For the cost of 8 years of $200 Claude subs, you can force your team to use shitty quantized versions of DeepSeek v4 Flash running at 30tps.
Incredible deal!
TDavid HendricksonXView on XLocal AI is good enough for a startup. A 10-person startup built its own 128GB-VRAM AI server for $18,000.
One Threadripper box sitting locally with
🎮 4× Radeon AI PRO R9700 32GB
💾 128GB total VRAM
🧠 Qwen3.8 + DeepSeek V4 Flash
👥 16 concurrent AI sessionsAnd they're actually pushing it.
Qwen3.8-Flash-Next with 16 simultaneous sessions:
📚 48K context/user
⚡ 538.5 tps aggregate decodeAt 128K context per user
⚡ 365 tps aggregate decodeThe GPUs are even undervolted to under 210W each.
Link in ALT
1
2 comments
And what do you access those Claude subs on? A $7000 mac mini?
On a random $500 laptop I installed Ubuntu on and leave in a closet :)