Impact per GPU-hour
Spread scarce AI compute where it does the most good.
AI compute is the new scarce resource — and the public-good projects that need it most get out-competed by commercial demand. Allocate a pool of donated GPU-hours below; Compute for Good ranks projects by impact per GPU-hour so every hour lands where it matters.
Donated compute pool
100k GPU-hoursAllocated 0 of 100k100k free
0
🌍 Climate
Regional climate downscaling📚 Education
Open tutors for low-resource languages🌊 Disaster
Flood & wildfire early warning🧬 Disease
Antibiotic resistance screening0
Ranked by impact per GPU-hour
Where each donated hour does the most good. Transparent model — see the weights.
Efficiency vs optimum 0%
Tip: tap ★ Optimize for impact to close the gap.
| # | Project | Impact / GPU-hr | Hours | Share |
|---|---|---|---|---|
| 1 | 🧬 Protein folding for neglected diseases | 0.0 | — | — |
| 2 | 🌊 Flood & wildfire early warning | 0.0 | — | — |
| 3 | 🌍 Regional climate downscaling | 0.0 | — | — |
| 4 | 🧬 Antibiotic resistance screening | 0.0 | — | — |
| 5 | 🌍 Crop disease detection for smallholders | 0.0 | — | — |
| 6 | 📚 Open tutors for low-resource languages | 0.0 | — | — |
Total impact delivered: 0 points across 0 funded project(s). Illustrative.
Ask AI
AI can be wrong. Not medical, legal, or financial advice.
Commit to the public ledger
Each commit is hash-chained to the last, so the record of how scarce compute was spent is tamper-evident.
allocation-ledger.chain
$ ledger --statusno allocations committed yetcommit an allocation to start the tamper-evident chain.