Integrated with the Self-Developed Altos aiGeni Platform for One-Click Development and Accelerated AI Innovation
Mumbai/Bengaluru : Altos Computing Inc., a subsidiary of Acer, today announced the launch of the new Altos BrainSphere™ GB10 F1 AI workstation, fully built on the NVIDIA Grace Blackwell architecture. Bringing revolutionary performance to desktop-level AI computing, the GB10 F1 also comes with the newly self-developed Altos aiGeni software, which simplifies AI development with one-click deployment and intelligent tools. With the integration of Altos aiGeni and real-time resource management, the Altos BrainSphere™ GB10 F1 provides a brand-new AI computing experience for developers, researchers, and the education sector—enabling high-performance AI modeling at the desktop level and rapid deployment of intelligent development environments, while lowering barriers and increasing efficiency for AI applications.
“Generative AI has become the driving force of global industries, and the real challenge lies in turning creativity into reality quickly and bringing innovation to life faster. By offering an all-in-one AI solution with both hardware and software integration, Altos delivers an efficient, flexible, and secure computing foundation for industries, research, and academia. This will help customers across different fields accelerate innovation and drive AI technology towards wider adoption. We will also continue collaborating with global ecosystem partners to advance intelligent infrastructure and next-generation applications, further accelerating digital transformation in academia and industry.” Jackie Lee, CEO of Altos Computing Inc., said.
Altos BrainSphere™ GB10 F1: AI Workstation, Mini Size, Massive power
The Altos BrainSphere™ GB10 F1 adopts the NVIDIA Grace Blackwell architecture, featuring the NVIDIA GB10 Grace Blackwell Superchip that includes a Blackwell GPU and a 20-core Arm CPU. It comes with 128GB of unified system memory and a 4TB NVMe SSD, and leverages NVIDIA NVLink-C2C technology to significantly enhance data transfer between CPU and GPU. Despite its compact form factor, it delivers up to 1 PetaFLOP of computing power, meeting the demands of large language model (LLM) fine-tuning, generative AI, and AI agent development.
The GB10 F1 also comes equipped with NVIDIA ConnectX-7 Ethernet NIC, enabling high-speed dual-system interconnection and supporting models with up to 405 billion parameters. With the full NVIDIA AI software stack pre-installed, users can quickly deploy workloads locally or move work to NVIDIA DGX Cloud or any accelerated cloud or datacenter infrastructure. Supporting mainstream tools such as PyTorch, Jupyter, and Ollama, the Altos BrainSphere™ GB10 F1 redefines desktop AI computing with its “compact size, massive computing power,” accelerating AI innovation across prototyping, inference, and edge computing.
Altos aiGeni: A New One-Click AI Development Tool
Alongside the GB10 F1, Altos also introduces its self-developed Altos aiGeni AI development platform, pre-installed in the workstation, delivering a one-click, intelligent AI development experience. Designed specifically for AI users, the platform not only greatly simplifies environment setup but also provides hardware resource monitoring and automated backup functions, ensuring both data security and stable operations.
With one click, Altos aiGeni enables fast deployment, allowing users to focus on algorithms and model innovation instead of time-consuming setup of frameworks such as TensorFlow or PyTorch. The platform also features real-time hardware monitoring, providing intuitive graphical displays of CPU, GPU, and memory usage, helping developers optimize configurations instantly to avoid overloads and system crashes. In addition, its automated backup mechanism saves data periodically according to preset schedules, reducing risks of data loss and preserving research integrity. Altos aiGeni offers a user-friendly, flexible, and secure development environment for education, research, and enterprise R&D teams alike.



Leave a Reply