IceRiver KS2

The IceRiver KS2 is a 2 TH/s kHeavyHash ASIC that mines Kaspa (KAS), pulling roughly 1,200 W at the wall for an efficiency near 600 J/TH. It is a first-generation, air-cooled prosumer unit — effectively a doubled-up KS1 in the same chassis, with twice the hashrate and twice the power draw.

Specifications at a glance

The numbers below are the figures D-Central tracks for the KS2 in our miner database. IceRiver, unlike Bitmain or MicroBT, does not publish a full silicon datasheet, so the spec sheet stays at the system level rather than the chip level.

Chip and hashboard architecture

The KS2 runs on a custom IceRiver ASIC built specifically for kHeavyHash. This is where the honesty matters: IceRiver keeps its die specifications proprietary. It does not publish the process node, the per-board chip count, or the voltage-domain layout the way Bitmain and MicroBT do for their Bitcoin hardware. Any « chip count » you see quoted for the KS series online is unverified, so we don’t repeat it as fact. What we can state with confidence is that the KS2 is a self-contained appliance: a single air-cooled chassis with the controller, hashing silicon, and an integrated power supply behind one 200–240 V AC inlet.

Architecturally, a Kaspa ASIC is a different animal from a Bitcoin SHA-256 miner. kHeavyHash is not a single hash function — it sandwiches a matrix-multiplication step (the « heavy » part) between Keccak-based hashing rounds. In silicon terms that means the datapath has to combine hashing cores with multiply-accumulate arrays, rather than the long, repetitive Boolean-logic pipelines that make up an SHA-256 engine. That difference is exactly why a Bitcoin miner cannot mine Kaspa and vice versa, and why the KS2’s custom ASIC has no cross-compatibility with the BM-series chips found in Antminers.

Power and control inside the box follow the same broad principles as any modern ASIC: the chips are grouped into chains, each chain is fed by regulated voltage rails, and the controller polls the chain for live status. As with every ASIC, voltage is regulated at the domain level across groups of chips, not individually per chip — a detail that matters once you get into board-level diagnostics.

Real-world power and efficiency

At a nameplate 2 TH/s and roughly 1,200 W, the KS2 lands at about 600 J/TH. Because IceRiver’s KS units carry their power supply inside the chassis, the ~1,200 W figure is close to true wall draw rather than a DC-side rating, so plan your circuit around 1,200 W plus headroom. On a 240 V circuit that is a modest 5 A; on shared 120 V infrastructure you’ll want a dedicated 20 A circuit and to respect the 80% continuous-load rule.

One important caveat on the efficiency number: J/TH for kHeavyHash is not comparable to J/TH for SHA-256. They are entirely different algorithms with their own difficulty scales, so 600 J/TH here tells you nothing about how a 16 J/TH Bitcoin miner performs — the units only share a name. Within the Kaspa hardware market, the KS2 was a first-generation efficiency point; later IceRiver generations have pushed the J/TH figure lower, which is the main reason the KS2 now competes on purchase price rather than on running cost.

Tuning headroom is limited compared with the Bitcoin world. Our ASIC power-profiles database catalogs hundreds of undervolt and overclock profiles, but those live in the Antminer ecosystem where mature autotuning firmware exists. The KS2 runs IceRiver’s stock firmware, which does not expose the same per-domain frequency-and-voltage tuning, so in practice the unit operates at or near its factory point. If you want to lower its cost per hash, that comes from cheaper power and good cooling, not from a tuning preset.

Heat is a real output: ~1,200 W of electricity becomes roughly 4,094 BTU/h of heat. That can be put to work as supplemental space heating in a cold shop or garage during winter, turning a heating bill into hashrate. At ~65 dB, though, the KS2 is genuinely loud — closer to a vacuum cleaner than a fan — so it belongs in a garage, basement, or outbuilding rather than a living space.

Firmware and software compatibility

The KS2 ships with IceRiver’s own stock firmware, managed through a built-in web dashboard over Ethernet, where you set pool, worker, and network details and watch per-chain status. There is no honest third-party firmware story here, and we won’t pretend otherwise: the well-known custom Bitcoin firmwares — BraiinsOS+, VNish, LuxOS — are SHA-256 projects and do not run on a Kaspa ASIC. D-Central’s own DCENT_OS is likewise focused on Bitcoin Antminer hardware, so it is not a fit for the KS2 either. For a Kaspa miner, stock firmware is the firmware.

On the operations side there is better news. The open-source pyasic library recognizes IceRiver as a first-class vendor, so the KS2 can be polled for hashrate, temperature, and fan data inside fleet-monitoring tooling alongside Antminers and Whatsminers. For anyone running a mixed-coin farm, that means the KS2 doesn’t have to be a monitoring blind spot. Pool connectivity itself is straightforward stratum to any Kaspa pool, configured directly in the stock dashboard.

Common faults and troubleshooting

The KS2’s failure modes are the familiar ASIC ones. The most common starting points:

• Hashrate below target or a dead chain — the controller reports fewer active chips than expected, usually pointing to one underperforming or failed hashboard chain.
• Fan errors or thermal shutdown — air-cooled units live and die by airflow; dust-clogged heatsinks and worn fan bearings are the usual culprits, especially in a heater-duty install.
• No-boot or no-network — controller, PSU, or Ethernet/config faults that stop the unit from reaching a pool.
• Power-supply faults — because the PSU is integrated, a supply fault takes the whole appliance offline rather than letting you swap a separate unit.

Start with the structured, symptom-first walkthrough in our ASIC fault finder, which maps observed behavior to likely root cause. For chain- and chip-level faults, the KS series uses a dedicated test fixture (the IceRiver KS hashboard tester) that connects to the board’s data port, powers the chain, and scans for the specific chip where the signal drops — the same family of bench tooling repair shops use to localize a fault before reflow.

Repair and longevity

D-Central has run an in-house ASIC repair bench since 2016, and a Kaspa miner is repairable like any other ASIC: faults are localized to a board, then to a chain, then to a component, and addressed with board-level and BGA rework rather than thrown away. The KS2 in particular benefits from professional repair because the IceRiver after-market is thinner than Bitmain’s — replacement boards and chips are harder to source on the open market, and the KS-specific test fixture is its own purchase. That makes a working repair relationship more valuable, not less.

If your KS2 is down or degraded, our ASIC repair service covers diagnosis through component-level fix. Day to day, the levers that extend its life are the unglamorous ones: keep the intake clean, keep ambient temperature inside the 5–40 °C window, give it clean stable power, and don’t run it pinned in a sealed, dusty room. Heat and dust kill ASICs; airflow and clean power keep them earning.

Who the KS2 is for, and buying it

The KS2 makes sense for someone who specifically wants exposure to Kaspa and has a spot that can absorb a loud, hot, 1,200 W appliance — a garage, shop, or basement, not a bedroom. As a home unit it scores in the middle of our scale: viable, but it is a heater-grade citizen rather than a quiet living-room device. Profitability is entirely a function of the KAS price, network difficulty, and your power cost; at 600 J/TH the margin is thin when power is expensive and far healthier when you are heating a space you’d otherwise pay to warm anyway.

It is not the right tool if your goal is Bitcoin — the KS2 cannot touch SHA-256, so a Kaspa ASIC and a Bitcoin ASIC are not substitutes. If you’re weighing the KS2 against other hardware, or against a Bitcoin-first strategy, browse the full lineup and current availability in our ASIC miner catalog. The KS2 launched around a 5,000 USD MSRP, but Kaspa-hardware pricing is volatile and second-hand units trade well below launch, so judge any purchase on today’s price against today’s KAS economics.

Where the KS2 sits in the IceRiver lineup

Credit where it’s due: IceRiver was one of the manufacturers that opened the Kaspa ASIC market, taking KAS mining off GPUs and onto purpose-built silicon across a full product ladder — KS0, KS1, KS2, KS3 (and KS3L), and the later KS5L/KS5M. The KS2’s place on that ladder is clear from its own DNA: it is the KS1 doubled, in the same chassis.

Stepping up from the KS1 to the KS2 buys you density — twice the hashrate from one chassis and one network drop — without changing the efficiency. The later KS3 and KS5 generations are where IceRiver moved the efficiency needle, so today the KS2 is best understood as a capable first-generation Kaspa workhorse: a sensible entry point when bought right and kept cool, and a machine worth repairing rather than retiring when a board goes down.