{
"$type": "site.standard.document",
"bskyPostRef": {
"cid": "bafyreiccxlpz5pzkuoiixxyo4f23c3hdiw4twfcg5hpycccfhwgpcukbo4",
"uri": "at://did:plc:ajcrkmnlj6rxdk7rltijv227/app.bsky.feed.post/3moarufydgbx2"
},
"coverImage": {
"$type": "blob",
"ref": {
"$link": "bafkreiblwlfx4cx5zruocl6ve2agniiboifcqdyr7iwgmcjateyr2e5yi4"
},
"mimeType": "image/png",
"size": 1517267
},
"path": "/tech-industry/artificial-intelligence/ai-cryptomining-networks-320-000-rtx-3090-class-gpus-allegedly-burn-112-megawatts-of-power-on-zero-useful-ai-computation-pearls-gpus-are-doing-random-matrix-math-study-claims",
"publishedAt": "2026-06-14T11:30:00.000Z",
"site": "https://www.tomshardware.com",
"tags": [
"Artificial Intelligence",
"Tech Industry"
],
"textContent": "A preprint claims Pearl’s AI mining network consumes 320,000 GPU-equivalents and 112 MW while producing no verified useful AI computation.",
"title": "AI cryptomining network's 320,000 RTX 3090-class GPUs allegedly burn 112 megawatts of power on ‘zero useful AI computation’ — GPU rental costs jump 38%, but Pearl’s cards are doing random matrix math, study claims"
}