Kaspa, the KS5 ASIC and ZKAS: assessing merged mining

Kaspa uses GHOSTDAG to order blocks found in parallel instead of discarding every simultaneous block as a conventional chain often does. Its goal is rapid payment confirmation while leaving transaction validation and mining open to participants. Proof of work, rather than an appointed block producer, helps determine the ordering of transactions. The practical use case is a fast, publicly verifiable settlement network that people and applications can access without permission.

Kaspa's mainnet launched in November 2021 without a premine or presale allocation. A Rust rewrite of its node enabled the Crescendo upgrade and ten blocks per second on mainnet. Those are deployed capabilities, not a future target. Whether they translate into sustained value depends on users, wallets, applications and demand for transactions. A fast protocol alone cannot guarantee either appreciation of KAS or revenue from an ASIC.

The Toccata upgrade activated on 30 June 2026 and added programmable UTXOs to mainnet. Covenants, proof verification and application sequencing create a basis for vaults, flexible wallets and other services. Yet a protocol feature is not the same as a mature app ecosystem. Developer tooling and some application models are still developing, so their eventual adoption should be assessed separately from what the network can technically do today.

DAGKnight is a proposed next consensus upgrade, while still higher throughput belongs to later plans. Kaspa's long-term case rests on combining quick confirmation, open proof-of-work security and useful programmability. The counterweights are competing networks, software complexity, possible defects, concentration of mining power and the need for real transaction demand as miner subsidies decline. A roadmap milestone should never be inserted into a mining forecast as though it had already shipped.

BITMAIN lists the KS5, version KS1-20, as a kHeavyHash miner rated at 20 TH/s and 3000 W at the wall under specified conditions. Its nominal efficiency is 150 J/TH. It needs 220–277 V power, Ethernet and suitable industrial airflow; the typical noise rating is 76 dB. The manufacturer allows approximately 3% variation in hashrate and 5% in power and efficiency. Actual performance of an individual unit must therefore be checked on site.

Do not confuse the 20 TH/s KS5 with the 21 TH/s KS5 Pro already shown separately in the catalogue. They have different specifications and prices. Before buying a particular KS5, request its serial number, condition, measured hashrate and wall power, firmware, repair history and warranty terms. The ASIC is built for kHeavyHash; changing its pool address cannot make it mine Bitcoin's SHA-256 or Equihash coins.

ZKas is a separate network and coin whose mainnet launched on 26 July 2026. It combines a blockDAG and kHeavyHash with mandatory private transfers based on Orchard and Halo 2. Its implementation supports merged mining: a compatible pool can apply work performed by a kHeavyHash ASIC to blocks on both networks. That creates the possibility of a second reward stream, but it is not an automatic feature of every ordinary Kaspa pool.

The project's own pool specifies a concrete configuration: the worker name carries a zkas: address and the password carries a kaspa: address. Under that pool's rules, ZKAS goes to the shielded wallet and any KAS blocks found are paid to the supplied Kaspa address. Leaving the password as x or omitting the KAS address can redirect that KAS to the pool. Check firmware and Stratum compatibility, fees, maturity and both payout addresses with a small live test. Other pools can apply different rules.

Kaspa reduces its block subsidy smoothly each month, adding up to roughly a halving per year. If the KAS price stays flat, that lowers new coins available per unit of hashrate over time; growth in total network hashrate further reduces one KS5's share. Pool luck, fees, rejected shares and uptime also matter. ZKAS adds the difficulty of a second network, its own emission path, wallet availability and the depth of actual exchange order books.

At 3000 W, a continuously running unit consumes 72 kWh per day. At €0.085/kWh, electricity costs €6.12 per day or €183.60 for 30 uninterrupted days. The executable value of KAS and ZKAS actually received must exceed that amount plus any other contracted costs before an operating surplus exists; the €1700 device price is separate. Model several price and difficulty scenarios. For ZKAS, use the price at which your intended volume could actually be sold, rather than the last trade on a thin market.

As of 29 September 2026, ForMiners offers a limited number of Antminer KS5 20 TH/s units for €1700 each together with hosting in Estonia. The stated electricity tariff for this offer is €0.085/kWh. At the rated 3000 W, that implies €6.12 per day of electricity during uninterrupted operation. The purchase price and metered power are separate charges; actual use and invoicing depend on the specific unit and hosting agreement.

Availability is limited, so request written confirmation before payment. The quote should identify each unit's condition, taxes, delivery, commissioning, minimum hosting term, what the tariff covers, servicing, access to pool statistics and the return procedure. We can help choose KAS-only or KAS/ZKAS merged mining, configure addresses and check initial payouts when those services are included in the agreed scope. That turns a headline price into an operational plan.

A KS5 may suit someone seeking kHeavyHash capacity with a defined entry price and a known hosting power tariff. One owner may prefer accumulating KAS with fewer operational decisions; another may want additional exposure to ZKAS. The choice depends on risk tolerance, custody arrangements and the ability to verify pools and markets. ZKAS is an independent project, not an upgrade to Kaspa, and its rewards do not automatically improve every KAS miner's economics.

ZKas is young. Its own site says the novel core has not yet had an independent external security audit. Shielded payouts are harder to inspect through a public explorer, while small exchanges can have wide spreads and constrained withdrawals. Before assigning a payback period, test the wallet, a small payout, deposits and withdrawals, accepted shares and the ability to sell the intended volume. Repeat the calculation regularly: the power tariff can be known, while price, difficulty and rewards change.