Post
Hansato Global Sourcing 🌐
hansato.com
did:plc:7h2dbwwdjvycpskqweupw7pg
DeepSeek's efficiency has real consequences for Nvidia's chip demand. Training V3 for $5.6M vs GPT-4's $50–100M is rewriting the cost curve entirely.
When AI is 10x cheaper to run, do you care who's "best"—or who makes your unit economics work?
That's not an AI story. That's a #supplychain story.
2026-07-19T03:31:00.383Z