ASUS announced the XA P2N-E2 server at OCP Global Summit 2026. Its press page carries an October 13 date above an October 12 San Jose dateline. This brief preserves both dates rather than treating the page as proof of a separate October 13 launch event.
The company describes the XA P2N-E2 as a 2U server based on NVIDIA MGX architecture and powered by NVIDIA Vera CPUs. It targets the execution layer of AI applications, including code execution, tool use, sandboxing, data processing and orchestration. Those are intended workload uses, not evidence that a particular model or customer's agent runs reliably on it.
ASUS says the server supports up to two dual-slot NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs. Supported configurations are different from measured throughput or a retail package that already includes every supported component. The release supplies no independent latency comparison or price for this server.
The announcement also describes larger systems shown at the event. The XA VR721-E3 AI POD integrates 36 Vera CPUs and 72 Rubin GPUs through the Vera Rubin NVL72 platform with a fully liquid-cooled architecture. A separate XR R1I12B-M1BHFS combines HGX Rubin NVL8 with liquid cooling for its GPU baseboard and air cooling for Intel Xeon 6 processors. These are distinct configurations, not interchangeable specifications for the 2U server.
Other showcase systems use Intel and AMD platforms. ASUS lists 1U and 2U dual-socket Xeon 6 servers and an EPYC 9006-series platform with PCIe 6.0 and support for up to 16 E3.S drives. The range illustrates the vendor's announced portfolio rather than proving one architecture is best for every cloud or AI workload.
Limits: the source is an ASUS product announcement, not a hands-on test. Performance-per-watt, deployment speed and responsiveness descriptions are manufacturer claims. This brief does not infer shipping inventory, a public price, a benchmark lead or a consumer PC launch from systems displayed at a trade event.