two ideas.
The first is called gossip about gossip, where each computer passes on not just new transactions but also the record of who told what to whom and when. The second is virtual voting, where every computer works out how the others would vote on the order of events from that shared history, so no actual vote messages need to be sent.
The practical effect is speed and efficiency. Because the voting is calculated rather than transmitted, the network can process many transactions per second with modest energy use compared with mining-based systems.
The design is also described as asynchronous Byzantine fault tolerant, which means it can reach agreement even when some participants fail or behave dishonestly, provided fewer than one third of them do. For a business reader, the interest lies in what it makes possible.
Fast, final settlement and predictable low fees suit use cases such as micropayments, tracking goods through a supply chain and recording tokenised assets. A finance team evaluating such a platform should still ask who runs the network, how much control any one party has and whether the legal rights attached to recorded assets are clear.
An important nuance is that hashgraph is a technology, not an investment or a guarantee of quality. Governance, meaning who decides how the network changes, varies by platform and can matter more commercially than the algorithm itself.
Different networks also make different trade-offs between openness and control.
In practice
Real-world examples.
Example
A logistics company records each handover of a refrigerated shipment on a hashgraph-based ledger. The recorded order of events is agreed by the network and cannot be quietly rearranged by any single party. When a claim arises for spoiled goods, the insurer and the shipper both rely on the same timeline.
Example
A media business wants to pay small royalties to thousands of creators every day. A network with low per-transaction fees makes payments of a few cents commercially sensible, where a card network would cost more than the payment itself.
Example
A bank piloting tokenised bonds tests a permissioned network where a fixed group of institutions run the nodes. The pilot team compares settlement time and cost against its current process and also asks the legal team who is liable if a node operator fails.
Formula
Calculation
Maximum faulty participants tolerated (f) = the largest whole number below n divided by 3, where n is the number of participants, which means n must be at least (3 x f) + 1.
Suppose a network is run by 100 independent nodes (the computers that validate transactions).
Step 1: 100 divided by 3 = 33.33.
Step 2: The largest whole number of faulty nodes that keeps n at least (3 x f) + 1 is 33, because (3 x 33) + 1 = 100.
Step 3: Check 34 faulty nodes: (3 x 34) + 1 = 103, which is more than 100, so 34 would break the guarantee.
Result: the network keeps working correctly with up to 33 faulty or dishonest nodes, and agreement needs more than two thirds of nodes, which is at least 67 of the 100, to be behaving properly.Case study
Seen in the real world.
Harbourline Freight is a fictional shipping broker that struggled with disputes over delivery times, which cost it about $400,000 a year in credits and staff time. It trialled a hashgraph-based ledger shared with five partner carriers, so every pickup and delivery was time-stamped in an agreed order.
Disputes fell by more than half in the illustrative trial because both sides were looking at the same record. The finance director noted, however, that the saving depended on carriers actually using the system, and that governance of the shared network took more negotiation than the technology did.
Watch out
Common mistakes.
- Treating hashgraph as the same thing as a blockchain, when it organises its records as a graph of events and not as a chain of blocks.
- Assuming a fast network is automatically safe, when governance, legal clarity and the quality of the applications built on it matter just as much.
- Confusing the consensus mechanism with a cryptocurrency, when it is only the method used to agree on order and any token is a separate matter.
Questions
People also ask.
What does gossip about gossip mean?
It means each computer shares both new information and the history of how it learned it, which lets others reconstruct who knew what and when.
What is virtual voting?
It is a way of working out how every participant would vote on the order of events by calculation from the shared history, so no voting messages are sent.
Is hashgraph decentralised?
That depends on the network using it; some platforms spread control across many independent operators, while others rely on a limited council or a permissioned group.
From the founder's library
Take it further with the book.
Build your financial confidence beyond this definition. Shihan's full-length guide, Accounting Fundamentals, takes the same plain-English approach and turns it into a complete, practical playbook for non-finance managers, business owners and students - with chapter-end quiz answers and presentation slides included.
25% off with code MMHQ25, applied at checkout. Priced in USD - checkout may show the equivalent in your local currency.
View the book and save 25%Related