Key Takeaways

  • KONST plans to use its Series B funding to expand AI data-center capacity across Asia.
  • ADATA Technology’s memory and storage expertise supports infrastructure designed around data-intensive AI workloads.
  • KONST’s Token Factory strategy combines compute supply, GPU cloud access, and enterprise AI governance.

KONST has secured a $30 million Series B led by ADATA Technology, giving the Taipei-based company fresh capital to expand its AI data-center footprint across Asia and develop its Token Factory operating model.

The round, announced on October 9, 2026, also included M Mobility, Pegatron Venture Capital, and other corporate investors. According to PR Newswire, KONST intends to direct the funding toward additional data-center capacity and an operating model combining computing infrastructure, GPU cloud services, and enterprise AI governance.

AI infrastructure requires more than just accelerator access. Deploying GPUs at production scale depends heavily on electricity, cooling, networking, memory, storage, and software to coordinate workloads. Clusters can severely underperform if data cannot move quickly between storage, memory, and accelerators.

ADATA Technology’s background in memory and storage addresses critical components of AI data-center design. As model sizes and enterprise datasets grow, storage throughput, memory capacity, and data locality directly influence both hardware utilization and operating costs.

Purchasing compute hardware is relatively straightforward compared with keeping it productive. Operators need sustained power, effective thermal management, high-speed networking, and resilient storage. These constraints compound when deploying capacity across multiple Asian markets with diverse energy systems, regulations, and data-residency expectations.

The financing, also reported by Reuters, underscores investor interest in businesses positioned between semiconductor supply and enterprise AI deployment. KONST operates in a market alongside GPU cloud specialists and hyperscale offerings such as NVIDIA DGX Cloud, CoreWeave, and Lambda. Its differentiation relies on integrating infrastructure, cloud access, and governance rather than competing on raw GPU inventory alone.

The Token Factory framework treats computing infrastructure as a system that turns hardware into usable AI capacity. This involves supplying the physical infrastructure, making GPU resources accessible via cloud delivery, and establishing governance controls like accountability, lifecycle management, and risk mitigation for enterprise deployments.

Integrating governance into an infrastructure strategy addresses barriers beyond compute access. Organizations require controls over data use, model behavior, access permissions, and accountability when AI-generated outputs affect business decisions. Bundling these concerns into the service model allows enterprises to secure regional capacity without assembling every operational and governance component internally.

Pulse2 highlighted KONST’s plan to advance the Token Factory concept amidst broader industry challenges. With data-center electricity demand projected to more than double by 2030 due to AI workloads, operators face intense pressure to secure power and deploy capacity efficiently. Capital alone cannot resolve these underlying grid constraints or shorten equipment lead times.

The investor mix provides KONST with strategic industrial connections. ADATA Technology brings memory and storage expertise, while Pegatron Venture Capital links the company to a wider electronics manufacturing ecosystem. The practical measure of success will be whether these relationships translate into dependable capacity, predictable performance, and robust governance controls suited to production AI.

This funding allows KONST to scale its operations across Asia. If the integrated model maintains high compute utilization while actively addressing enterprise oversight, the company is positioned to capitalize on a market increasingly defined by complete AI production systems rather than isolated GPU availability.