KONST, a startup building AI compute infrastructure, has closed US$3 million (about NT$96 million) in funding led by memory maker ADATA. The round is part of KONST's Series A, and the two companies say it backs their push into the sovereign AI market, where enterprises and governments run AI models in local data centers to keep data within national borders and meet each country's regulatory and compliance requirements.
Key points
- Vertical integration from memory to compute. ADATA contributes its DRAM and eSSD memory products, while KONST handles compute scheduling, data center construction, and operations. The two plan to offer an integrated hardware-to-operations AI compute offering for the Asia-Pacific region and beyond.
- A mining-to-AI pivot. KONST's team started in cryptocurrency mining in 2017 and redirected its GPU and server assets to AI computing during the 2022 crypto downturn, giving it an edge in GPU procurement, data center buildout, and large-scale compute scheduling. It is a path comparable to that of CoreWeave in the US.
- Faster, denser buildouts. KONST's in-house Glows.ai platform addresses data storage, migration, and environment setup, and its modular design lets an AI compute center be built in 6 to 8 months, lowering the time and capital barrier for smaller companies and application developers.
- Memory as the missing piece. As model sizes grow, memory bandwidth and capacity have become a bottleneck for training and inference. ADATA's DRAM and storage products will be integrated directly into KONST's compute architecture.
- Buildout target. KONST plans to complete 10 ESG-compliant AI compute centers worldwide by the end of 2027.
Read the full article in Chinese: link
Start with a free AI readiness assessment
In one hour, we map your data, systems, and use cases and give you an actionable AI adoption blueprint.
Book an assessment