+34 641 10 27 16

Home/Cryptocurrencies/Handshake (HNS)

Handshake (HNS): guide to the coin, the naming network and its mining

Coin encyclopedia

Handshake (HNS): the DNS root on chain and mining after the halving

The second halving landed on July 27, 2026 at block 340,000 and the reward fell from 1,000 to 500 HNS. Most calculators have not caught up yet and still double the estimated income.

Blake2B plus SHA3Second halving July 27, 2026Vickrey name auctions500 HNS per block

Denys Klimchukupdated 7 August 202625 sections

About the coin

Handshake is a proof of work blockchain built for exactly one job: ownership of top level domains. The network does not try to replace the whole of DNS. It replaces only the root zone and the list of TLDs that ICANN administers today, and moves the distribution of names into open on chain auctions. Mainnet went live on February 3, 2020 at 17:00 UTC, and the full set of transaction and name operations opened on February 17, 2020.

The project was put together by a group of well known industry developers: Joseph Poon, co author of the Lightning Network, Andrew Lee, head of Purse and founder of Private Internet Access, Christopher Jeffrey, author of the bcoin implementation, plus Boyma Fahnbulleh and Steven McKie. In August 2018 the project raised 10.2 million dollars from a16z crypto, Sequoia Capital, Founders Fund, Draper Associates and Polychain Capital. The money went into development, after which the launch happened with no ICO and no presale of the coin to outside investors.

The defining organisational trait of this network: it has no owner. No foundation, no treasury, no company answering for development, marketing or listings. The hsd code is open and changes come from a scattered community through soft forks. Ideologically that is honest for a protocol, but the other side of the coin became decisive by 2026: when the infrastructure around the network started shutting down, there was nobody left to replace it.

This page is written as a reference for hardware owners, not as a promotional card. Below you will find the algorithm broken down from the source code, the mechanics of Vickrey auctions, the full timeline up to and including the latest halving, the state of the machine and pool markets, and a separate list of the figures that the major reference sites on this coin were still publishing in outdated form at the end of July 2026. The honest verdict on profitability is stated in plain words, with nothing softened.

Coin card

Parameter	Value
Name	Handshake, a network for decentralised ownership of top level names
Ticker	HNS
Network launch	February 3, 2020, 17:00 UTC. Full functionality from February 17, 2020
Founders	Joseph Poon, Andrew Lee, Christopher Jeffrey, Boyma Fahnbulleh, Steven McKie
Funding	10.2 million dollars in August 2018: a16z crypto, Sequoia, Founders Fund, Draper, Polychain
Owning organisation	None. No foundation, no treasury, no development company
Purpose	Replacing the DNS root zone and the TLD registry, auction based ownership of top level domains
Algorithm	Blake2B plus SHA3, three hash calls per attempt and a final XOR with the pool mask
Consensus	Proof of work, Nakamoto consensus
Block time	10 minutes target
Block reward	500 HNS since July 27, 2026, block 340,000
Halving	Every 170,000 blocks, roughly 3.23 years
Original supply cap	2,040,000,000 HNS, two thirds of it reserved for the airdrop to open source developers
Cap after the fork	Around 961,271,000 HNS. The soft fork of August 28, 2025 closed the unclaimed airdrop
Name mechanics	Vickrey second price auction, the winning bid is burned
Merge mining	None. The Blake2B plus SHA3 pair is used by this network only

Project history

2018In August the project raises 10.2 million dollars from a16z crypto, Sequoia Capital, Founders Fund, Draper Associates and Polychain Capital. There is no public coin sale: two thirds of the future supply are reserved for an airdrop to open source developers.
2020On February 3 at 17:00 UTC mainnet goes live. On February 17 the full set of operations opens: transactions, auction openings, bids and name registration. The initial block reward is 2,000 HNS and mining starts on graphics cards and on the hs-miner software implementation with CUDA and OpenCL backends.
2020The first machines built specifically for Blake2B plus SHA3 reach the market. Bitmain ships the Antminer HS3, Goldshell enters the niche with its HS line. Graphics cards are pushed off the algorithm within a few months.
2021In May HNS sets its all time price high, 0.8519 dollars. By 2026 the coin has lost more than 99.8 percent from that mark. It was the peak of interest in the idea of decentralised domains, and it has never been repeated.
2022Namecheap buys Namebase, the main venue for buying, selling and storing Handshake names. HNS domain registration appears in the interface of an ordinary large registrar. This is the deepest the project ever reaches into the mainstream domain industry.
2022In December the last batch of industrial machines for this algorithm ships, the Goldshell HS6 and HS6 SE line. After that date and through the end of July 2026 no manufacturer has released a new model for Blake2B plus SHA3.
2023On April 28 at 10:30:38 the first halving takes place at block 170,000. The block reward falls from 2,000 to 1,000 HNS and daily network issuance from 288,000 to 144,000 HNS. The date is confirmed by the direct block card in the explorer.
2024In February the community ships a soft fork that the project calls "The Happening". Reserved but never claimed names from the Alexa Top 10,000 list and from the ICANN TLD registry are unlocked and go to open auction. This noticeably revived the name market inside the network.
2024On December 30 at 00:00 UTC F2Pool shuts down its HNS and RVN pools. The announcement came on December 25, balances were paid out over the following week and access to reports stayed open until February 5, 2025. The exit of one of the largest multi coin pools was the first signal that the infrastructure around the coin was winding down.
2025On August 28 the second major soft fork activates. It closes the ability to claim the airdrop, which over more than five years had been claimed by only about 2.9 percent of recipients. The supply cap is cut by roughly 56 percent, from 2.04 billion down to something on the order of 961 million HNS. The network finally becomes a pure PoW chain: new coins now reach only miners.
2026On January 28 Namecheap sells Namebase. The buyers are not named publicly. The built in Namebase exchange closes, open orders are cancelled, funds are returned to user accounts and the platform goes into migration.
2026On June 8, UTC plus 8, ViaBTC stops mining HNS. The announcement came on June 1, report exports stayed available until June 7, and settlements, withdrawals and conversions closed on June 30. After that, public pools on the algorithm are practically gone.
2026On June 10 Namecheap stops working with Handshake domains altogether: new registrations, renewals, transfers and management are unavailable, with an upstream provider winding down given as the reason. Already registered domains keep working in their current configuration but are not renewed. In the same month Namebase announces its own wind down: no name renewals, BTC and HNS withdrawals open, staked TLDs frozen pending a decision on the registry.
2026On July 27 at 08:59:28, at block 340,000, the second halving takes place. The block subsidy falls from 1,000 to 500 HNS and daily network issuance from 144,000 to 72,000 HNS. The next halving is expected at block 510,000, roughly in 2029, taking the reward down to 250 HNS.

How the network works

The network splits into two independent layers. The first layer is an ordinary PoW block chain: miners iterate the nonce, find a header below the target value, and collect the reward and the fees. The second layer is the name registry, which lives in a separate data structure, an urkel Merkle tree. The root of that tree is written straight into the block header, alongside the transaction root.

The key difference from a typical blockchain is that the name tree is not updated every block but roughly once every six hours. This is done for light clients: a resolver needs to prove that a record exists for a name, and if the root changed every ten minutes, those proofs would have to be refreshed constantly. Once every six hours the proof stays valid long enough for an SPV node to run on weak hardware and answer DNS queries quickly.

A name cannot be bought at a list price, it can only be won at auction. The mechanism is a Vickrey auction, or second price auction: all bids are blind and the winner pays not their own bid but the second highest one. The difference goes back to the winner, and the amount paid goes to nobody at all: it is burned and leaves circulation forever. According to the explorer, by the end of July 2026 some 109.7 million HNS had been burned this way, a visible share of the supply.

After winning, the owner gains the right to run the zone under their TLD and to publish DNS records on chain, including DNSSEC keys and DANE records for certificates. The right is not permanent: a name has to be renewed periodically with a transaction, otherwise it returns to circulation. The legacy of the old internet was handled separately: names from the top of the Alexa list were reserved for the owners of the matching domains, who could claim them through a DNSSEC proof. The unclaimed part of that reserve was unlocked by the community with a soft fork in February 2024.

Building the headerThe node assembles a block and computes four roots: merkle for transactions, witness for witness data, tree for the urkel name tree and reserved for future extensions. All four go into the header.
Splitting the headerThe header is split into a 128 byte pre header and a sub header. The pre header carries only the essentials, including the name tree root. This is an optimisation for light resolvers, which do not need transactions but do need names.
Iterating the nonceThe miner changes the 32 bit nonce at the start of the pre header and on every attempt computes three hashes: Blake2B-512 of the pre header, SHA3-256 of the pre header with eight bytes of padding, and Blake2B-256 of the first two concatenated with 32 bytes of padding between them.
Applying the pool maskThe result is XORed with a 32 byte mask that the pool operator sets and does not disclose to individual miners. This is protection against block withholding: a miner cannot tell that they have found an actual block rather than an ordinary share.
Checking and broadcastingIf the result does not exceed the target value, the block counts as found and goes out to the network. Nodes verify it by recomputing the same chain of three hashes and restoring the mask from the maskHash field in the header.
Retargeting difficultyDifficulty is adjusted after every single block rather than once per interval. The protocol target is 144 blocks a day, that is a ten minute block. Step by step adjustment smooths out the swings typical of networks with a small and unstable hashrate.
Updating the name registryName operations, opening an auction, bidding, revealing, registering and updating records, land in blocks as ordinary transactions, but they are folded into the urkel tree and the tree root roughly once every six hours.

Mining algorithm

The easiest way to describe Handshake mining is straight from the hsd source, file lib/primitives/abstractblock.js, method shareHash. Every attempt runs exactly three hash function calls. First two branches are computed in parallel: the left one is Blake2B with a 512 bit output over the pre header, the right one is SHA3-256 over the pre header with eight bytes of padding appended. Then a third hash is taken, Blake2B with a 256 bit output, over the concatenation of the left branch, 32 bytes of padding and the right branch. That is why the algorithm is called a Blake2B plus SHA3 pair rather than a single function.

The padding here is neither random nor chosen by the miner. In the code it comes from a byte wise XOR of the previous block hash with the name tree root. In other words, the padding is deterministically tied to the chain state and changes from block to block. Separate from it there is another value that is often confused with the padding: the 32 byte pool mask. The final result is XORed with that mask before it is compared against the target, and what goes into the header is not the mask itself but its hash together with the previous block.

The point of the mask is spelled out in a comment in the code: it exists to counter block withholding. The pool operator keeps the mask secret and does not hand it to miners. A miner computes a hash and sees that the share cleared the pool threshold, but cannot tell an ordinary share from a real block, because without the mask the final value is unknown to them. That makes the scheme pointless in which a dishonest participant submits every share except the ones that turned out to be blocks.

Why ASICs dominate this algorithm. Both functions, Blake2B and SHA3, are cheap in silicon, need no memory and pipeline beautifully. There is no anti ASIC protection in the design and none was intended: judging by early discussions in the repository, the authors only wanted to devalue the SHA3 chips that already existed so that starting conditions would be equal, not to rule out ASICs as a class. That goal was met, and purpose built machines appeared in the network first year.

Graphics cards do formally work. The official hs-miner ships CUDA and OpenCL backends, and mining on a GPU is technically possible. It has made no practical sense since 2020: the efficiency gap against ASICs is orders of magnitude, and at the current coin price even purpose built hardware does not pay for its electricity. The only scenario left for a GPU here is a learning run on testnet.

Full description of the Blake2B plus SHA3 algorithm →

Economics and issuance

The original scheme looked like this: a cap of 2,040,000,000 HNS, of which 1,360,000,000, exactly two thirds, were reserved for an airdrop to open source developers. Eligibility came from a GitHub account with a minimum level of contribution, a Hackerbase account or a PGP key in the strong set of the web of trust. The remaining 680 million HNS or so were meant for miners through the block reward. The initial reward was 2,000 HNS with a halving every 170,000 blocks, which is about 3.23 years.

The airdrop effectively failed. Over more than five years it was claimed, by most estimates, by only about 2.9 percent of recipients: the overwhelming majority of developers either never heard about it or saw no reason to go through the procedure. On August 28, 2025 the community shipped a soft fork that closed the window for claiming those coins. The supply cap was cut by roughly 56 percent. Handle the number with care: press releases from that time quote a figure of about 889 million HNS, while market aggregators at the end of July 2026 show a cap on the order of 961.3 million. Keep that gap in mind in any calculation.

Halvings are tied to block height, and both dates so far are confirmed by direct block cards in the explorer. First halving: block 170,000, April 28, 2023, 10:30:38, reward from 2,000 to 1,000 HNS. Second halving: block 340,000, July 27, 2026, 08:59:28, with the block subsidy shown as 500.00 HNS. The third halving is expected at block 510,000, roughly in 2029, taking the reward to 250 HNS.

A separate mechanism, and a very important one for this network, is burning. The amount an auction winner pays for a name goes neither to miners nor to any fund: it is destroyed. According to the explorer, by the end of July 2026 roughly 791.6 million HNS had been issued in total, roughly 109.7 million had been burned and about 681.9 million remained in circulation. Put differently, about every seventh coin ever issued has been destroyed. This is a rare case where a deflationary mechanism genuinely works at a visible scale.

What the halving means arithmetically. Before July 27, 2026 the network issued 144 blocks of 1,000 HNS, that is 144,000 HNS a day. After the halving it is 72,000 HNS a day for the entire global fleet. At a price on the order of a sixth of a cent, the whole daily issuance of the network is worth a little over a hundred dollars. It helps to keep that figure in view in any conversation about profitability: that is the number everyone is dividing up.

Network parameters

Parameter	Value
Mining algorithm	Blake2B plus SHA3, three hash calls per attempt
Hash chain	Blake2B-512 of the pre header, SHA3-256 of the pre header with padding, then Blake2B-256 of the concatenation
Padding	Byte wise XOR of the previous block hash with the name tree root, deterministic
Final step	XOR with a 32 byte pool mask, the header stores its hash together with the previous block
Block time	10 minutes target, about 9.85 minutes in practice at the end of July 2026
Difficulty retarget	After every block, target 144 blocks a day
Block size	1,000,000 bytes of base and 4,000,000 weight units
SegWit	100 percent, the witness commitment is moved into the header
Roots in the header	merkle, witness, tree for names and reserved for extensions
Name tree	An urkel structure, the root is updated roughly once every six hours
Address format	bech32 with the hs1 prefix on mainnet
Smallest unit	dollarydoo, 1 HNS equals 1,000,000 of them
Special operations	Covenants on outputs, opcodes OP_TYPE, OP_BLAKE160, OP_BLAKE256, OP_SHA3, OP_KECCAK
Merge mining	Not supported, no other network uses the Blake2B plus SHA3 pair

Which ASICs work

Fewer than ten models have ever been built for Blake2B plus SHA3, and not a single new one since December 2022.

Model	Hashrate	Power draw	Efficiency
Antminer HS3	9 TH/s	2079 W	231 J/TH

Browse miners in the catalogue →

Best machines for this coin

The asic.es catalogue holds exactly one machine on this algorithm, so the ranking has a single entry and the rest of the fleet is covered in the comparison table below.

1Antminer HS3

Pros. 9 TH/s at 2,079 W, which works out to 231 J/TH. The most efficient machine ever built for Blake2B plus SHA3: the Goldshell HS6 behind it trails by about a third, and the smaller Goldshell models by a factor of two or more. It is a Bitmain platform, so the web interface, monitoring and pool setup are familiar to anyone who has run an Antminer.

Cons. The machine can compute Handshake and nothing else. Switching it to another algorithm or another coin is impossible, so once mining HNS stops making sense, the hardware turns into scrap metal. That is the fundamental difference from the Goldshell HS line, which switches over to Siacoin. There are no new units from the factory and no warranty on the secondary market either. No third party firmware exists for the platform.

Best suited for. Anyone with an industrial tariff of around two cents per kilowatt hour or their own generation, buying the machine for token money to take part in the network rather than to earn.

Payback. The break even point on electricity alone sits around 0.029 dollars per kilowatt hour by the end of July 2026 calculation, with the halving already included. At a Spanish household tariff of roughly 0.10 to 0.15 euro the machine runs at a loss of about 4 to 7 dollars a day. A payback period is not calculated under those conditions: there is none.

Cooling. Air. More than two kilowatts of heat and an industrial noise level, so a living space is ruled out, you need intake and organised removal of the hot stream

Hardware manufacturers

BitmainAntminer HS3, 9 TH/s at 2,079 W, 231 J/TH

The only Bitmain model for this algorithm and at the same time the most efficient machine the network has ever seen. It arrived on the wave of interest in 2020 and never got a successor: the company never released a second Handshake model. One important trait that sellers usually keep quiet about: the HS3 is single purpose and computes Handshake only, with no fallback use.

GoldshellThe HS line: HS Box, HS5, HS6, HS6 SE, from roughly 307 to 979 J/TH

The main supplier by number of models and the only one to cover both the industrial segment and the desktop box niche. The key advantage of the line is that the machines are dual algorithm and switch over to Siacoin. That, rather than Handshake mining, is what defines their residual value today. The last models in the line, the HS6 and HS6 SE, shipped in December 2022, after which Goldshell left the algorithm too.

Graphics cardsThe official hs-miner with CUDA and OpenCL backends

A branch that formally exists but is economically dead. The project software miner can run on GPUs, and in the first weeks after the network launched in February 2020 that was exactly how mining was done. Once purpose built machines appeared, the gap became unbridgeable. Today a GPU on Handshake is purely a developer tool for testnet.

The market as a wholeZero new models since December 2022

The main fact about Handshake hardware: by the end of July 2026 more than three and a half years had passed without a single new machine. Neither Bitmain nor Goldshell nor any of the smaller manufacturers announced a line refresh. For a buyer that means the entire available fleet is the secondary market, with hardware four years old and older, no warranty and awkward spare part supply.

Firmware and overclocking

Straight to the point, with no illusions: no firmware exists for Blake2B plus SHA3. There are no machines on this algorithm in the HashCore Toolkit support lists, and we neither build nor plan AsicBoost images for the Antminer HS3 or for the Goldshell HS line. Our tooling targets SHA-256 and Scrypt, where the installed base runs into hundreds of thousands of machines and development pays for itself. A single catalogue model on Handshake is not a market worth developing and supporting firmware for.

There is no third party market either, and that is a statement of fact rather than an opinion: in six years of the network not one mature public project for this algorithm has appeared that would compare with what exists on SHA-256. The reason is simple and has nothing to do with engineering. Building firmware means months of work and ongoing support, and there is nothing to recoup it from: the entire daily issuance of the network after the halving is worth about a hundred dollars. If someone offers to sell you firmware for the HS3, treat it as a reason to be suspicious.

What an owner actually has instead of firmware. Three levers, and all of them are external to the machine: the quality of the power supply and the stability of the line, the air temperature at the intake, and the choice of a pool with a lower fee. None of them will give you the effect that careful undervolting gives on SHA-256, because the platform is not built for fine tuning of frequencies and voltages. Run the official manufacturer firmware.

A word on risk. Any unverified firmware on a machine drawing more than two kilowatts is not an experiment, it is a fire load. An error in the voltage table kills a hash board in minutes and voids whatever is left of the warranty. In a market where no new machines have shipped since December 2022, a broken Antminer HS3 is next to impossible to replace.

How to start mining

Technically, starting to mine Handshake is simple. You need three things: a machine for Blake2B plus SHA3, a wallet address in hs1 format and a working pool. Setup is no different from any other ASIC: you enter the stratum address in the web interface, put your wallet address or pool login in the worker field, and the machine picks up the job and starts submitting shares. This network has no quirks such as a mandatory full node or an upfront deposit.

The difficulty lies elsewhere. Almost no working public pools are left. F2Pool left on December 30, 2024 and ViaBTC on June 8, 2026. Among the large operators with working documentation and a live stratum, at the end of July 2026 essentially one remains, Antpool, at stratum plus tcp colon hns.antpool.com port 8025 with a minimum payout of 10 HNS. Before buying hardware, check that the pool is accepting connections today: in a market this narrow that can change without notice.

On noise and siting. The Antminer HS3 at 2,079 W is a full size industrial machine with the airflow and noise level to match. It cannot go into a flat, no matter what enclosure tricks you try. You need a dedicated technical space with cold air intake and organised hot air removal, a dedicated breaker on the line and headroom on the incoming supply. Of the Goldshell line, the only thing acceptable at home is the desktop HS Box at 230 W, and even then with a caveat about the constant whine of small fans.

Home mining versus industrial mining does not even compete here: the two sit on opposite sides of the break even line. The electricity threshold for the best machine on the network is around three cents per kilowatt hour, while household tariffs in Spain and in the EU generally start at about ten. An industrial site with its own generation or a very cheap contract can in theory break even, a home miner cannot. Better to accept that before the purchase than after the first bill.

Profitability and what drives it

Income on this network is the product of four factors, and all four currently work against the miner. The first is daily issuance, which after the halving of July 27, 2026 stands at 72,000 HNS for the whole world. The second is your share of the network hashrate. The third is the coin price, sitting close to all time lows. The fourth is your electricity tariff. Network fees can be left out of the calculation: in blocks they amount to fractions of a coin and do not move the result.

The order of magnitude takes one step to work out. Divide daily issuance by the current network hashrate and you get the output per terahash per day. At a hashrate of roughly 720 to 740 TH/s that is about a hundred HNS per terahash per day. Then multiply by your own hashrate and by the price, and compare against your electricity cost, which for current machines on this algorithm runs from five and a half to over eleven kilowatt hours a day per terahash. The ratio between those two numbers is the entire economics of Handshake.

The honest conclusion, unsoftened. At any European electricity tariff, mining Handshake at the end of July 2026 is unprofitable for every model without exception. The break even point for the best machine on the network sits at around three cents per kilowatt hour, and the other models need one and a half to two and a half cents. Tariffs like that do not exist at retail in the EU. We will not calculate a payback period for you where there is none: if a hardware seller shows you a flattering Handshake projection, almost certainly it uses the old reward of 1,000 HNS per block.

Open the profitability calculator →

Mining pools

Pool	Payout scheme	Fee	Note
Antpool	Per pool schedule	Not published separately for HNS	Effectively the only large operator with live documentation and a working stratum at the end of July 2026. The address is stratum plus tcp colon hns.antpool.com port 8025, the default minimum payout is 10 HNS and credits are daily to the address you give. The exact payout scheme and fee for HNS specifically are not stated in the public documentation, so it is worth checking with support.
F2Pool	Closed	Not applicable	The HNS pool closed on December 30, 2024 at 00:00 UTC, announced on December 25. Balances were paid out within a week and access to earnings reports stayed open until February 5, 2025. Reference sites that still list F2Pool as a working option for HNS are more than a year and a half out of date.
ViaBTC	Closed	Not applicable	Mining and deposits stopped on June 8, 2026 in UTC plus 8, announced on June 1. Report exports were available until June 7, and payouts, withdrawals and conversions closed on June 30. Funds not withdrawn by that date were declared unrecoverable by the operator.
Poolin	Listed as PPS	About 1 percent per old data	It appears in HNS reference lists, but we could not confirm that the connection was live at the end of July 2026. The data in the aggregators that list it is stale: the same pages still show the reward of 1,000 HNS per block that was cut in July 2026.
DxPool	Listed as PPS	About 3 percent per old data	Same situation: the pool is in the old lists for Blake2B plus SHA3, but there is no live confirmation that it works as of the build date of this page. Check stratum availability yourself before configuring hardware.
HashPool	Listed as PPLNS	About 1 percent per old data	A specialist pool from the first wave of Handshake support. As in the previous cases its current status is unconfirmed, and the reference data about it dates from before the halving.
Solo mode	Solo	Depends on the operator	Running your own hsd node and mining straight into it. At a network hashrate of roughly 720 to 740 TH/s a single 9 TH/s machine finds a block on average about once every two months. That is useful for the network but useless for income planning: the spread between payouts is measured in months.

Pools are the main practical problem of Handshake mining at the end of July 2026. In a year and a half the network lost its two biggest operators, F2Pool in December 2024 and ViaBTC in June 2026, and essentially one reliably confirmed option is left. The figures for Poolin, DxPool and HashPool come from aggregators that show the cancelled reward of 1,000 HNS per block on the same page, which is why we flag them as unconfirmed. Check stratum availability before you buy hardware, not after. A reminder on the schemes: PPS pays for submitted shares regardless of the pool luck and suits a small fleet, while PPLNS pays for blocks actually found and gives a wider spread over a short horizon.

Wallets

Category	Wallets
Main desktop	Bob Wallet, the official graphical wallet of the project with a built in full hsd node, DNS record management and participation in name auctions
Full node and CLI	hsd, the daemon and full network node with a built in command line wallet. The canonical option for anyone who wants full control and solo mining
Hardware	Ledger Nano S and Nano X through Bob Wallet. Trezor is not supported
Alternatives	Apex, a wallet and toolset for resolving Handshake names
Exchange accounts	Keeping coins on an exchange is acceptable only for the duration of a trade. Given how few liquid venues for HNS are left, the concentration risk here is higher than usual
What is missing	Mainstream wallets such as Trust, Exodus and MetaMask do not support HNS and, judging by the trajectory of the project, never will

A practical point that sets this network apart: the wallet here is needed not only for coins but for names. A top level domain lives on your address as a special output with a covenant, and every operation on it, bidding at auction, revealing the bid, registering, updating DNS records and renewing, is done through transactions from the wallet. That is why Bob Wallet is built not like a payment client but like a zone control panel with a node inside.

The second point concerns anyone who kept names on external services. After the Namecheap registrar closed the direction on June 10, 2026 and Namebase wound down in the same month, automatic name renewal stopped. A name left on a shuttered platform without renewal eventually returns to circulation and goes to a fresh auction. If you own Handshake domains, move them into your own wallet and renew them yourself: there is nobody else to rely on.

Where to buy and how to store

This section is for reference only, we do not trade the coin and do not give investment advice. Even so, the liquidity situation on HNS has to be stated plainly, because it hits every miner directly: what you mine has to be sold. At the end of July 2026, according to a major aggregator, the coin has only two tracked trading pairs, and both are on one exchange, CoinEx: HNS against USDT and HNS against BTC. Their combined daily turnover is measured in thousands of dollars, not millions.

What that means in practice. With an order book like that, even one machine daily output can move the price, and trying to sell a month of accumulation will almost certainly produce noticeable slippage. This is a separate risk that large coins do not carry: you can calculate your yield in coins correctly and still fail to turn that money into fiat. The venue that historically offered the best HNS liquidity, the built in Namebase exchange, closed in early 2026.

On security the rules are standard and particularly apt here. Do not leave what you mine on an exchange any longer than a trade requires: with all of the coin liquidity concentrated on one venue, its problems automatically become yours. Withdraw to your own wallet, keep the seed phrase offline, and for meaningful amounts use a hardware wallet through Bob Wallet. Exchange names appear here for reference and are not a recommendation.

Strengths and weaknesses

Pros

The idea behind the network is original and addresses a real problem: the DNS root zone is governed today by a single organisation, and a proof of work alternative to that model genuinely makes sense.
No owning organisation, no treasury and no premine for investors. The network cannot be switched off or forced to change policy: there is nobody to give the order.
The deflationary mechanism works off paper. Around 109.7 million HNS have already been burned in name auctions, roughly every seventh coin ever issued.
The hashrate barrier to entry is extremely low: a network at 720 to 740 TH/s is the equivalent of about eighty Antminer HS3 units. Solo mining is theoretically meaningful here, unlike on large networks.
Hardware on the secondary market is very cheap precisely because mining is unprofitable. For an experiment, for learning or for using up surplus heat, that is a cheap way in.
Retargeting difficulty step by step after every block smooths out hashrate swings, which matters for a small network and guards against sharp drops in block time.

Cons

Mining is unprofitable at any European tariff. The break even point for the best machine is around three cents per kilowatt hour, and lower still for the rest.
After the halving of July 27, 2026 the entire daily issuance of the network is worth about a hundred dollars, shared across the whole global fleet.
Liquidity is almost absent: two trading pairs on one exchange and daily turnover in the thousands of dollars. Selling what you mine without slippage is hard.
The infrastructure is winding down. In a year and a half F2Pool, ViaBTC, the Namecheap registrar and the Namebase platform all left. One working pool is effectively left.
The Antminer HS3 is single purpose and switches to nothing else. If mining HNS stops making sense for good, the machine has no fallback use.
No new hardware models since December 2022. The entire fleet is the secondary market, without warranty and with spare part problems.
A Handshake name does not open in an ordinary browser without a special resolver. In six years that limitation has not been removed, and the domains have no mass use.

What the coin is used for

The primary and, in essence, the only use of the network is owning a top level domain. Win the auction and you get the whole TLD, free to run a zone under it and hand out second level domains on your own terms. There is no intermediary, no accreditation and no annual contract with a registry: the right is proven by a record in the chain and renewal is an ordinary transaction. According to the explorer, at the end of July 2026 the chain holds roughly 12.9 million unique names, about 6.5 million of them registered, with more than 101,000 auctions running at once.

The second use is technical: certificates without certificate authorities. The owner of a name publishes DNSSEC and DANE records directly in the chain, binding a specific key to their domain. A client that trusts the chain verifies the match on its own and needs no signature from a commercial certificate authority. That closes off a whole class of attacks based on compromising or coercing an authority.

And the limitation that rules out mass adoption. A Handshake name does not resolve over ordinary DNS. To open such a site you need either your own hsd node, or the lightweight hnsd resolver, or a public resolver such as HDNS.io or NextDNS with support enabled, or the hns.to gateway, or a browser extension. In six years not one major browser and not one mass market provider has added support by default, and there is no sign of that changing.

Real world cases do exist, but there are not many: personal sites and blogs on Handshake names, reachable through the resolvers listed above, and infrastructure integrations where Handshake zones are served on ordinary DNS hosting. There was an attempt to bring names to mass users through a large registrar: Namecheap sold them from 2022 and shut the direction down completely on June 10, 2026. That is the best indicator of where the project stands on practical demand right now.

Network statistics

Parameter	Value
Price	About 0.0016 dollars, aggregators spread from 0.00159 to 0.00165
Market cap	On the order of 1.08 to 1.11 million dollars
Daily exchange turnover	About 6,000 to 6,700 dollars across all venues together
Trading pairs	Two, both on CoinEx: against USDT and against BTC
Network hashrate	From 720 to 737 TH/s depending on the tracker
Difficulty	About 99.17 T
Chain height	On the order of 340,590 blocks
Block reward	500 HNS, cut on July 27, 2026 at block 340,000
Daily issuance	72,000 HNS, about 115 dollars for the entire global fleet
In circulation	About 681.9 million HNS against a total issue of roughly 791.6 million
Burned in auctions	Roughly 109.7 million HNS
All time price high	0.8519 dollars in May 2021, down by more than 99.8 percent

All the figures were taken on July 31, 2026 from the chain explorer, difficulty trackers and market aggregators, and they change constantly: price and turnover within hours, hashrate and difficulty within a day. The spread between sources is about two percent on hashrate and about four on price, which is why we give ranges rather than a single value. A separate warning about market cap: one of the large aggregators showed a price and turnover several times higher than the rest on the date of capture, and we treat that data as polluted and leave it out of the table. We do not publish the number of public network nodes: it could not be confirmed from a trusted source. Daily transactions estimated from recent blocks come to roughly 1,100 to 1,200, but that is an estimate from observed block fill, not a figure from official statistics. For income calculations use fresh calculator data rather than this snapshot.

Network security

Handshake security rests on the same work economics as any PoW network, but the scale here is very modest. A hashrate of roughly 720 to 740 TH/s is the equivalent of about eighty Antminer HS3 machines. A fleet like that costs tens of thousands of dollars on the secondary market, not hundreds of millions, and the low coin price only pushes that barrier down further. The honest conclusion: a 51 percent attack here is affordable in principle, and what protects the network is the absence of a motive rather than the price of the attack.

A separate and, today, sharper risk is pool concentration. After F2Pool left in December 2024 and ViaBTC in June 2026, essentially one large operator is confirmed to be working. That means a significant share of the network hashrate physically converges on a single point, and the question is no longer the theoretical cost of an attack but the simple fact that a miner has no alternative. For the network this is the worst state of decentralisation in its history.

At the protocol level the design is nonetheless careful. Step by step difficulty retargeting after every block makes it impossible to rock the network with bursts of power, and the XOR with a secret pool mask closes off block withholding: a dishonest participant cannot submit shares to the pool while holding back found blocks, because they cannot tell one from the other. Splitting the header into a pre header and a sub header lets light resolvers verify name records without downloading the whole chain.

No major protocol level incidents involving loss of funds have been publicly recorded in the whole life of the network since February 2020. Both soft forks, the one in February 2024 and the one in August 2025, went through without a chain split. We do not quote the number of public nodes here: the public node tracker was unreachable on the build date of this page, and we will not invent a figure. Indirect signs of network health are the chain size, roughly 23.87 gigabytes, and the stable block time of about 9.85 minutes.

Compatibility and requirements

Parameter	Value
Hardware type	Only ASICs for Blake2B plus SHA3. Graphics cards formally work but have made no economic sense since 2020
Models in the catalogue	One, the Antminer HS3 at 9 TH/s and 2,079 W
Models outside the catalogue	The Goldshell HS line: HS Box, HS5, HS6, HS6 SE
Single purpose hardware	The Antminer HS3 computes Handshake only. Goldshell HS units switch over to Siacoin
Our firmware	None. The algorithm is not on the support list for our tooling
Third party firmware	No mature public builds exist for this algorithm
Power supply	Factory unit included. The Antminer HS3 draws up to 2,079 W on the line, the Goldshell HS6 up to 3,250 W
Electrical line	A dedicated breaker per machine, three phase supply from three machines up
Cooling	Air on every model. No hydro or immersion versions were ever built for this algorithm
Noise and siting	Industrial level on every full size model, so a living space is ruled out. The only home option is the Goldshell HS Box at 230 W

Comparison with similar coins

Model	Hashrate	Power draw	Efficiency	Electricity threshold	Second coin
Antminer HS3	9 TH/s	2,079 W	231 J/TH	About 0.029 dollars per kWh	None, Handshake only
Goldshell HS6	10.6 TH/s	3,250 W	307 J/TH	About 0.022 dollars per kWh	Siacoin
Goldshell HS6 SE	8.2 TH/s	3,400 W	415 J/TH	About 0.016 dollars per kWh	Siacoin
Goldshell HS5	5.4 TH/s	2,650 W	491 J/TH	About 0.014 dollars per kWh	Siacoin
Goldshell HS Box	0.235 TH/s	230 W	979 J/TH	About 0.007 dollars per kWh	None, Handshake only

The electricity threshold is the tariff at which a machine breaks even on the power bill alone, without counting the cost of the hardware itself, the premises or maintenance. The calculation is made for the end of July 2026 at a reward of 500 HNS per block and a price of about 0.0016 dollars, so the halving is already included. Pay attention to the last column: it matters more than efficiency today. The Antminer HS3 is the most efficient of them all but can only do Handshake, whereas the Goldshell HS machines lose on joules yet switch over to Siacoin and therefore have a meaningful use beyond this network. The Goldshell figures come from manufacturer specifications, as these models are not in the asic.es catalogue.

Outdated data and common errors

The block reward is 1,000 HNS. The most widespread and most expensive error at the end of July 2026. The second halving happened on July 27, 2026 at block 340,000 and the reward was cut to 500 HNS. Major profitability calculators and pool aggregators were still counting on the old reward on the date we checked, which means they were doubling the estimated income. If someone shows you a Handshake projection, the first thing to look at is which reward it uses.

The halving is in April or July, the date is approximate. Different reference sites gave different forecasts: some named April 28, 2026, others July 22 or 23. Every forecast turned out to be inaccurate. The actual date is confirmed by the direct card for block 340,000 in the chain explorer: July 27, 2026, 08:59:28, block subsidy 500.00 HNS. The same explorer dates the first halving to April 28, 2023, 10:30:38.

The supply cap is 2.04 billion HNS. Almost a year out of date. The soft fork of August 28, 2025 closed the unclaimed developer airdrop and cut the cap by roughly 56 percent. Market aggregators at the end of July 2026 show a cap on the order of 961.3 million HNS. The figure of 889 million from the press releases of the time also circulates and does not match the exchange data, so any of these numbers deserves caution.

F2Pool and ViaBTC mine HNS. Both pools are closed. F2Pool stopped HNS on December 30, 2024 at 00:00 UTC and ViaBTC on June 8, 2026. Even so, pool lists on popular reference sites still show them as working options, sometimes on the same page that carries the cancelled reward of 1,000 HNS per block.

Handshake domains are bought through Namecheap. The direction closed on June 10, 2026: new registrations, renewals, transfers and management are unavailable. Yet plenty of reviews and guides, including recent ones, still describe buying Handshake names through the ordinary registrar interface. Existing domains keep working but are not renewed automatically.

Namebase is the main liquidity venue for HNS. Namecheap sold Namebase on January 28, 2026, the built in exchange is closed, orders were cancelled and funds returned to users, and in June of the same year the platform announced its wind down along with the end of name renewals. Pointing at Namebase as a working HNS market today is misleading.

Network hashrate is around 1.1 PH/s and difficulty is 155 M. Numbers like these are still sitting on the algorithm pages of some aggregators. That is the state of the network in earlier years. According to trackers at the end of July 2026 the hashrate is 720 to 737 TH/s and the difficulty is around 99.17 T. The gap against the stale data is more than one and a half times on hashrate.

The Goldshell HS6 has an efficiency of about 756 J/TH. That figure appears in reference sites and does not agree with the manufacturer specification. 10.6 TH/s at 3,250 W gives 307 J/TH. A value around 490 J/TH belongs to the older HS5, and only the desktop HS Box comes anywhere near a thousand. The error matters in practice: it shifts the break even threshold by almost two and a half times.

Handshake uses a single Blake2B hash function. Inaccurate. Per the hsd source, every attempt runs three calls: Blake2B with a 512 bit output, SHA3-256, and a final Blake2B with a 256 bit output over the concatenation of the first two. On top of that the result is XORed with a 32 byte pool mask. Describing the algorithm as plain Blake2B or plain SHA3 does not match the implementation.

Two thirds of open source developers received the airdrop. This confuses the allocated share with the share actually claimed. Two thirds of the supply, 1.36 billion HNS, were reserved for the distribution, but over more than five years about 2.9 percent of recipients claimed it. That is precisely why the 2025 soft fork closed the unclaimed part, rather than because the distribution finished successfully.

Questions and answers

What is Handshake and what is it for

It is a proof of work blockchain that replaces the DNS root zone and the registry of top level domains. Names are distributed not through ICANN and registrars but through open on chain auctions.

Who created Handshake

The project was put together by Joseph Poon, co author of the Lightning Network, Andrew Lee of Purse and Private Internet Access, Christopher Jeffrey, author of bcoin, plus Boyma Fahnbulleh and Steven McKie. Mainnet launched on February 3, 2020.

Does the network have an owner or a foundation

No. There is no development company, no foundation and no treasury. The hsd code is open and changes come from the community through soft forks. That is both a strength of the project and the reason nobody was left to replace the infrastructure that shut down.

Which algorithm does Handshake use

The Blake2B plus SHA3 pair. Every attempt computes three hashes: Blake2B-512 and SHA3-256 over the pre header, then Blake2B-256 over their concatenation. The result is XORed with a 32 byte pool mask.

Why is the pool mask needed

It protects against block withholding. The pool operator keeps the mask secret, so a miner cannot tell a real found block from an ordinary share and cannot hold back a block while continuing to collect share payouts.

When was the last halving and what is the reward now

The second halving happened on July 27, 2026 at block 340,000. The reward fell from 1,000 to 500 HNS. The first halving was on April 28, 2023 at block 170,000.

When is the next halving

At block 510,000, roughly in 2029. The reward will fall from 500 to 250 HNS. The interval between halvings is 170,000 blocks, which is about 3.23 years.

Why does difficulty change

Difficulty adapts to the actual power on the network in order to hold the target of 144 blocks a day. Unlike many networks, the retarget here happens after every block rather than once per interval, which smooths out hashrate swings.

Can Handshake be mined at home

Technically yes, economically no. The Antminer HS3 at 2,079 W is an industrial machine with the noise to match, and the break even point sits at around three cents per kilowatt hour against European household tariffs that start at ten.

Which ASIC is best for Handshake

On efficiency the best is the Antminer HS3, 9 TH/s at 2,079 W, that is 231 J/TH. But it can only do Handshake. Goldshell HS machines are worse on joules yet switch over to Siacoin, and today that counts for more.

Is it true that the HS3 cannot switch to other coins

True. The Antminer HS3 computes Handshake exclusively and has no fallback use. That is the key difference from the Goldshell HS line, which also runs Siacoin. Keep it in mind when buying on the secondary market.

Have any new machines come out for this algorithm

No. The last models, the Goldshell HS6 and HS6 SE, shipped in December 2022. As of the end of July 2026 no new machine for Blake2B plus SHA3 has appeared in more than three and a half years.

Which pool should I choose

There is effectively no choice. F2Pool closed the direction on December 30, 2024 and ViaBTC on June 8, 2026. The confirmed working option is Antpool, with a minimum payout of 10 HNS. Check stratum availability before buying hardware.

Can I install third party firmware

No firmware for this algorithm exists, neither from us nor on the third party market. In six years of the network no mature public builds have appeared, because there is nothing to recoup the investment from. Treat offers to sell you firmware for the HS3 as a reason to be suspicious.

How can I reduce the power draw of the machine

Without firmware only external levers are available: stable power quality, a lower air temperature at the intake and a pool with a lower fee. The platform gives nothing comparable to undervolting on SHA-256.

How loud is the Antminer HS3

As loud as any full size industrial machine drawing over two kilowatts: a constant noise at the level of an industrial fan. Unacceptable in a living space with any amount of enclosure soundproofing.

What power supply do I need

The factory unit that comes with the machine. The Antminer HS3 draws up to 2,079 W on the line, the larger Goldshell HS models up to 3,250 W. Each machine needs a dedicated breaker, and from three machines up a three phase supply makes sense.

Will mining Handshake pay for itself

At European tariffs, no. The break even point on electricity alone for the best machine on the network is about 0.029 dollars per kilowatt hour. We do not calculate a payback period where a machine does not even break even on power.

What happens to income if the price falls

It falls in direct proportion, because the reward is fixed in coins. At a price of about 0.0016 dollars there is no cushion left: any drop in the rate increases the daily loss by the same share.

Can Handshake be mined on a graphics card

Technically yes, the official hs-miner has CUDA and OpenCL backends. Practically no: since 2020 the gap against ASICs has been orders of magnitude, and at the current price even purpose built machines do not pay for their electricity.

Does Handshake have merge mining

No. No other network uses the Blake2B plus SHA3 pair in this form, so there is no combined mining with a second coin at the protocol level.

How many confirmations does a transaction need

There is no single protocol standard, each venue sets its own requirement. With a ten minute block, six confirmations is about an hour. Check the requirement of the specific exchange or service before withdrawing.

What are network fees like on Handshake

Extremely low. On recent blocks the total fees amount to fractions of a single coin per block, which in dollars is well under a cent. The network is very lightly loaded and there is no competition for block space.

How should I store mined HNS

The main option is Bob Wallet with a built in full hsd node, which you also need for name operations. For large amounts use a Ledger through Bob Wallet. Mainstream wallets such as Trust or MetaMask do not support HNS.

Where can I buy hardware for Handshake

Only on the secondary market: no new machines for this algorithm have shipped from the factory since December 2022. The asic.es catalogue carries the Antminer HS3. Before buying, assess the economics soberly and remember that the HS3 does not switch to other coins.