AI Transformation in Energy Systems
From the grid to AI data centers, we engineer advanced power systems that bring together proven expertise in power conversion, grid protection, and intelligent energy technologies — enabling the reliable, efficient, and scalable infrastructure required for the next era of AI.
AI does not just need more power.It needs intelligent power.
As AI workloads grow, power infrastructure is becoming more dynamic, more complex, and more demanding.
Voltage fluctuations, power flows, grid protection, and energy efficiency can no longer be addressed as separate challenges.
CROCUS sees them as one intelligent power system.
Acelo SST
Where a conventional transformer simply converts voltage, a solid-state transformer controls voltage, current and power flow in real time. Closer to a router than to a passive component.
Connects on the secondary side without removing the existing main transformer. Voltage optimisation cuts industrial site consumption by an average of 5%.
A single conversion stage for AI data centers. Cuts infrastructure space by 40–50% and TCO by about 30%.
Applied at the normally-open point between distribution feeders to redistribute power, improving grid efficiency and utilisation.
Engineering & R&D
Research floor, laboratories and production line on one site at Uiwang — so a design decision and the machine that has to build it are never far apart.
The core of our engineering. Power electronics and Edge AI are designed, prototyped, tested and built here — research floor, laboratories and production line on one site.
67 Malgeunnae-gil, Uiwang-si, Gyeonggi-do — Bldg. C, Units 102 & 603
Contact
Technical questions go straight to the engineers who work on the product. Tell us the application, the voltage level and the timeline, and we will answer with specifics.
/* 오버레이에 -z-10 을 주면 안 된다. 이 컨테이너가 -z-20 으로 쌓임 맥락을 만들고, 그 안에서 음수 z 는 z-index:auto 인 형제(= next/image 가 그리는 img)보다 뒤로 간다. 실제로 오버레이를 빨강으로 칠해도 화면에 보이지 않았다 — 값을 아무리 조절해도 아무 효과가 없었다. 위치만으로 충분하다: DOM 순서상 이미지 뒤, EnergyFlow 앞이다. */