Electricity has a peculiar talent for disappearing into the background. It is noticed when it fails, cursed when the bill arrives and otherwise left to hum behind walls. Daniel Jesuph Lim built his company by refusing to let it be so shy. His question was less romantic than it was useful: if industrial facilities already measure enormous amounts of power, why does so little of that information alter the next decision?
Lim is the CEO and co-founder of Crocus, an energy technology company with roots in Silicon Valley and operations in South Korea. The company has moved through several chapters since its 2016 founding - factory voltage optimization, energy-management software, electric-vehicle charging and, more recently, advanced transformer systems. The connecting idea is remarkably consistent. Data should not end its life as a handsome graph. It should become an instruction.
A statistician walks into a substation
Lim studied statistics at Carnegie Mellon University from 2005 to 2009. He later completed a master's degree in engineering and industrial management at Brown University in 2014. The pairing matters. Statistics teaches respect for uncertainty; engineering insists that the answer eventually survive contact with a machine.
Before Crocus, he worked at Samsung SDI's American operation and at Utilidata, a U.S. grid-software company. Those roles put him close to batteries, utilities and the awkward seam where software meets electrical equipment. The grid was becoming measurable, but measurement alone did not make it responsive. There were meters on one side, voltage-control hardware on the other and a useful business waiting in between.
Crocus began in Silicon Valley in June 2016. Its Korean corporate presence followed, giving the company a working bridge between Redwood City and Seoul. The geography was practical. California offered a dense network of energy and software talent; South Korea offered an industrial economy where electricity is a visible operating cost and factories make unusually good laboratories for efficiency.
The expensive gap in the middle
When Lim described the early business in a 2019 interview, he did not begin with artificial intelligence. He began with a steelmaker's electricity bill. A large factory could spend hundreds of billions of won on power. A modest percentage improvement, achieved without changing the production process, became a very serious number. This was technical storytelling with a calculator in its pocket.
The product was then called Power Saver. Crocus gathered industrial power data, used machine learning to identify an optimized voltage and coordinated the equipment that could make the change. The pitch was deliberately undramatic: keep making the same glass or steel, but waste less power while doing it. Crocus reported that deployments could cut electricity costs by 5 to 7 percent under the conditions it described, and Lim cited Corning's Korean operation as an early customer.
Power efficiency is the beginning of automation and the smart factory.Daniel Lim, translated from a 2019 interview
There is a founder's lesson tucked inside the transformer cabinet. Lim did not ask customers to purchase “AI” as an abstract virtue. He attached software to an existing bill, an existing control mechanism and an outcome the buyer could check. The algorithm earned its place by producing a physical difference.
The early factory proposition
The dashboard grows hands
Lim's larger ambition was visible early. While visiting big factories, he encountered facilities still managed with aging software and improvised routines. He told ChosunBiz that Crocus wanted to become a broadly used enterprise system, comparable in function to the way Salesforce organized customer information, and eventually a core operating system for industrial automation.
The comparison was revealing. A smart factory is usually marketed as a collection of impressive nouns: sensors, twins, models, robots. Lim's formulation focused on the verb between them. A useful operating system coordinates. It turns scattered information into a shared rhythm. In energy, that means knowing what a site is consuming, what the grid can supply, what equipment can safely change and when the economics justify doing so.
At COMEUP 2020, Lim appeared on a panel about AI-based next-generation infrastructure alongside founders working on autonomous drones and smart ports. The grouping was apt. These were businesses in which machine data had to cross the border into the physical world. The panel discussed the importance of industry-specific processes, a sober point that tends to get lost whenever AI is treated as a universal solvent.
By the 2022 Climate Energy Summit Seoul, Crocus's public description had widened to grid control and EV charging in both Korea and North America. Lim's speaker biography offered a phrase that sounds almost musical for a subject ruled by electrical standards: “energy micro-balancing.” It captures the move from monitoring a large system after the fact to balancing smaller pieces of it continuously.
Follow the wire
Crocus's product range expanded along the electrical path. Energy-management software could help a factory understand and optimize its load. Charging systems introduced vehicles, access control and power allocation. Grid products connected renewable generation, storage and local demand. Transformer work moved the company deeper into the machinery that changes and routes electricity.
This expansion can look broad on a product page, but the adjacencies are physical. A charger is a demanding electrical load. A depot full of chargers is a grid-management problem. Solar panels and batteries make that problem more dynamic. A transformer sits where the entire argument passes through. Crocus followed the wire from one point of control to the next.
In July 2026, the company said it had been selected to lead a South Korean government-supported project for an AI-controlled solid-state transformer system aimed at data centers. The proposed system would use silicon-carbide power electronics and provide a medium-voltage input with an 800-volt direct-current output. Lim linked the work to rising data-center power requirements and said Crocus had treated power conversion as core next-generation infrastructure since 2020.
A solid-state transformer is a neat culmination of his thesis. The old transformer is stout, passive and mostly indifferent. Its software-controlled successor promises to be more observable and adaptable. It is the dashboard finally growing hands.
The less visible work is compliance. Charging infrastructure must speak agreed languages, behave safely and survive the scrutiny of utilities and laboratories. Open Charge Alliance documents from 2024 list Lim as Crocus's CEO on certifications for charging-system software. A certificate is not the sort of artifact usually pinned above a founder's desk, but it marks an important transition. A prototype can be persuasive in a room. A standardized product has permission to enter a network where somebody else's car, building and power supply are already waiting.
That distinction helps explain the patience required in energy technology. Consumer software can ship on Tuesday and apologize on Wednesday. Electrical infrastructure prefers the apology before anything is energized. Development moves through pilots, certifications, equipment partnerships and long customer decisions. The glamour is rationed; the dependencies are not. Lim's decade at Crocus is therefore also a story about staying with a problem long enough for the institutional machinery to move.
A practical form of ambition
People who worked with Lim before Crocus described him on LinkedIn as proactive, attentive and good at translating concerns into workable relationships. One recommendation recalled his role connecting a patent-sales team with potential Korean purchasers; another praised his ability to understand business needs and mitigate issues before they hardened into obstacles. Those are operator's virtues. Infrastructure companies are assembled from standards, equipment vendors, utilities, industrial buyers and government programs. Charm helps. Follow-through keeps the electrons moving.
Lim's public style is more engineer than oracle. He talks about voltage, cost, control and deployment. Even his broader aspiration - an operating system for industrial automation - begins at the factory floor rather than in a cloud of abstraction. The approach has allowed Crocus to change its hardware and markets while keeping a stable idea of value.
The cross-border nature of the company adds another kind of translation. Redwood City and Seoul do not share a power market, industrial culture or startup vocabulary. Yet Crocus has used both as working terrain. Lim can discuss an American grid-software lineage, Korean factory economics and international charging standards without treating any one of them as scenery. For an infrastructure founder, this is a useful temperament: respect the local constraint, then look for the underlying system that travels.
Ten years into the company, that idea can be written in four short commands: sense, predict, control, verify. It applies to a motor, a charger, a transformer or a grid. It also explains Lim's career. Statistics taught him to see the signal. Engineering taught him to respect the system. The business is what happened when he insisted the two should speak.