Eaton and Singapore Polytechnic Team Up to Train Engineers for AI Data Centres' Power Crunch

Singapore's push to dominate AI infrastructure runs into a specific bottleneck: not enough people who know how to power the machines. Eaton has signed a Memorandum of Understanding with Singapore Polytechnic (SP) to teach the next generation of engineers about emerging power technologies — most notably 800-volt direct current (800VDC) systems — as AI data centres strain the limits of conventional electrical design.

The agreement, signed on 19 August at the 800V Direct Current Data Centres Symposium 2026 convened by the Institute of Technical Education (ITE), SP and ST Telemedia Global Data Centres, covers curriculum design, seminars, workshops and practical projects. Eaton brings the technical grounding in emerging power systems; SP supplies the teaching environment and coursework. The scope also extends to industrial attachments and internships, giving lecturers and students hands-on contact with live projects rather than theory alone.

The urgency comes from the electrical load AI hardware now places on data centre white space. Racks that once drew single-digit kilowatts have moved beyond 100kW, and next-generation AI systems are projected to approach or exceed a megawatt each. Conventional low-voltage distribution runs out of headroom at that scale, which is why operators and vendors are turning to 800VDC — it carries the same power at lower current, trimming copper content and cutting the number of conversion stages between grid and chip. Yet 800VDC is not yet a settled discipline with decades of teaching material behind it, so engineers who can specify, commission and maintain these systems are scarcer than the equipment itself.

Eaton frames the tie-up as support for Singapore's wider effort to build engineering talent for digital infrastructure and electrification — a lane the country is banking on as part of its national AI strategy. The composition of the symposium points to the same logic: ITE and SP sit on the training side, STT GDC on the operator side, and the technologies under discussion cross straight between them.

Why it matters for Singapore: The hardware story of AI tends to dominate — chips, models, algorithms — but the physical infrastructure underneath is where a real skills shortage is emerging. By bringing a manufacturer's engineering knowledge into coursework while the architecture is still being defined, Singapore is trying to shorten the lag between what industry deploys and what graduates have been taught, keeping its position as a regional data centre hub credible as rack densities climb toward the megawatt mark.