HP Z2 Mini G1a Nero corvino 2025

The HP Z2 Mini G1a packs a 16-core Ryzen CPU and 128GB of RAM into a tiny workstation. It's a memory monster built for AI inference and heavy multitasking, but the integrated GPU holds it back for rendering and LLM training. Prices range from $3,734 to over $900k depending on the vendor, so shop carefully. If you need massive unified memory in a compact box, this is one of the best options out there.

The Ryzen AI Max+ PRO 395 is a beast for multi-threaded work. With 16 cores running at 3.0GHz, it chews through code compilation and data processing tasks that would make a standard desktop sweat. In our database, this CPU sits in the 91st percentile, meaning it outperforms the vast majority of workstations we've tested. The real magic is the unified memory architecture. That 128GB of 8533 MHz LPDDR5x isn't just system RAM, the integrated GPU can access it directly. This lets you work with massive datasets or AI models that simply won't fit on a discrete GPU with 8GB or 12GB of VRAM.

The Radeon 8060S graphics are solid for an integrated solution, landing in the 64th percentile overall. It handles 3D design and rendering tasks well enough, but it's not going to replace a dedicated RTX card for heavy GPU compute. The 1TB NVMe SSD is middle of the pack at the 73rd percentile, fast enough for most workflows but not breaking any records. Where this machine stumbles is pure AI LLM performance, which we scored at 47.7 out of 100. The NPU helps with some AI tasks, but for large-scale model training or inference, the lack of a high-end discrete GPU holds it back.

The Z2 Mini G1a occupies a weird niche, and its competitors reflect that. The Lenovo Legion 34IAS10 and ASUS Republic of Gamers GM700TZ are more traditional gaming and creator desktops. They'll smoke the G1a in GPU-heavy tasks thanks to their discrete graphics cards, but they can't touch the 128GB of unified memory. If you're doing 3D rendering or video editing, those towers make more sense. The MSI EdgeXpert and Dell Tower Plus are closer to traditional workstations, offering better expandability and discrete GPU options, but they're physically much larger.

The CLX SET is a wildcard, likely a custom-built machine that can be configured to match or exceed the G1a's specs in a larger chassis. The trade-off is always the same: the HP gives you incredible memory capacity in a tiny box, while the competition gives you more GPU power and upgradeability in a bigger package. If desk space is at a premium and your workloads are memory-bound, the G1a wins. If you need raw GPU compute and don't mind a tower, look elsewhere.

Pricing on the G1a is all over the map depending on where you look. We've seen it listed anywhere from $3,734 to an eye-watering $922,201 across different vendors. That spread is wild, and it means you absolutely need to shop around. The lower end of that range puts it in a reasonable spot for a workstation with 128GB of RAM and a top-tier mobile CPU. At the high end, you're paying more than a car for something that fits in a backpack.

For what it offers, the value proposition is strongest if you specifically need that massive unified memory pool. There aren't many mini PCs that can allocate 96GB of RAM to the GPU for AI workloads. If that's your use case, the G1a is almost in a class of its own. But if you don't need that much memory bandwidth, a traditional tower with a discrete GPU will give you better raw compute per dollar.

HP crammed a full-blown workstation into a box that's basically the size of a thick book, and the result is the Z2 Mini G1a. This isn't your average office PC. It's built for developers, data scientists, and engineers who need serious compute but don't have room for a tower. The headliner is that AMD Ryzen AI Max+ PRO 395 chip with 16 cores and a dedicated NPU, paired with a frankly absurd 128GB of RAM. You can run large language models locally, compile massive codebases, or render 3D scenes without breaking a sweat.

We see a lot of mini PCs come through our database, and most of them compromise somewhere. The G1a makes a different trade. It goes all-in on CPU and memory, landing in the 91st and absolute top percentile respectively. The integrated Radeon 8060S graphics are no slouch either, but they're not the star of the show. This machine is for people whose workloads are more about cores and memory bandwidth than pushing polygons. Think AI inference, virtual machines, and heavy multitasking.

Connectivity is another strong point. You get three USB-C ports, five USB-A ports, dual Mini DisplayPort 2.1 outputs, Thunderbolt, and Wi-Fi 7. That port selection puts it in the 94th percentile, which is rare for a box this small. You can drive multiple high-res monitors and still have room for your peripherals. Just don't expect to upgrade the RAM later, it's soldered LPDDR5x, so that 128GB is what you're stuck with.

If your work revolves around GPU rendering, video editing, or 3D animation, the Z2 Mini G1a is probably not for you. The integrated Radeon 8060S just can't keep up with discrete workstation GPUs from NVIDIA or even AMD's own dedicated cards. You'd be better served by a tower like the Dell Tower Plus or a custom build with an RTX 4070 or better. The GPU sits in the 64th percentile, which is fine for general use but underwhelming for graphics professionals.

Also, if you like to upgrade your machines over time, look elsewhere. The soldered RAM and limited internal expansion mean this is a buy-it-and-use-it-as-is proposition. Anyone who wants to swap out components or add more memory in a couple years should consider a traditional tower workstation instead.

For developers and data scientists working with large language models locally, the Z2 Mini G1a is a compelling option. That 128GB of unified memory means you can run 70B parameter models without needing multiple GPUs, and the compact size means it won't dominate your desk. The CPU performance is top-notch, and the port selection means you can connect pretty much anything. Just know that the integrated graphics will limit you if your work involves heavy GPU rendering or training.

If you're a CAD designer or video editor, this probably isn't your machine. The GPU just isn't strong enough to compete with workstations packing discrete RTX cards. But for a specific slice of the market, people who need massive memory in a small footprint, the G1a is basically unmatched. It's a specialized tool that excels at its intended job, even if it stumbles at tasks outside that narrow focus.