AMD introduced the Threadripper Halo Station at IFA 2026, positioning it as a high-performance AI workstation. The system integrates a Threadripper Pro 9995WX processor and two liquid-cooled Instinct MI350P accelerators.
The Threadripper Halo Station is equipped with a 96-core, 192-thread Zen 5 Threadripper Pro 9995WX CPU, capable of boosting up to 5.4 GHz. It also includes 2TB of DDR5 memory and 288GB of HBM3E memory from the dual MI350P accelerators, with potential for expansion to four accelerators and 576GB HBM3E. AMD states this configuration allows the workstation to run trillion-parameter models.
The Threadripper Pro 9995WX CPU features 384 MB of L3 cache and a 350W TDP. Each MI350P accelerator contains 128 CDNA 4 compute units built on TSMC N3 and provides 144GB of HBM3E memory, with each accelerator rated for up to 600W. The system shown at IFA supports two liquid-cooled accelerators, though AMD mentions a path to four.
The estimated street price for the core components (CPU, memory, dual GPUs) is over $100,000, potentially exceeding $150,000 with additional storage, power, and cooling. AMD plans for OEM partners to build and ship these systems, though no partners have been announced yet. The design reconfigures server tray components into a tower form factor.
✨ This summary was generated by AI from the outlets' reporting listed below. It is not independently verified and may contain errors — check the original sources. How BrevFeed works →
One email each morning: the day's tech stories, clustered across outlets and summarized. No account needed.
One email a day. Unsubscribe in one click, any time.
Spend a few minutes, get the whole day. Every topic's top stories in one hands-free rundown — listen, watch, or read the transcript.
▶ Play today's briefNew every morning, and the back catalogue is archived by date.
AMD announced the Threadripper Halo Station, a new AI workstation featuring a 96-core Threadripper Pro 9995WX CPU and dual liquid-cooled Instinct MI350P accelerators, at IFA 2026. This workstation is designed for running large AI models, with AMD claiming it can handle trillion-parameter models, and is expected to be available through OEM partners.