Nvidia H200 SXM
Nvidia H200 SXM earns $78.45 per day renting on the AI GPU marketplace at a median rate of $3.34/h, drawing 700W. At $0.1/kWh electricity, the daily power cost is $1.68.
Daily projection RENTAL
| Period | /Day | /Month |
|---|---|---|
| Income | $80.13 | $2,403.86 |
|
Cost
$0.1/kWh
|
$1.68 | $50.40 |
| Profit | $78.45 | $2,353.46 |
| Coin | Algorithm | Income | Cost | Profit |
|---|---|---|---|---|
RVN
Ravencoin
|
KAWPOW
88Mh · 700.0W
|
$0.47 | $1.68 | $-1.21 |
|
—
|
Ethash
260Mh · 700.0W
|
— | $1.68 | — |
- Model
- H200 SXM
- TDP
- 700 W
- Type
- GPU
- Vendor
- Nvidia
How Nvidia H200 SXM earns renting on the AI GPU marketplace
| Provider | GPU | Income | Cost | Profit |
|---|---|---|---|---|
|
RunPod
GPU marketplace
|
H200 SXM
$3.590/h ·
1 offer
|
$80.13
|
$1.68 |
$78.45
★
Visit →
|
Nvidia H200 SXM earns $78.45/day renting to AI workloads — meaningfully more than the $-1.21/day it would make mining KAWPOW. Mining and rental are mutually exclusive use modes, so the AI marketplace is the obvious play here.
| Market | Algorithm | Profit /day | |||
|---|---|---|---|---|---|
NiceHash
seller 24h-weighted avg
|
KAWPOW
0.00000006854 BTC/M/d
|
$-1.19
$0.49 income · $1.68 cost
Visit →
|
|||
|
MRR floor
10% rented · matches cheapest seller
|
KAWPOW
0.00000012151 BTC/M/d
|
$-0.81
$0.87 income · $1.68 cost
Visit →
|
|||
|
MRR recent
last 10 rentals · actual clearing price
|
KAWPOW
0.00000018736 BTC/M/d
|
$-0.33
★
$1.35 income · $1.68 cost
Visit →
|
|||
|
MRR asking
aspirational — seller wish, not matched
|
KAWPOW
0.00000012865 BTC/M/d
|
$-0.76
$0.92 income · $1.68 cost
|
|||
|
Ethash
|
|||||
|
NiceHash
seller 24h-weighted avg
|
Ethash
0.00000001451 BTC/M/d
|
$-1.37
$0.31 income · $1.68 cost
Visit →
|
|||
| Rigs × Qty | Share | Rev /rig/day | Cost /rig/day | Profit /rig/day | Total profit /day |
|---|---|---|---|---|---|
| — | — | — | — | — | — |
ROI calculator for Nvidia H200 SXM
Model payback, electricity, and first-year return for this rig.
The line crosses $0 on the day you break even. Everything above is pure profit.
| Month | Earned (mo) | Cost burned (mo) | Cumulative earned | Cumulative cost | Net | % ROI |
|---|
Yearly emissions by energy source
Based on the rig's annual power draw and the carbon intensity of common grid mixes.
| Energy source | CO₂e / yr |
|---|---|
| Wind | 66.53 kg |
| Nuclear | 72.58 kg |
| Hydroelectric | 145.15 kg |
| Geothermal | 229.82 kg |
| Solar | 272.16 kg |
| Biofuels | 1,391.04 kg |
| Gas | 2,963.52 kg |
| Coal | 4,959.36 kg |
Estimates only — actual emissions vary by hardware, cooling, and grid mix.
What does that actually mean?
At the world-average grid intensity of about 475 g CO₂e/kWh, Nvidia H200 SXM running 24/7 for a year releases about 2,873 kg of carbon dioxide equivalent. Here's what that looks like in everyday terms:
Where you plug in matters
Electricity is not one thing. A kilowatt-hour from a coal plant carries roughly 820 g of CO₂; the same kilowatt-hour from a hydro reservoir carries about 24 g. That's a 34× difference — large enough that Nvidia H200 SXM's annual footprint swings from roughly 4,959 kg on coal-heavy grids down to about 145 kg on hydro-dominated grids. The single biggest lever a miner has on their carbon footprint is choosing where to plug in.
Regions commonly used for low-carbon crypto mining include Quebec and British Columbia (hydro-dominated, typically <50 g CO₂/kWh), Iceland and Norway (geothermal + hydro, often <30 g), Paraguay (Itaipú hydro), and parts of the US Pacific Northwest. Coal-heavy grids — Kazakhstan, Inner Mongolia, Poland, parts of Australia — sit at the opposite end, often above 700 g CO₂/kWh.
Some operators also reduce their net impact by using otherwise-wasted energy: flare gas at oil wells (burning methane that would be vented anyway), curtailed renewables (wind or solar that the grid can't absorb), or behind-the-meter hydro during off-peak hours. These arrangements can drop effective emissions below the local grid average because the energy would have been wasted or flared without the mining load.
How to reduce this rig's footprint
- Pick a greener ASIC. The efficiency column above matters as much as the grid: a 15 J/TH rig emits roughly half the CO₂ of a 30 J/TH rig for the same hashrate.
- Choose a low-carbon host. Data centres advertising hydro, geothermal, or nuclear power typically sit at <100 g CO₂/kWh.
- Look for stranded or curtailed energy. Flare-gas miners, wind-curtailment co-location, and off-peak hydro arrangements use energy that would otherwise be wasted.
- Use heat recovery. Capturing the heat for greenhouse agriculture, pool heating, or district warmth offsets fossil-fuel heating that would have been burned anyway.
- Time-shift your uptime. In grids with high daytime solar, running more during the day and less at night lowers your effective intensity even if you don't switch providers.
- Purchase verifiable offsets. Treat this as a last resort, not a substitute — and favour additional, permanent, third-party-verified projects (Gold Standard, Verra VCS).
Frequently asked questions
Yearly electricity use = rig power (W) × 24 × 365 ÷ 1000. We multiply that by each row's grid intensity in grams CO₂-equivalent per kWh and convert to kilograms. Intensities are representative averages — real emissions depend on your specific utility mix, time of day, and local transmission losses.
It depends almost entirely on where the electricity comes from. A single rig plugged into hydro in Quebec emits less over a year than an average family's two cars in a month. The same rig on a coal-dominated grid can exceed that in a few days. The hardware is the same — the grid is what changes the answer.
Network-wide estimates vary by methodology; the Cambridge Centre for Alternative Finance's Bitcoin Electricity Consumption Index is the most widely cited reference. As of recent reporting, the network's sustainable-energy share has grown as more hashrate migrates to hydro, wind, solar, and stranded-gas sites. This page just estimates a single rig — for the big picture, CCAF's dashboard is the best source.
Not directly. The rig draws the same wattage regardless of which pool it joins or how difficulty trends — so its electricity use, and therefore its emissions, stay constant. Those factors change revenue, not power consumption.
Daily projection RENTAL
| Period | /Day | /Month |
|---|---|---|
| Income | $80.13 | $2,403.86 |
|
Cost
$0.1/kWh
|
$1.68 | $50.40 |
| Profit | $78.45 | $2,353.46 |
| Coin | Algorithm | Income | Cost | Profit |
|---|---|---|---|---|
|
RVN
Ravencoin
|
KAWPOW
88Mh · 700.0W
|
$0.47 | $1.68 | $-1.21 |
|
—
|
Ethash
260Mh · 700.0W
|
— | $1.68 | — |
- Model
- H200 SXM
- TDP
- 700 W
- Type
- GPU
- Vendor
- Nvidia
How Nvidia H200 SXM earns renting on the AI GPU marketplace
| Provider | GPU | Income | Cost | Profit |
|---|---|---|---|---|
|
RunPod
GPU marketplace
|
H200 SXM
$3.590/h ·
1 offer
|
$80.13
|
$1.68 |
$78.45
★
Visit →
|
Nvidia H200 SXM earns $78.45/day renting to AI workloads — meaningfully more than the $-1.21/day it would make mining KAWPOW. Mining and rental are mutually exclusive use modes, so the AI marketplace is the obvious play here.
| Market | Algorithm | Profit /day | |||
|---|---|---|---|---|---|
|
NiceHash
seller 24h-weighted avg
|
KAWPOW
0.00000006854 BTC/M/d
|
$-1.19
$0.49 income · $1.68 cost
Visit →
|
|||
|
MRR floor
10% rented · matches cheapest seller
|
KAWPOW
0.00000012151 BTC/M/d
|
$-0.81
$0.87 income · $1.68 cost
Visit →
|
|||
|
MRR recent
last 10 rentals · actual clearing price
|
KAWPOW
0.00000018736 BTC/M/d
|
$-0.33
★
$1.35 income · $1.68 cost
Visit →
|
|||
|
MRR asking
aspirational — seller wish, not matched
|
KAWPOW
0.00000012865 BTC/M/d
|
$-0.76
$0.92 income · $1.68 cost
|
|||
|
Ethash
|
|||||
|
NiceHash
seller 24h-weighted avg
|
Ethash
0.00000001451 BTC/M/d
|
$-1.37
$0.31 income · $1.68 cost
Visit →
|
|||
| Rigs × Qty | Share | Rev /rig/day | Cost /rig/day | Profit /rig/day | Total profit /day |
|---|---|---|---|---|---|
| — | — | — | — | — | — |
ROI calculator for Nvidia H200 SXM
Model payback, electricity, and first-year return for this rig.
The line crosses $0 on the day you break even. Everything above is pure profit.
| Month | Earned (mo) | Cost burned (mo) | Cumulative earned | Cumulative cost | Net | % ROI |
|---|
Yearly emissions by energy source
Based on the rig's annual power draw and the carbon intensity of common grid mixes.
| Energy source | CO₂e / yr |
|---|---|
| Wind | 66.53 kg |
| Nuclear | 72.58 kg |
| Hydroelectric | 145.15 kg |
| Geothermal | 229.82 kg |
| Solar | 272.16 kg |
| Biofuels | 1,391.04 kg |
| Gas | 2,963.52 kg |
| Coal | 4,959.36 kg |
Estimates only — actual emissions vary by hardware, cooling, and grid mix.
What does that actually mean?
At the world-average grid intensity of about 475 g CO₂e/kWh, Nvidia H200 SXM running 24/7 for a year releases about 2,873 kg of carbon dioxide equivalent. Here's what that looks like in everyday terms:
Where you plug in matters
Electricity is not one thing. A kilowatt-hour from a coal plant carries roughly 820 g of CO₂; the same kilowatt-hour from a hydro reservoir carries about 24 g. That's a 34× difference — large enough that Nvidia H200 SXM's annual footprint swings from roughly 4,959 kg on coal-heavy grids down to about 145 kg on hydro-dominated grids. The single biggest lever a miner has on their carbon footprint is choosing where to plug in.
Regions commonly used for low-carbon crypto mining include Quebec and British Columbia (hydro-dominated, typically <50 g CO₂/kWh), Iceland and Norway (geothermal + hydro, often <30 g), Paraguay (Itaipú hydro), and parts of the US Pacific Northwest. Coal-heavy grids — Kazakhstan, Inner Mongolia, Poland, parts of Australia — sit at the opposite end, often above 700 g CO₂/kWh.
Some operators also reduce their net impact by using otherwise-wasted energy: flare gas at oil wells (burning methane that would be vented anyway), curtailed renewables (wind or solar that the grid can't absorb), or behind-the-meter hydro during off-peak hours. These arrangements can drop effective emissions below the local grid average because the energy would have been wasted or flared without the mining load.
How to reduce this rig's footprint
- Pick a greener ASIC. The efficiency column above matters as much as the grid: a 15 J/TH rig emits roughly half the CO₂ of a 30 J/TH rig for the same hashrate.
- Choose a low-carbon host. Data centres advertising hydro, geothermal, or nuclear power typically sit at <100 g CO₂/kWh.
- Look for stranded or curtailed energy. Flare-gas miners, wind-curtailment co-location, and off-peak hydro arrangements use energy that would otherwise be wasted.
- Use heat recovery. Capturing the heat for greenhouse agriculture, pool heating, or district warmth offsets fossil-fuel heating that would have been burned anyway.
- Time-shift your uptime. In grids with high daytime solar, running more during the day and less at night lowers your effective intensity even if you don't switch providers.
- Purchase verifiable offsets. Treat this as a last resort, not a substitute — and favour additional, permanent, third-party-verified projects (Gold Standard, Verra VCS).
Frequently asked questions
Yearly electricity use = rig power (W) × 24 × 365 ÷ 1000. We multiply that by each row's grid intensity in grams CO₂-equivalent per kWh and convert to kilograms. Intensities are representative averages — real emissions depend on your specific utility mix, time of day, and local transmission losses.
It depends almost entirely on where the electricity comes from. A single rig plugged into hydro in Quebec emits less over a year than an average family's two cars in a month. The same rig on a coal-dominated grid can exceed that in a few days. The hardware is the same — the grid is what changes the answer.
Network-wide estimates vary by methodology; the Cambridge Centre for Alternative Finance's Bitcoin Electricity Consumption Index is the most widely cited reference. As of recent reporting, the network's sustainable-energy share has grown as more hashrate migrates to hydro, wind, solar, and stranded-gas sites. This page just estimates a single rig — for the big picture, CCAF's dashboard is the best source.
Not directly. The rig draws the same wattage regardless of which pool it joins or how difficulty trends — so its electricity use, and therefore its emissions, stay constant. Those factors change revenue, not power consumption.