The Veriton GN100 was revealed at IFA 2025 in Berlin.
In today’s time, technology has advanced significantly. In this era of AI, you used to need bulky computers for AI processing, but since Nvidia launched the GB10 Superchip, there has been a major change in the AI workstation industry. This Superchip has been developed by Nvidia in collaboration with Mediatek, which means Mediatek developed the CPU cores of this chip. This chip is based on the Grace Blackwell architecture. The special feature of this chip is that it is an APU, and you get a powerful iGPU within it. You get 5th generation Tensor Cores developed on Blackwell architecture, along with DLSS-4 and Ray Tracing. This powerful APU chip from Nvidia will be a game changer. With this powerful superchip, Nvidia has launched its powerful mini AI workstation, DGX Spark. After the launch of this GB10 Superchip, other major companies like MSI, Asus, Dell, and Lenovo have also launched their powerful Mini AI Workstations in the market, and now Acer has also launched its AI Workstation, the ‘Veriton GN100.’ Let’s see what you get in it.
Table of Contents
Design:
If we talk about the design, all the mini AI workstations coming with the Nvidia GB10 Superchip have almost the same design, and in terms of specs, they are also exactly similar, as they all feature the Nvidia GB10 Superchip APU. However, the total weight of this Acer Veriton GN100 mini AI workstation will be approximately 1.5 kg, which is 300 grams more than Nvidia’s DGX Spark.
CPU/GPU:
If we talk about the CPU, the Nvidia GB 10 is an APU that features 20 special cores designed by Mediatek, which are powerful and power-efficient (10x Cortex-X925 + 10x Cortex-A725). This means 2 clusters with 10 cores each, along with a 16 MB L3 cache per cluster. The special thing is that there is a private L2 cache reserved for each core.
Image Source: Nvidia
Nvidia’s GB10 has its S-dielet and G-dielet fabricated using advanced 3nm technology. The die of the Nvidia GB10 is internally divided into two parts: the S-dielet and the G-dielet. In the S-dielet, you’ll find MediaTek CPU cores, memory subsystems, and more, while the G-dielet contains GPU cores, including fifth-generation Tensor Cores with FP4 (4-bit) precision. The Nvidia GB10 Grace Blackwell superchip is capable of delivering up to 1 petaFLOP of AI performance (1,000 TOPS) using FP4/sparsity.
For detailed information about the NVIDIA GB10, you can check out the article on Nvidia DGX Spark mini AI workstations.
RAM/Storage:
The Veriton GN100 features superfast massive 128GB LPDDR5X memory, just like other GB10-based AI workstations, which will be shared with the CPU and iGPU. This massive memory enables local handling of large generative models—up to 200 billion parameters. The Veriton GN100 offers you up to 4TB Gen 4 NVMe.
OS/Software:
Veriton GN100 Mini AI Workstations provide server-grade performance despite their small size, just like Nvidia DGX Spark. It comes with Nvidia’s AI software stack and DGX Base OS, which supports commonly used developer tools such as PyTorch, Jupyter, and Ollama.
Connectivity:
For connectivity, you get a total of four USB 3.2 Type-C ports, HDMI 2.1b, and an Ethernet RJ-45 (10 GbE) port, while for wireless connectivity, it offers Wi-Fi 7 and Bluetooth 5.1.
Nvidia GB10 Spec:
| Specifications | Details |
|---|---|
| CPU | 20-core Arm CPU (10x Cortex-X925 + 10x Cortex-A725) |
| GPU | Blackwell-based 5th-generation Tensor Cores |
| AI Performance | Up to 1 PFLOPS of FP4 AI performance. |
| Memory | 128 GB unified LPDDR5x memory. |
| Storage | Up to 4 TB NVMe M.2 SSD, self-encrypting. |
| IO-Ports | 4 × USB 3.2 Type-C, HDMI 2.1b, Ethernet RJ-45, NVIDIA ConnectX-7 SmartNIC. |
| Connectivity | Wi-Fi 7, Bluetooth 5.1. |
| Expandability | You can link two GN100 units (via ConnectX-7 NIC) to handle larger models up to ~405 billion parameters. |
| Security & Other Features | Self-encrypting storage, Kensington lock, support for local model execution. |
| Physical Size & Weight | Approx. 150 × 150 × 50.5 mm; under 1.5 kg. |
| OS/Software | Ships with NVIDIA AI software stack / DGX Base OS. Supports developer tools like PyTorch, Jupyter, Ollama, etc. |
| Price | Starts at US $3,999 (also €3,999 in EMEA and AUD 6,499 in Australia). |
Special features:
You can double the memory and compute capacity by connecting two Veriton GN100 AI Workstation units together (through the ConnectX-7 Smart NIC), which allows you to handle workloads on models with up to 405 billion parameters. Security features include a Kensington lock and the ability to execute AI models locally to protect sensitive data.
Price:
If we talk about the price, the starting price of the Acer Veriton GN100 AI mini workstation will be up to 3999 USD.
