The Hashgraph consensus mechanism achieves Byzantine Fault Tolerance (BFT) using voting on a braided tree of hashed events (the "hashgraph") that is faster and more lightweight than Nakamoto consensus on blockchain using Proof of Work (PoW). However, it is being positioned for different use cases than cryptocurrency. Whether its consensus mechanism is better PoW really depends on the use case. Hashgraph actually layers nicely on top of blockchain, e.g., by using Bitcoin for Proof of Stake or Proof of Burn. That said, it will certainly be more suitable than blockchain for many commercial applications that need distributed shared memory and high transactions per second on consumer grade nodes.

In addition to the consensus algorithm, when comparing the two it also makes sense to consider the full range of features that would be supported by a bonafide cryptocurrency platform, which for Bitcoin would mean duplicating the features in its proposed payment platform, Lightning. Is it Hashgraph's express goal to replace Bitcoin+Lightning? Clearly not at this time. Can it? That is not yet clear, and whether it can will be determined as much by additional factors including licensing and developer community support as by the consensus algorithm. Lightning will support blockchain contracts, bidirectional payment transactions and payment routing. Hashgraph could conceivably be used to implement the off-chain transaction ledgers, though that could require dozens and possibly up to thousands of such Hashgraph ledgers to satisfy the millions to billions of transactions per second that Lightning is targeting.

I hope that helps, best of luck to you.