Regret Selling RTX 5090s Early | Generated by AI
Question: Why does my friend regret selling several hundred RTX 5090s at just over 30,000 (CNY each)?
Answer:
Your friend regrets it because the RTX 5090 price spiked hard after they sold. They sold at ~30K+ CNY each, but prices climbed well above that afterward — so on “several hundred” units, they left a large amount of money on the table (potentially millions of CNY in missed upside).
The driver is the AI/compute demand surge. The 5090’s 32GB GDDR7 and strong FP/tensor throughput make it attractive for local LLM inference and small-scale training, especially in China where datacenter-grade cards (H100/H200/B200) are export-restricted. That pushed consumer high-VRAM cards into a quasi-prosumer-AI market, tightening supply and inflating street prices. Memory price increases across the board (DRAM/GDDR cost pressure in 2025–2026) compounded the squeeze.
So the regret is straightforward: he sold into a rising market and watched the same cards trade meaningfully higher within months. Classic seller’s remorse — exited a scarce, appreciating, demand-driven asset too early.
I’d want to verify the current street price before putting a number on the missed gain. Want me to search for the latest China retail/grey-market 5090 pricing so you can quantify how much he actually “lost”?