Taiwan-based Altos Computing is expanding beyond its traditional hardware business as it pushes deeper into enterprise AI infrastructure, targeting growth opportunities in the Philippines and the broader Asia-Pacific region.
The company, which operates as the enterprise arm of Acer, said it has been building its AI infrastructure business since early 2025 as demand for GPU servers and enterprise AI systems continues to rise.
Founded in 1977 and acquired by Acer in 1990, Altos became an independent entity within the Acer Group in 2017.
The company said its headquarters workforce has grown from 30 to more than 80 employees following its formal entry into the AI infrastructure market this year.
Altos said it is also expanding operations outside Taiwan, including in India where more than 40 employees are now supporting its regional growth efforts.
The company added that it is working with partners across Malaysia, the Philippines, Singapore, Thailand, Vietnam, and South Korea.
The company cited projections showing the Philippine AI hardware market growing from $37 million in 2024 to $105 million by 2028, driven largely by demand for GPU servers and AI-related infrastructure projects.
Altos said more than 80% of projected GPU server sales in 2024 are expected to support roadside AI data center projects.
It added that enterprise customers, government agencies, and educational institutions are among the sectors seeking AI infrastructure deployments.
Among the products being positioned for the market is the Altos BrainSphere R880 F7 server, equipped with eight NVIDIA B200 GPUs for AI training and large-scale AI workloads.
The company also offers PCIe-based systems such as the Altos BrainSphere R785 F6 and Altos BrainSphere R680 F7, which support configurations ranging from two to eight GPUs.
Aside from servers, Altos is also offering AI-focused workstations including the BrainSphere P150 F10 and P130 F10, which the company said are aimed at organizations with limited space and technical resources, including schools.
Another product highlighted by the company is the compact BrainSphere GB10 F1 system, which supports up to 128GB of VRAM and can be paired with external NAS storage systems for expanded capacity.
Altos also argued that on-premise AI infrastructure could offer lower long-term costs compared to cloud-based or leased GPU services, citing a three-year total cost of ownership analysis favoring local deployments.
