I'm looking to put together a budget estimate for deploying a private version of DeepSeek R1 and V3 that would serve a company with 3,500 employees. Any suggestions or guidance on how to approach this? Thanks!
4 Answers
One option is to invest in a couple of Mac Studios with 512GB of VRAM to handle the processing needs.
I found this article that might be helpful: https://www.aifun.cc/en/deepseek-local-deployment-cost.html. It shares insights into local deployment costs for DeepSeek.
Based on my experience at Open Router, I’d recommend checking out DeepInfra or Novita as they provide solid services. For a company your size, connecting directly to the DeepSeek API could be a good approach too.
You might want to reach out to providers like DeepInfra, Modal, or other AI companies; several of them offer private deployment solutions. They could give you specific quotes based on your needs.
OP meant Private inference.