DeepSeek discloses the cost of its V3/R1 inference system

Made by AIWritten by AI (Claude) from the source below. Check anything important against the source.How it's made

Disclosed the cost of its V3/R1 inference system — a theoretical cost-profit margin of 545%

Why it matters

The first primary source on the unit economics of inference. DeepSeek itself cautioned that this is an upper bound assuming all usage is paid, and that actual revenue is much lower

Source — Primary sources first; secondary reports are labelled “Press”

Officialgithub.comhttps://github.com/deepseek-ai/open-infra-index/blob/main/202502OpenSourceWeek/day_6_one_more_thing_deepseekV3R1_inference_system_overview.md

Added in Week 35, 2026 · See that week

On the map — ⑧ Inference Infrastructure line

  1. Previous stopV1 engine in alpha
  2. This recordDeepSeek discloses the cost of its V3/R1 inference system
  3. Next stopGemma 3 QAT int4
See it on the map

Same week

OpenAI releases GPT-4.5 as a research previewOfficialKakao open-sources Kanana Nano 2.1BOfficialZoom researchers present Chain of DraftPaperClaude 3.7 SonnetOfficial

Nearby on the same axis

DeepSeek holds Open Source WeekOfficialEpoch AI measures how fast inference prices are fallingOfficialNVIDIA announces Dynamo at GTCOfficialvLLM releases its V1 engine in alphaOfficial