Singapore’s AI Boom Has a Training Gap, and Employers Own It

Eighty-six percent of workers in Singapore now use AI at work. Yet among workers who skipped AI training in the past 12 months, around 35% said their employer did not nominate them.

Both numbers come from figures the Infocomm Media Development Authority (IMDA) released Oct. 5. For chief data officers, the space between them matters more than the headline.

The headline is strong. According to the Singapore Digital Economy Report 2026, Singapore’s digital economy reached SGD144.1 billion in 2025, or 19.3% of GDP, up from 18.8% in 2024. Tech employment grew 3.8% to 222,200 jobs. AI adoption rose from 14.5% to 23.4% among small and midsize enterprises (SMEs) and from 62.5% to 70.4% among large local enterprises.

Beyond the totals, the report answers four questions for data leaders: Who competes for your talent? What do you control? How do you prove value? Who trains your people?

Your talent rivals are not only tech firms

Most of the digital economy sits outside the tech industry. Sectors other than information and communications (I&C) contributed more than two-thirds of the total, or SGD97.0 billion, up 3.6%. The I&C sector reached SGD47.1 billion, up 6.1%.

Tech jobs follow the same pattern. Tech employment grew 5.3% outside I&C and 1.8% inside it. Roles in AI and data, and in cybersecurity, were among the fastest-growing. The median resident tech worker earned SGD8,000 a month, 60% above the overall resident median of SGD5,000.

So the data engineer a CDO wants to hire may weigh offers from any sector. Budget for it.

Those employers are also choosing how to bring in AI. Most that use it buy it ready-made.

You rent the model, so you control the data

Close to nine in 10 AI-adopting enterprises use commercial off-the-shelf tools, meaning ready-made products. IMDA’s examples are ChatGPT, Claude, Gemini and Microsoft Copilot. IMDA found that 28.3% customize such tools, for example by building AI agents or apps on commercial models.

When a company rents the model, what it controls is the data that goes in. That puts data classification, access rules and retention policies, the CDO’s core work, at the center of AI strategy.

Use is wide but not yet deep. Two in five AI adopters are expanding deployment or have fully deployed or integrated AI across the enterprise. Across digital tools more broadly, 96.4% of enterprises used at least one of six digital areas in 2025, including cloud, data analytics and AI. The average SME used 2.5 of the six, up from 2.3.

Renting tools is quick. Proving they pay off is harder.

Productivity is easier to claim than savings

Among AI adopters, 87.6% named productivity and process improvement as a key benefit; 44% reported cost reduction and resource optimization. SMEs that adopted AI-enabled, pre-approved digital solutions through the Productivity Solutions Grant reported average cost savings of 49.5% in 2025. IMDA notes those results are self-declared by grant applicants.

Workers report gains, too. Of those who use AI, 73% cited higher productivity and 69% better work quality. Fewer reported saving time: Around 33% said the time they spent on tasks fell, and about six in 10 of them saved up to an hour a day.

The lesson for CDOs: Measure how long a task takes before the AI tool arrives. Without that baseline, “it feels faster” never becomes “it costs less.”

Proof also needs people who know how to use the tools. On that, the report finds a gap between need and training.

The training gap starts with the employer

Around 68% of workers said they need to upskill or reskill in AI, but only 37% attended AI-related training in the past 12 months. Among those who did not, around 35% said their employer did not nominate them. A similar share cited lack of time, and about one in four were unsure which course to take.

IMDA says clearer signposting and broader support could help. The first barrier, though, sits with employers: Someone has to put a name forward.

What workers do with AI raises the stakes. Content creation is the most common use, at 77%. Around two-thirds of workers use AI for process automation, and more than half for data exploration and analytics. The report does not say how much of that analysis runs on governed data, meaning data with clear owners, agreed definitions and access rules. A company’s own usage logs may show it.

Public programs will help. IMDA’s TechSkills Accelerator will upskill 40,000 tech professionals over the next three years to become full-stack developers with agentic AI capabilities, the skills to build AI that carries out multistep tasks on its own. Under the National AI Impact Programme, 100,000 professionals will receive domain-specific AI training by 2029, with rollout underway in sectors such as accountancy and law.

“Singapore’s digital economy continues to demonstrate encouraging growth as both enterprises and workers embrace new solutions in their digital transformation journeys,” said Ng Cher Pong, IMDA’s chief executive. “As a result of the strong underlying momentum, our digital economy has now grown to 19.3% of Singapore’s overall economy. IMDA remains committed to investing in continuous learning — equipping workers with practical, domain-specific AI skills — to ensure our workforce maintains Singapore’s competitive edge.”

The government can fund the courses. Only employers can nominate staff, give them time to learn and make sure the data they query is governed.

Image credit: iStockphoto/Jacob Wackerhausen