Singapore AI Demand: EDB Survey Shows Where Hiring Grows

The Economic Development Board published its quarterly manufacturing survey on 2 November, and the headline jumped. A net weighted balance of 24 per cent of manufacturers expect business conditions to improve between October 2026 and March 2027, compared with the third quarter. In the previous round the same measure sat at 12 per cent.

The survey ran across September and October, covered 397 manufacturing establishments, and drew an 83 per cent response. EDB put the reason in the opening line of its release: sustained strength in global AI-related demand.

Singapore does not host the models. It supplies the hardware they run on, and this quarter those order books are why a sentiment index sits at its highest in four rounds.

Precision engineering came in at +48, with its machinery and systems segment at +58. Those firms build the equipment that fabricates and tests chips. Electronics scored +42, and inside that cluster semiconductors reached +47 while computer peripherals and data storage hit +49.

If your mental picture of "AI jobs in Singapore" is prompt engineers and chatbot startups, the survey points somewhere more physical: silicon, bonding wires, precision components, storage arrays.

The tail end of the same table matters too. Chemicals landed at -13, with the petroleum segment at -28 on uncertainty over crude feedstock supply from the Middle East. Transport engineering managed +1, held back by material, fuel and freight costs plus overseas competition.

Expectations are cheap to state. The output forecast in the full survey release is the firmer number. A net weighted balance of 34 per cent of firms expect production to rise in the fourth quarter of 2026, up from 26 per cent for the third. Semiconductors lead at +64, precision engineering at +54, data storage at +49.

One reading note, because this trips people up. The net weighted balance is the weighted share of firms saying conditions will improve minus the share saying they will worsen. A +34 is not a 34 per cent rise in output. It means optimists outnumber pessimists by that margin, weighted by size. In plain terms, roughly two thirds of manufacturing output in Singapore expects to be busier in the last three months of the year than in the third.

The hiring detail is narrower. A weighted 72 per cent of manufacturers expect employment in the fourth quarter to stay broadly unchanged from the third. Electronics is the clear outlier at +50 for headcount, semiconductors +56, computer peripherals and data storage +72, precision engineering +20. Biomedical manufacturing at -3 and transport engineering at -8 expect slightly smaller teams.

Put the two sets of numbers side by side and the story sharpens. AI demand is lifting output faster than it lifts headcount. A plant that adds capacity automates the new line, and the people it hires are equipment, process and yield specialists, not extra pairs of hands on the floor.

The roles that follow from this survey are unglamorous and steady: equipment engineer, process technician, quality and reliability engineer, test and automation engineer, facilities and utilities engineer. Most are diploma or ITE entry points, not degree-only. The ITE Work-Study Diploma in Microelectronics and the various SkillsFuture-funded WSQ courses in precision engineering exist precisely because these plants struggle to fill them.

Two practical moves. First, search MyCareersFuture for "semiconductor", "precision engineering" and "data storage" rather than for "AI", because the postings are written for the trade and not for the hype. Second, when you apply, show the adjacent skill the plants now assume: reading data, running a simple Python or Excel analysis on yield logs, and using an AI assistant to draft reports, specs and shift handovers without pasting confidential process data into it.

If you run a small supplier into these clusters, the survey is a demand signal with a cost problem attached. A weighted 23 per cent of firms expect difficulty securing export orders in the fourth quarter, and they name overseas price competition first, then geopolitical developments, tariffs and inflation. Your customer is busy; your margin is the thing under pressure. The Champions of AI programme and the SME support that came with it is where the government is putting its money for firms in that position.

Sentiment is not revenue. Three things could turn this table over by the next round: a wobble in AI capital spending that reaches equipment orders, a tariff or export-control change touching semiconductor trade, or crude prices moving the chemicals cluster further into the red.

The wider context is Singapore's AI adoption and policy track, which has been running hotter than the global average on diffusion measures, and the entry-level hiring squeeze that sits alongside it. Both are consistent with what EDB found: more machines doing more work, and fewer new desks than the output numbers would suggest.