{
  "id": 14286635,
  "title": "Tinygrad refreshes Tinybox with an 'old' EPYC CPU running DDR4 RAM to feed four AMD 9070XT — slaps a $11,000 price tag on a new 'computer for deep learning'",
  "url": "https://urgent.news/2026/10/13/tinygrad-refreshes-tinybox-with-an-old-epyc-cpu-running-ddr4-ram-to",
  "topic": "tech",
  "section": "Tech",
  "published": "2026-10-13T20:30:00.000Z",
  "source": {
    "name": "TechRadar",
    "slug": "techradar",
    "url": "https://www.techradar.com/pro/tinygrad-refreshes-tinybox-with-an-old-epyc-cpu-running-ddr4-ram-to-feed-four-amd-9070xt-slaps-a-usd11-000-price-tag-on-a-new-computer-for-deep-learning"
  },
  "original_language": "en",
  "account": "Tinygrad unveils Tinybox, a workstation equipped with four AMD Radeon RX 9070 XT graphics cards, priced at $11,000. This new model keeps DDR4 memory and PCIe 4.0, alongside a water-cooled server processor for deep-learning tasks. The standard configuration features a 32-core AMD EPYC processor, 32 GB of DDR4 memory, and a 1 TB boot SSD. Its enclosure, measuring 12\" x 19.25\" x 18\", weighs 42 pounds, and boasts a 1,700 W power supply. Customers can also opt for a G5 platform with DDR5 memory and PCIe 5.0 connectivity. The company justifies its size by recalling the era of underpowered PCs. Noise levels remain below 50 dB through low-speed fans and liquid cooling. The G4 platform starts at $7,000, while the G5 configuration is $12,000. Adding four Radeon RX 9070 XT graphics cards adds $4,000, making the complete G4 configuration $11,000 before optional upgrades.",
  "summary": "The Tinybox offers DDR4 and DDR5 platforms, Nvidia graphics options, optional RAID storage and water cooling, with configurations tailored to demanding deep-learning workloads.",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}