Asus's AI server business hit its five-fold revenue growth target a year early in 2024. By 2025, server operations are projected to contribute 15% of the company's total revenue. Asus's early move to set up a manufacturing line in California—initially to serve US and European clients—paid off as US tariff enforcement tightened in late 2024. The facility's launch helped reduce the company's exposure to rising trade tensions.
Jackie Hsu, Asus SVP and Co-Head of the Open Platform and AIoT Business Groups, noted that Asus began shifting its manufacturing footprint in 2018 in response to the US-China trade war. All motherboards bound for the US are now produced entirely outside China.
Asus actively tracks trade and policy changes in major markets like the US and China, adapting its production and sales strategies to stay ahead of global disruptions.
AI as a core growth engine
Unlike many of its peers, Asus has been investing in AI for years, across hardware, software, and systems, and has recently accelerated efforts to integrate AI throughout its operations.
"AI is no longer just for data crunching—it's now embedded in every device," Hsu said. Asus is weaving AI into its full product lineup, from infrastructure and applications to AI PCs and smartphones.
Asus provides a comprehensive AI server lineup capable of handling edge AI, inference, fine-tuning, training, generative AI, and HPC workloads—using liquid-cooling systems optimized for scale.
Hsu sees 2024 as the launchpad for mass AI adoption. Since rolling out its first Copilot+ PC in May 2024, Asus has aimed to lead the AI PC segment within three to five years.
AI PC adoption is set to drive overall market growth, led by commercial upgrades triggered by Windows 10's October 2025 sunset and pandemic-era replacement cycles, with consumer and gaming segments to follow.
Asus is deepening its commercial reach through partnerships with governments, enterprises, and resellers, while expanding its CSP customer base. Its AIoT unit is deploying solutions across various industrial sectors.
Product portfolio, integration strategy, and global build-outs
Hsu spotlighted Asus's GB300 NVL72-based AI POD platform, which has secured major orders and reflects tight integration with Nvidia.
The current AI server lineup includes models with Blackwell Ultra GPUs, along with the B300 NVL16, DGX B200 NVL8, HGX H200 series, and MGX servers supporting RTX PRO 6000 Blackwell Server Edition.
Asus established a server team of over 800 employees and, in 2024, merged its data center and enterprise divisions into the Infrastructure Solutions Group (ISG) to streamline operations and boost collaboration.
Beyond hardware, Asus delivers full-stack AI solutions—combining platforms, services, and its expertise in compute infrastructure, including LLM training.
Asus helped build Taiwania 2, which ranked 20th on the TOP500 and 10th on the Green500—and is now planning Taiwania 4. In late 2024, the company also built Taiwan's largest supercomputing facility for Ubilink.AI Co., Ltd.
Asus has secured global AI data center deals across the US, Europe, and Asia—including Vietnam and India—offering complete deployments from hardware and design to cooling, rack integration, and platform rollout.
A standout project is Vietnam's FPT AI factory, built in partnership with FPT and Nvidia. Powered by Nvidia HGX H100 GPUs and 5th Gen Intel Xeon processors, the facility supports LLM training and generative AI workloads with optimized efficiency and round-the-clock reliability.
Article translated by Levi Li and edited by Jerry Chen
