Skip to main content

AOOSTAR EG02 Thunderbolt 5 / OCuLink eGPU Dock

The AOOSTAR EG02 is a dual-mode eGPU dock that connects a desktop graphics card over either Thunderbolt 5 or a native OCuLink (PCIe 4.0 x4) link, using an Intel JHL9480 Thunderbolt 5 controller. It is an open-frame dock that accepts ATX/SFX power supplies (not included) and provides 140W host charging over the Thunderbolt 5 connection. It is distinct from the AOOSTAR AG02, which is USB4-only12.

Specifications

SpecValue
ModelEG02
Price~$109–149 USD (dock only, no GPU/PSU)
Host connectionThunderbolt 5 or OCuLink (PCIe 4.0 x4)
ControllerIntel JHL9480 (Thunderbolt 5)
GPU slotPCIe x16
Host power delivery140W (via TB5)
Downstream TB51x
OCuLinkBuilt-in redrivers
Power supplyBYO ATX/SFX (not included)
EnclosureOpen-frame

Framework Compatibility

ModelStatusNotes
Framework Laptop 16 (AMD Ryzen 7040 / AI 300)WorksVia TB5/USB4, or OCuLink with an Expansion Bay adapter
Framework Laptop 13 (Intel Core Ultra)WorksTB4/USB4
Framework Laptop 13 (AMD Ryzen AI 300)WorksUSB4 port
Framework Laptop 13 (Intel 11th/12th/13th Gen)WorksTB3/TB4
Framework Laptop 12IncompatibleNo USB4/Thunderbolt
Two ways to connect: Thunderbolt 5 or OCuLink

The EG02 exposes both a Thunderbolt 5 host port and an OCuLink input. Over Thunderbolt/USB4 it behaves like any eGPU enclosure (hot-pluggable, PCIe-tunneled). Over OCuLink it gets native PCIe 4.0 x4 bandwidth with no protocol overhead — but, as with all OCuLink setups, no hot-plugging; you must reboot to connect or disconnect1.

Distinct from the AG02

The AG02 is a USB4 (ASMedia) enclosure. The EG02 upgrades to an Intel JHL9480 Thunderbolt 5 controller and adds the OCuLink option, giving higher peak bandwidth and a fallback to native PCIe12.

Where to Buy

StoreLink
AOOSTAREG02
Amazon.comSearch
Amazon.plSearch
Amazon.deSearch
Amazon.frSearch
Amazon.esSearch
Amazon.itSearch
Amazon.co.ukSearch

Footnotes

  1. AOOSTAR EG02 TB5 + OCuLink eGPU Dock — AOOSTAR 2 3

  2. 2025 AOOSTAR EG02 eGPU Dock with Thunderbolt 5 / OCuLink — Discussion — eGPU.io 2