Skip to content
View original post on X: Yuchen Jin
Published
Pick60/100

DeepSeek open-sources DeepInfer and reports 155 trillion tokens/day per cluster

AIDeepSeek open-sourced DeepInfer, a large-scale cloud inference framework, and used it to run a DeepSeek V4.1 Flash service. The image states that one cluster serves 155 trillion tokens per day, with $38 revenue per card per hour at off-peak prices.

AIWhy it matters.

The post links DeepSeek's disclosed throughput and revenue per GPU-hour to its open-sourced DeepInfer framework, showing what a large inference service reports about its economics.

Yuchen JinVerified on X
@Yuchenj_UW
View on X

“A single cluster currently serves 155 trillion tokens/day, generating $38 in revenue per GPU-hour (at off-peak pricing).”

Huge flex from DeepSeek. Not only did they open-source their inference engine, but they also open-sourced their revenue numbers.

They didn't disclose which GPUs they're using here, but DeepInfer supports both NVIDIA Blackwell and Huawei Ascend.

Even assuming they're renting B300s at $7 per GPU-hour, that's ~82% gross margin.

Source details

Yuchen Jin · x.comPublished · added here

Original title: “A single cluster currently serves 155 trillion tokens/day, generating $38 in revenue per GPU-hour (at off-peak pricing).”

Topics

DeepSeekChina AIBusiness news