{"product_id":"atlas-300i-duo-ai-inference-accelerator-card-96gb-280-tops","title":"Atlas 300I Duo AI Inference Accelerator Card — 96GB, 280 TOPS INT8, PCIe Gen4","description":"\u003cp\u003eThe Atlas 300I Duo is a PCIe Gen4 inference accelerator for servers running AI vision, video analytics and media workloads. Each card carries 96GB of on-card memory and rates at 280 TOPS of INT8 compute, so a single slot handles large batch sizes and many concurrent inference streams without spilling to host memory. On-card AI compute cores, an integrated CPU and dedicated hardware video codec engines sit on the same board, which lets it decode, analyze and encode video in one pass rather than passing frames back to the host.\u003c\/p\u003e\u003cp\u003eIt suits teams building smart-video platforms, retail and traffic analytics, content moderation and other high-throughput inference services where rack density and power budget matter. Because the codec, CPU and AI cores are integrated, you consolidate what would otherwise be separate transcoding and inference hardware into one PCIe card.\u003c\/p\u003e\u003cul\u003e\n\u003cli\u003e96GB on-card memory for large models and high concurrency\u003c\/li\u003e\n\u003cli\u003e280 TOPS INT8 inference performance\u003c\/li\u003e\n\u003cli\u003ePCIe Gen4 host interface\u003c\/li\u003e\n\u003cli\u003eIntegrated AI cores, on-card CPU and hardware video codec engines\u003c\/li\u003e\n\u003c\/ul\u003e\u003cp\u003eBacked by a 1-year manufacturer warranty.\u003c\/p\u003e","brand":"BarEquip.Pro","offers":[{"title":"Default Title","offer_id":67947233149027,"sku":"BEP-0865","price":58612.0,"currency_code":"THB","in_stock":true}],"thumbnail_url":"\/\/cdn.shopify.com\/s\/files\/1\/0643\/6726\/3843\/files\/4efd6160-a8f6-479d-9bec-95014bd5f9f2.png?v=1794867860","url":"https:\/\/barequip.pro\/ru\/products\/atlas-300i-duo-ai-inference-accelerator-card-96gb-280-tops","provider":"BarEquip.Pro","version":"1.0","type":"link"}