{"version":1,"type":"story","url":"https://digestai.news/story/nvidias-gb300-fp4-submission-trails-googles-tpuv7-on-cost-efficiency","json":"https://digestai.news/story/nvidias-gb300-fp4-submission-trails-googles-tpuv7-on-cost-efficiency.json","markdown":"https://digestai.news/story/nvidias-gb300-fp4-submission-trails-googles-tpuv7-on-cost-efficiency.md","slug":"nvidias-gb300-fp4-submission-trails-googles-tpuv7-on-cost-efficiency","headline":"Nvidia’s GB300 FP4 submission trails Google’s TPUv7 on cost efficiency","summary":"SemiAnalysis’s October 19, 2026 report compares Nvidia’s GB300 FP4 entry with Google’s TPUv7 on an agentic inference benchmark used by external customers. The analysis finds Nvidia’s submission 1.64 times worse in performance per dollar, while Google’s TPUv7 scores between 1.42 and 1.67 times better. Nvidia’s entry runs in FP4 (four‑bit) format, whereas Google’s uses FP8 (eight‑bit), making FP4 the more aggressive, lower‑precision option. The GB300 NVL72 chip can deliver up to 15 PFLOPS of dense FP4 compute per GPU and 1.1 exaFLOPS per rack, but the benchmark used a full 72‑GPU configuration. Google achieved its results with only four TPUv7 chips.\n\nThe report notes that the cost‑efficiency gap is most pronounced on medium‑to‑high interactivity tasks typical of agentic AI, where long conversations and multi‑step reasoning drive many inference calls. SemiAnalysis expects further counter‑submissions from Nvidia as it works to close the gap. For investors and enterprises, the findings suggest that Google’s architecture may offer a cheaper scaling path for agentic workloads, especially when starting with a small chip count.","keyPoints":["Nvidia’s GB300 FP4 submission is 1.64× worse in performance per dollar than Google’s TPUv7 on agentic inference, per SemiAnalysis.","Google’s TPUv7 achieved 1.42–1.67× better cost efficiency using four chips, while Nvidia used a full 72‑GPU rack.","GB300 NVL72 offers up to 15 PFLOPS FP4 per GPU (1.1 exaFLOPS per rack) but still trails Google’s FP8‑based TPUv7."],"whyItMatters":"The cost‑efficiency gap shows that Google’s TPUv7 can run agentic AI workloads cheaper and with far fewer chips, prompting enterprises to consider alternatives to Nvidia’s dominant GPUs for long‑context inference.","category":{"slug":"hardware","name":"Hardware & Compute","url":"https://digestai.news/category/hardware"},"entities":{"companies":["Nvidia","Google"],"models":["GB300 NVL72","TPUv7"],"people":[]},"firstPublishedAt":"2026-10-19T12:11:00Z","updatedAt":"2026-10-19T12:11:00Z","sourceCount":1,"hasPrimarySource":false,"sources":[{"outlet":"cryptobriefing.com","title":"Nvidia’s GB300 FP4 submission still trails Google’s TPUv7 on cost efficiency","url":"https://cryptobriefing.com/nvidia-gb300-fp4-trails-google-tpuv7","publishedAt":"2026-10-19T12:11:00Z","type":"press","primary":false,"lead":true}],"sourceNotes":null,"discussions":[],"thread":null,"cite":{"text":"Digest AI, \"Nvidia’s GB300 FP4 submission trails Google’s TPUv7 on cost efficiency\", 19 October 2026, https://digestai.news/story/nvidias-gb300-fp4-submission-trails-googles-tpuv7-on-cost-efficiency","publisher":"Digest AI","title":"Nvidia’s GB300 FP4 submission trails Google’s TPUv7 on cost efficiency","datePublished":"2026-10-19T12:11:00Z","url":"https://digestai.news/story/nvidias-gb300-fp4-submission-trails-googles-tpuv7-on-cost-efficiency"},"generatedBy":"Written by Digest AI's editorial model from the linked sources; the sources are the record.","license":"Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse"}