Doosan Enerbility's discussions with NVIDIA could mark a pivotal moment, transforming it into a key infrastructure partner for the AI era.
At the heart of this news is the concept of an 'AI factory'. Think of it not just as a data center, but as a complete, software-defined system for building and operating AI on a massive scale. These factories require enormous amounts of stable power, far beyond what traditional data centers needed. This is where NVIDIA's DSX platform comes in—it acts as the standardized operating system for these AI factories, and partners are needed to build the physical infrastructure around it.
So, why is this happening now? A few key events set the stage. First, industry giants like Siemens recently unveiled 100 MW-class power architecture designs specifically for AI factories. This validated the need for utility-scale power solutions—gas turbines, nuclear components, and hydrogen fuel cells—which are precisely Doosan Enerbility's areas of expertise. It shifted the conversation from IT hardware to industrial-grade energy systems.
Second, NVIDIA announced that its new Vera Rubin platform is in full production and that its DSX software is being adopted by major server manufacturers. This move codifies DSX as the industry standard, creating a clear technological target for companies like Doosan to integrate their products with. It's no longer a hypothetical concept; it's a tangible ecosystem they can plug into.
Third, there's a strong geographic catalyst in Korea. Major Korean corporations like SK hynix, NAVER, and Coupang are actively building out their AI factory infrastructure with NVIDIA. This creates a robust domestic market for Doosan, and NVIDIA CEO Jensen Huang's recent visit to Korea to recruit 'physical AI' partners further signals that these discussions are likely to convert into concrete projects.
In conclusion, this collaboration is more than a simple supply agreement. It represents a strategic alignment that could reposition Doosan Enerbility from a cyclical heavy industry manufacturer to a standardized, essential supplier for the global AI infrastructure buildout. The key now is to watch whether these talks translate into a formal pilot program in the second half of 2026.
- AI Factory: A term for an end-to-end system, including hardware and software, designed to produce AI models and applications at an industrial scale, much like a physical factory produces goods.
- DSX: NVIDIA's platform for building and managing AI factories. It provides the software-defined infrastructure layer that controls everything from the servers to the power and cooling systems.
- SMR (Small Modular Reactor): A type of nuclear fission reactor that is smaller than conventional reactors. They can be manufactured in a factory and transported to a site for installation, making them a flexible option for powering large facilities like AI factories.