UBB Store

Sipeed K3 Pico-ITX 32GB LPDDR5 RISC-V AI Board 60 TOPS - USA Stock Ships Fast
When you click on links to various merchants on this site and make a purchase, this can result in this site earning a commission. Affiliate programs and affiliations include, but are not limited to, the eBay Partner Network.
Sipeed K3 Pico-ITX 32GB LPDDR5 RISC-V AI Board 60 TOPS - USA Stock Ships Fast
Price: US $750.00
[US Seller - Ships Fast] Sipeed K3 Pico-ITX 32GB LPDDR5 AI Inference Board (60 TOPS / RISC-V)

⚡ IN HAND & READY TO SHIP FROM THE USA ⚡

Skip the 2-month wait from China, custom delays, and import taxes. Sold out on the official site!

Condition & Package Includes
  • Condition: Excellent / Like New

  • Model: Sipeed Key Stone K3 Pico-ITX Development Board

  • RAM Configuration: 32GB LPDDR5 (Maximum Capacity)

  • In the Box: Sipeed K3 Pico-ITX 32GB Board

  • Shipping: Ships same or next business day via USPS.

Key Features & Highlights
  • Compute Power Awakens: Powered by the SpacemiT Key Stone K3—the World's First RVA23 AI CPU featuring 8+8 Cores Fusion Compute and 60 TOPS total AI capacity.

  • LLM Native (32GB RAM): Run 30B parameter models locally and privately. Inference speeds over 10 Tokens/s—no cloud connection required.

  • Ultra-Fast Memory: 32GB LPDDR5 (6400Mbps) with 51GB/s bandwidth, paired with full-speed PCIe Gen3 and USB 3.0.

  • Pico-ITX Form Factor (100mm x 86mm): Unmatched desktop AI performance in a compact footprint.

  • Jetson Orin Nano Carrier Compatible: Modular design allows you to reuse existing cases and ecosystems, migrating seamlessly from ARM to RISC-V hardware.

  • 10GbE Networking: Onboard 10 Gigabit Ethernet + Gigabit LAN to eliminate I/O bottlenecks.

  • Modern I/O: Dual Type-C ports with USB-PD + Alt-DP support for single-cable power and display debugging.

Architecture Breakdown (SpacemiT K3)
  • X100 Big Cores: 4-issue out-of-order execution @ 2.4GHz for heavy single-threaded performance.

  • A100 AI Cores: Full support for up to 1024-bit RVV1.0 parallel computing, purpose-built for matrix operations.

  • Zero-Cost Deployment: Follows standard CPU programming paradigms; runs AI algorithms natively without complex proprietary NPU toolchains.

Shipping & Terms
  • Ships from Connecticut, United States.

  • All photos show the actual item you will receive. Please feel free to ask any questions prior to purchasing!



Buy Now