{
  "docId": "019dd923-5eff-723e-9be4-898b452b08b7",
  "docSlug": "e3a2aefd8126020d",
  "documentTitle": "2025 Q1FY25 Quarterly",
  "authorId": "NVIDIA",
  "authorName": "NVIDIA",
  "documentKindSlug": "consulting-deck",
  "documentKindLabel": "Consulting deck",
  "sourceTypeSlug": "strategy_consulting",
  "sourceTypeLabel": "Strategy consulting",
  "presentationDate": null,
  "orientation": "landscape",
  "aspectRatio": 1.778,
  "pageNumber": 19,
  "pageCount": 25,
  "prevPage": 18,
  "nextPage": 20,
  "slideType": "case_study",
  "function": "quantify_impact",
  "density": "dense",
  "nDataPoints": 4,
  "notes": "The slide uses a bar chart to compare H100 and H200 performance on Llama 2 70B.",
  "elementsJson": [
    "headline_text",
    "action_title",
    "bullet_list",
    "bar_chart_grouped",
    "callout_box",
    "footnote"
  ],
  "metadataConfidence": 1,
  "imagePath": null,
  "slideHref": "/slides/019dd923-5eff-723e-9be4-898b452b08b7/19",
  "deckHref": "/decks/019dd923-5eff-723e-9be4-898b452b08b7",
  "deckJsonHref": "/decks/019dd923-5eff-723e-9be4-898b452b08b7.json",
  "deckAnchorHref": "/decks/019dd923-5eff-723e-9be4-898b452b08b7#slide-19",
  "components": [
    {
      "bbox": {
        "h": 0.05,
        "w": 0.3,
        "x": 0.6,
        "y": 0.78
      },
      "kind": "callout",
      "text": "Up to 45% Faster Llama 2 70B Inference in MLPerf",
      "attrs": null,
      "subkind": "primary",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "deaf7c0f-39bf-4670-9986-262d22170dd4",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.5,
        "w": 0.3,
        "x": 0.6,
        "y": 0.18
      },
      "kind": "chart",
      "text": "New MLPerf Llama 2 70B Results",
      "attrs": null,
      "subkind": "bar-grouped",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "3b5f02b9-8ced-448c-83bc-b760713cac66",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.4,
        "w": 0.4,
        "x": 0.07,
        "y": 0.33
      },
      "kind": "list",
      "text": "In the latest MLPerf benchmark for AI inference, NVIDIA Hopper swept every test of AI inference on a per accelerator basis... NVIDIA was the only company to submit results on every workload... NVIDIA TensorRT-LLM... boosted the performance... NVIDIA H200 GPUs... set a record on the Llama 2 70B benchmark",
      "attrs": null,
      "subkind": "bullet",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "ff643bc2-d5c3-48b0-867c-69d116376625",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": null,
      "kind": "metric",
      "text": "Tokens/Second: 31,712",
      "attrs": null,
      "subkind": "primary",
      "toolName": "Quantification",
      "toolSlug": "quantification",
      "confidence": null,
      "componentId": "019dd952-78b5-733e-810b-15bd828d5dbe",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.05,
        "w": 0.35,
        "x": 0.58,
        "y": 0.92
      },
      "kind": "source-note",
      "text": "MLPerf Inference v4.0 data center results retrieved from www.mlperf.org on March 27, 2024, from entries 4.0-0062, and 4.0-0068.",
      "attrs": null,
      "subkind": null,
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "a8012766-5c0b-40a1-be47-28a2bfc2269c",
      "frameworkName": null,
      "frameworkSlug": null
    },
    {
      "bbox": {
        "h": 0.15,
        "w": 0.4,
        "x": 0.07,
        "y": 0.07
      },
      "kind": "title",
      "text": "NVIDIA H200 Tensor Core GPUs and NVIDIA TensorRT-LLM Set MLPerf LLM Inference Records",
      "attrs": null,
      "subkind": "headline",
      "toolName": null,
      "toolSlug": null,
      "confidence": null,
      "componentId": "8d7e4543-422b-43b6-9a4b-a6e49fdafa58",
      "frameworkName": null,
      "frameworkSlug": null
    }
  ],
  "metrics": [],
  "tools": [],
  "frameworks": [],
  "arcBeats": [
    {
      "to": 20,
      "from": 18,
      "beatId": "fd38dad8-ac50-4763-98e1-186588a588d3",
      "arcName": "The Sequoia Pitch",
      "arcSlug": "sequoia-pitch",
      "beatName": "Competition",
      "beatSlug": "sequoia-pitch-competition",
      "evidence": "The document presents case studies and key takeaways, highlighting NVIDIA's competitive advantage",
      "position": 3,
      "confidence": 0.8,
      "parentBeatName": "Evidence",
      "parentBeatSlug": "evidence"
    }
  ],
  "loops": [
    {
      "to": 20,
      "from": 18,
      "name": "Cost Of Inaction",
      "slug": "27-cost-of-inaction",
      "bestFor": "Urgent budget requests, compliance, risk mitigation",
      "matchId": "b5d8a448-7791-40b0-888f-57f0fac09c88",
      "evidence": "The document presents case studies and key takeaways, highlighting the benefits of NVIDIA's solutions",
      "position": 2,
      "objective": "To highlight the importance of NVIDIA's solutions",
      "structure": "The Status Quo -> The Hidden Costs Accumulating -> The Future State of Inaction -> The Tipping Point",
      "confidence": 0.6,
      "description": "Quantify what happens if the audience does nothing"
    }
  ],
  "imagePathAlt": null,
  "thumbSrc": null,
  "thumbSrcAlt": null,
  "locked": true
}