Machines fail sometimes.
Your mining keeps running.
An ASIC is industrial hardware pushed hard around the clock — sooner or later, something needs attention. The difference is what happens next. At home, a failure means weeks of silence: diagnosing, shipping, waiting on a repair shop. In a Cuverse data center, automated monitoring catches it in minutes and engineers repair the unit meters from the rack — and on Cloud Infrastructure, equivalent capacity keeps your output running while they do. You'll likely learn about the incident from your dashboard, after it's already being handled.
01
The same failure, three very different weeks
Mining output after a failure
¹ Illustrative, based on published turnaround times of independent ASIC repair services (typically 1–4 weeks of bench time and queue, before shipping in both directions). Manufacturer warranty repair including international shipping commonly runs 4–10 weeks. Actual times vary by fault, parts, and season.
Same failure, same physics — the unit breaks either way. On your own it costs 19 days and ≈ $185 of output; in Hosting Mode the repair happens meters from the rack in days; on Cloud Infrastructure it costs ≈ $2 — not because nothing broke, but because real equivalent capacity from our own fleet is connected the moment your unit goes down, and keeps mining while it's repaired.
02
Five layers between a failure and your output
This isn't one lucky mechanism — it's a stack. Each layer exists so the one above it rarely gets tested.
24/7 automated monitoring
Every unit reports continuously. Anomalies raise an alert in minutes — not when someone happens to look.
Engineers on site
Each data center runs its own service team, around the clock. Diagnosis starts in hours, not after a shipping label.
No shipping, no queue
Repair happens meters from the rack, with parts and bench on site — the weeks a courier and a workshop waitlist would eat simply don't exist.
Equivalent capacity during repair
If a unit breaks down, the lost computing power is replaced with equivalent capacity while it's fixed — output continues.
Cloud InfrastructureWarranty & included maintenance
Routine maintenance to manufacturer requirements is part of the service, with warranty coverage of 12 or 36 months by connection type.
03
Aging is just a failure in slow motion
Breakdowns are the loud problem. The quiet one is drift: a physical machine's output isn't a constant — it dips a few percent below nominal day to day and slowly sags as chips, fans, and thermal paste age. That's normal physics, and good maintenance slows it. But on Cloud Infrastructure there's a cleaner answer: you select a hashrate, and delivering that hashrate is our job. Which unit produces it, how it's maintained, when it's swapped — that's our side of the agreement. Your line stays your line.
Delivered hashrate over 24 months (illustrative)
Running a few percent under nominal never feels like an incident — no alert fires, nothing is "broken". But 2–3% under 270 TH is 5–8 TH quietly missing: roughly 0.0009–0.0014 BTC a year (≈ $55–85 at $60,000/BTC), per machine, every year — losses you're taking without ever seeing them. On Cloud Infrastructure that leak doesn't exist: the capacity you selected is the capacity that mines.
04
Written into the agreement, not the ad copy
Uptime here isn't a slogan — it's a committed minimum in your hosting agreement, measured over the calendar year.
annual uptime, MAX mode — Hosting Mode (12-month placement)
annual uptime, MAX mode — Cloud Infrastructure (36-month placement)
standard warranty by connection type; extended coverage available for Hosting Mode
monitoring and on-site engineering, included in the service — not an add-on
Running ECO? Its pauses are scheduled by design, not failures — ECO follows its own uptime targets. See the MAX & ECO page for how that works. Exact commitments for each mode and placement are set out in your agreement.
And you watch it all live. Hash rate, uptime, and incident status are in your dashboard in real time — the same numbers our monitoring works from. You're not taking anyone's word for it.
05
Guaranteed uptime, and more...
Whichever mining mode you choose, Cuverse platform works for you.
MAX & ECO modes
Two ways to run the same hardware — switch as the market moves.
Payouts to your wallet
Pool → your address, on your schedule. We hold no keys, no coins.
Hosting or cloud platform
Your own miner or one from our fleet — your choice.
Loyalty cashback
Tiers reduce your effective energy cost on both infrastructures.
06
FAQ
- Automated monitoring flags the anomaly within minutes and on-site engineers take the unit into repair — no shipping, no external queue. On Cloud Infrastructure, the lost computing power is additionally replaced with equivalent capacity right away, so your output continues while the unit is fixed. You can follow the status in your dashboard throughout.
- Depends on the connection type. On Cloud Infrastructure, no by design — equivalent capacity covers your selected hashrate while the unit is repaired. On Hosting Mode, your specific machine is assigned to you, so there is a pause while it's fixed — but it's an on-site repair measured in days, not the weeks of shipping and workshop queues you'd face on your own. The annual uptime minimums in your agreement account for this.
- Routine maintenance to manufacturer requirements is included in the hosting service. Breakdown repairs are handled under your warranty coverage — 12 months standard for Hosting Mode (extendable), 36 months for Cloud Infrastructure. The exact coverage terms are in your agreement.
- Hosting Mode: 12 months standard, with extended coverage available. Cloud Infrastructure: 36 months standard — the full placement term. So for Cloud Infrastructure, warranty simply spans everything.
- In MAX mode, the agreement commits to a minimum of 95% annual uptime for Hosting Mode and 99% for Cloud Infrastructure, with operational targets above that. ECO mode is different by design — it runs a scheduled off-peak pattern with its own targets. Full terms per mode are in your agreement.
- All physical hardware drifts with age — that's normal. Included maintenance (cleaning, thermal service, part replacement) keeps it near spec on either connection type. On Cloud Infrastructure you're insulated from it entirely: you selected a hashrate, and delivering that hashrate — whatever it takes on our side — is Cuverse's job under the agreement.
- Your dashboard shows hash rate, uptime, and unit status in real time, so incidents and their resolution are visible as they happen. You're seeing the same telemetry our monitoring acts on.
- Then it's resolved under your warranty terms rather than left as your problem — support contacts you with the resolution. On Cloud Infrastructure, equivalent capacity keeps your output running in the meantime.
- The stack — monitoring, on-site engineers, maintenance, warranty — protects the hardware in both modes. Keep in mind ECO's pauses are scheduled, part of how the mode works, and its uptime is accounted differently in the agreement. A failure is a failure in either mode, and it's handled the same way.
- You can — plenty of people do. But an ASIC at home means industrial noise and heat, residential electricity rates, and every failure being your project: diagnosing, shipping, waiting. Hosting exists to turn all of that into someone else's day job, with committed uptime and repairs happening meters from the rack.
Hardware that's looked after
Professional data centers, committed uptime, and a repair process you'll mostly read about after it's done.
Illustrative model. The simulator is a simplified illustration as of the page date for a single miner. Mining output is modelled as the current network reward of about 0.0000005 BTC per TH per day multiplied by the Bitcoin price you set, holding network difficulty constant within the period. Energy uses the base rates shown, reduced by Diamond-tier loyalty cashback to the effective rates noted; ECO running hours follow a representative off-peak schedule, and the on/off clock above is illustrative. The price path (dip then recovery) is an illustrative scenario, not a prediction. Actual results depend on Bitcoin price, network difficulty, mining output, energy tariffs, your loyalty tier, hardware condition, and configuration — all of which vary over time. ECO can extend the period during which a machine remains economically worth running, but does not guarantee profitability or change physical hardware lifespan. This is not a forecast, an offer, or a guarantee of any result. Energy rates and figures shown are model inputs; current rates and terms on cuverse.com prevail. Mode availability, uptime targets, and scheduling are governed by your service terms.