GB2-0030
① SA Source
- Source: 開啟完整 SA 文章
- Section:
NVLink Scale Up Interconnect - Line hint:
295
Context Before
Source: SemiAnalysis
Evidence
Nvidia claims that if they used optics with transceivers, they would have needed to add 20kW per NVL72 rack.
Context After
As such, this is the reason Nvidia chose to use 5184 copper cables, which is a much cheaper, less power hungry, and more reliable option. Each GPU has 900GB/s unidirectional bandwidth. Each differential pair (DP) is capable of transmitting 200Gb/s in 1 direction thus it will take 72 DP per GPU for both directions. Since there are 72 GPUs per NVL72 rack, that would mean there is 5184 differential pairs. Each NVLink cable contains 1 differential pair thus there are 5184 cables.
This is a massive increase in copper content generation on generation. In a bit of a whiplash, we have seen some investors estimate that there is around 216k/NVL72 rack but this is completely wrong.
② Atomic Claim
Nvidia 聲稱若 NVL72 使用 optical transceivers，將增加約 20kW/rack 的功耗。
- Epistemic Mode:
ATTRIBUTED - Mapping Status:
COMPLETE
③ Semantic Frame
{
"attribute": "optical_transceiver_power_addition",
"context_nodes": [
{
"id": "02_companies/NVDA",
"label": "Nvidia"
}
],
"entity": {
"id": "04_knowledge_base/GB200 NVL72",
"label": "GB200 NVL72"
},
"frame_type": "ATTRIBUTE",
"qualifiers": {
"condition_text": null,
"numeric_mentions": [
"20kW"
],
"temporal_mentions": []
},
"value": {
"numeric_mentions": [
"20kW"
],
"value_text": "Nvidia claim: +20kW per rack"
}
}④ Canonical Entity Mapping
| Role | Surface Label | Canonical Target |
|---|---|---|
| entity | GB200 NVL72 | 04_knowledge_base/GB200 NVL72 |
| context_0 | Nvidia | NVDA |
⑤ Human Review
請在 Properties 逐項確認：
- 原文 → Atomic Claim 是否忠實
- Atomic Claim → Semantic Frame 是否忠實
- Canonical Entity mapping 是否正確
- Epistemic mode 是否保留原文語氣
- 最後選擇
review_action
Review state
Markdown 內文不是正式 approval。只有 Apply bridge 寫入的 Decision Ledger event 才是正式決策。