DeepSeek inference system reveals: 545% amazing cost profit margin announced

DeepSeek opened an official account on Zhihu and released a technical article "Overview of DeepSeek-V3/R1 Inference System", which disclosed the optimization details and cost-profit margin information of its model inference system for the first time, marking the official end of the highly-watched "DeepSeek Open Source Week".

The article introduces two major optimization goals of the DeepSeek-V3/R1 inference system: "larger throughput and lower latency". To achieve these goals, DeepSeek adopts large-scale cross-node expert parallelism (EP) technology, although this increases system complexity. The article focuses on how to use EP technology to increase batch size, hide transmission time, and achieve load balancing.

Of particular note, DeepSeek rarely discloses its cost and profit margin data. The article disclosed: "Assuming the cost of GPU rental is US$2 per hour, the total cost is $87,072 per day. If all tokens are calculated based on the pricing of DeepSeek R1, the theoretical total revenue per day is $562,027, and the cost profit margin is 545%.