WiCi One wireless eGPU opens for preorder at $1,999, with Wi-Fi 7 and RTX 5060 Ti
The standalone base model includes an RTX 5060 Ti, while WiCi is developing an RTX 5090 version.
Hardware manufacturer WiCi has opened preorders for the WiCi One, a standalone external graphics system designed to deliver desktop-tier graphics rendering and artificial intelligence compute power to lightweight laptops over Wi-Fi 7, without a physical cable connection. The entry-level model has an introductory preorder price of $1,999, which WiCi says will rise to $2,599 when general sales begin. Initial shipments are scheduled for the fourth quarter of this year.
The base model pairs an Nvidia GeForce RTX 5060 Ti graphics card with 16GB of video memory, an Intel Core 3 100U processor and a 1TB NVMe solid-state drive. Unlike a conventional external graphics processing unit, or eGPU, the WiCi One is a standalone computer: its chassis contains an independent central processing unit, dedicated storage, a proprietary active cooling system and an integrated Wi-Fi 7 networking module alongside the graphics processor.
WiCi says the RTX 5060 Ti configuration reaches 80 frames per second in *Cyberpunk 2077* at 4K resolution on Ultra settings with Nvidia DLSS 4 enabled. The company also claims 60 frames per second in *Black Myth: Wukong* at 4K using Cinematic presets and DLSS 4.
Traditional eGPUs are passive enclosures that connect a desktop graphics card directly to a host computer’s PCIe lanes through a high-bandwidth cable, most commonly Thunderbolt 3, Thunderbolt 4 or OCuLink. WiCi instead uses Wi-Fi 7, or 802.11be. The standard uses 320 MHz channel bandwidths, 4K-QAM modulation and Multi-Link Operation (MLO) across the 2.4 GHz, 5 GHz and 6 GHz bands, enabling theoretical peak data rates of up to 46 Gbps while significantly reducing over-the-air latency.
For comparison, Thunderbolt 4 provides up to 40 Gbps of total bandwidth, with 32 Gbps allocated to PCIe data. Emerging interfaces Thunderbolt 5 and OCuLink offer theoretical bandwidths of up to 80 Gbps and 64 Gbps, respectively.
WiCi is also marketing the device as an external accelerator for local artificial intelligence workloads. Its on-device processing power is intended to run local large language models, automated generative workflows and persistent AI assistants directly over a local network without sending proprietary data to cloud-hosted servers.
An upgraded premium version is in development with an Nvidia GeForce RTX 5090 graphics card carrying 32GB of VRAM, an Intel Core Ultra 7 225H processor and a 4TB PCIe Gen 5 NVMe SSD. WiCi has not disclosed its price and says it will announce it closer to the device’s release date.
