My 10x 170HX GPU "Ai" Server Is Actually Making Money Now.

Описание: ⚡ Crypto Miner Bros https://www.cryptominerbros.com
⚡ Sign up for Simplepod here: https://simplepod.ai/ref?user=redpanda
💻 Sign up for VastAi here: https://cloud.vast.ai/?ref_id=502537
💻 RTX 5090's https://geni.us/nOmSUfh (PAID LINK)
💾 Tangem Cold Storage Crypto Wallet https://geni.us/rpmtangem use code RPM for 10% off!I just unlocked 10 CMP 170HX / 178X GPUs and turned them into a massive 640GB VRAM AI server — and now people are actually renting it for AI inference!

After getting the new CMP unlock working, these GPUs are now showing up as 64GB of unlocked VRAM each. With 10 GPUs in one server, that's a total of 640GB of VRAM available for workloads like large language models and AI inference.

I've now had this server online through SimplePod for several days, and the results have been pretty interesting. I've had people rent the full 10-GPU server, test different workloads, run AI inference, download hundreds of gigabytes of data, and even use the server through Hugging Face-related workloads.

At one point, the server was rented at 25¢ per GPU per hour. After SimplePod's 20% fee, the actual earnings are lower, but the potential is still very interesting for GPUs that I originally bought for roughly $150–$170 each.

In this video, I go over:

• Unlocking 10 CMP 170HX / 178X GPUs
• 640GB of total unlocked VRAM
• Using the GPUs for AI inference
• Renting the server on SimplePod
• How much the server actually earned
• SimplePod's 20% platform fee
• Full 10-GPU rentals
• 25¢ per GPU per hour rental pricing
• How much the server could potentially make per day
• Comparing the earnings to RTX 5090 GPUs
• A real AI inference workload running on the server
• Hugging Face workloads
• 400GB data downloads
• 3Gbps fiber internet
• Network bandwidth concerns
• Packet loss and potential networking issues
• PCIe 2.0 x4 limitations
• The possibility of PCIe 3.0 x4 support
• 2TB NVMe storage limitations
• Why I may need 4TB or 8TB storage
• My plans to put even more GPU servers online

This is one of the most interesting developments I've seen with these CMP GPUs. What started as cheap GPUs sitting around is turning into potentially useful AI inference hardware with a huge amount of VRAM.

The big question is: How much can these servers actually make when they're consistently rented?

Let me know what you think in the comments!

🔔 Subscribe for more GPU rental, AI infrastructure, crypto mining, and hardware testing videos.

Chapters
00:00 — I Built Another 10x CMP 170HX AI Server
00:28 — My First 10-GPU Server Is Already Rented
00:59 — The New CMP 170HX Unlock Changes Everything
01:30 — 640GB of Unlocked VRAM for AI Inference
02:02 — Installing 10x 170X GPUs Into the Server
02:24 — Sponsor: Crypto Miner Bros
03:39 — 10x 170X GPUs Finally Installed
04:01 — Getting the Server Ready for SimplePod
04:50 — How Much Is My 170HX Server Making?
05:13 — My 10x 170HX Server Has Been Rented
05:28 — 64GB CMP 170HX GPUs Are Now Unlocked
05:44 — PCIe 2.0 x4 vs PCIe 3.0 x4
06:14 — The 10GbE Network Card Problem
07:12 — How Much Money Is the 10-GPU Server Making?
07:43 — My First Full 10-GPU Rental
08:17 — Renting 10 GPUs for 25¢ Per Hour
08:28 — SimplePod Takes a 20% Fee
09:00 — How Much Can 10x 170HX GPUs Make Per Day?
09:34 — Comparing This to RTX 5090 GPU Rentals
10:17 — Driver Updates and Server Reliability
10:31 — Someone Is Running AI Inference on My GPUs
11:16 — What Is the Renter Actually Doing?
11:24 — 400GB Downloaded in Just 20 Minutes
12:18 — Do I Have Enough Internet Bandwidth?
12:34 — Why Am I Seeing Packet Loss?
13:25 — The Problem With Only 2TB of NVMe Storage
13:57 — Why Didn't They Rent All 10 GPUs?
14:10 — My Plans for More AI GPU Servers

#CMP170HX #GPU #AI #GPURental #AIInference #SimplePod #Mining #CryptoMining #GPUMining #NVIDIA