95% of Singapore Employers Cannot Fill AI Roles — What ManpowerGroup's 2026 Talent Shortage Survey Means for Your Hiring Pipeline

ManpowerGroup published the results of their 2026 Global Talent Shortage Survey in February, and the Singapore-specific data landed with the force of a headline that most CTOs and heads of engineering already knew in their gut but had not seen quantified: 95% of Singapore employers in the technology sector report difficulty hiring the talent they need. That number is not a rounding error. It is not a soft proxy for "hiring is a bit slower than last year." It means that in a room of twenty tech employers in Singapore, nineteen of them are operating with unfilled roles that are actively constraining their product roadmap, delivery capacity, or both.

The survey — conducted across 40,000 employers in 42 countries — reveals a structural shift in what is hardest to hire. For the first time in Singapore, AI Model & Application Development (26%) and AI Literacy (25%) top the hardest-to-fill skills list, displacing IT & Data, which was the top concern in 2025 and has now dropped to seventh position at 17%. The message is unambiguous: the skills gap has migrated from general technical competence to specialised AI capabilities, and Singapore employers who have not restructured their hiring pipelines around this reality are falling further behind every quarter.

This article breaks down every actionable data point from the survey, maps it against the Singapore-specific policy landscape (Budget 2026, SkillsFuture, IMDA, EDB), and gives you a concrete playbook for what to change in your hiring pipeline this week — not next quarter.

The Numbers: What ManpowerGroup Actually Found in Singapore

The headline figure — 95% of tech employers struggling to hire — needs context. The overall Singapore talent shortage figure is actually 71% across all sectors, down from 83% in 2025 and slightly below the global average of 72%. This is the lowest level since 2021. So the broader market is easing. The tech sector is not. The gap between the 71% all-sector figure and the 95% tech-sector figure is the widest it has been since ManpowerGroup began publishing sector breakdowns, and it tells you everything you need to know about where the market pressure is concentrated.

The skill-level breakdown is equally revealing. Here is the full ranking of the hardest-to-fill skills in Singapore for 2026:

1. AI Model & Application Development — 26%: This is the engineers who build, train, and deploy machine learning models. Not data analysts. Not prompt engineers. The people who can work with PyTorch, fine-tune foundation models, and deploy inference pipelines in production.
2. AI Literacy — 25%: This is broader than engineering — it includes product managers, project leads, and domain experts who can meaningfully specify, evaluate, and integrate AI capabilities into existing workflows. The shortage here is a signal that AI is moving from an engineering concern to a company-wide competency requirement.
3. Engineering — 24%: General software engineering, cloud infrastructure, DevOps — the baseline technical workforce that every AI team needs as a foundation.
4. Operations & Logistics — 19%
5. Sales & Marketing — 19%
6. Administration — 18%
7. IT & Data — 17%: The former number one. The fact that it has dropped from first to seventh in a single year does not mean IT & Data hiring has become easy — it means that AI-specific hiring has become so much harder that it has reframed the relative difficulty of everything else.

Separately, 58% of employers identified data analytics and data science roles as the single most difficult category to fill. This is not the same as the IT & Data skills category — it is a role-level finding rather than a skill-level finding. What it confirms is that the sharpest pain sits precisely at the intersection of AI capability and data fluency: the engineers who can work with both the models and the data infrastructure that feeds them.

The Surprising Finding: Human Skills Are Not an Afterthought

Buried in the ManpowerGroup survey data is a finding that most coverage has underplayed: the most sought-after attributes among Singapore employers are not technical at all. Professional and work ethic (34%), adaptability and willingness to learn (34%), and communication, collaboration and teamwork (33%) all rank above any specific technical skill on the employer wish list.

This is not a contradiction with the AI skills shortage. It is a complement. What employers are telling ManpowerGroup — and what we hear repeatedly from the hiring managers in our network — is that they need AI engineers who can also communicate, collaborate, and adapt. The unicorn candidate is not someone who can build a RAG pipeline. It is someone who can build a RAG pipeline, explain the tradeoffs to a product manager who has never heard of vector databases, and then adapt when the business requirements change in week three.

Industry Breakdown: Where the Pain Is Worst

The talent shortage is not distributed evenly across industries. ManpowerGroup breaks it down by sector, and the results show where Singapore employers are under the most pressure:

• Utilities & Natural Resources — 79%: The highest overall talent shortage rate in Singapore. These employers are competing for AI engineers who can work on predictive maintenance, smart grid optimisation, and environmental monitoring — niche applications where the talent pool is extremely thin.
• Construction & Real Estate — 77%: BIM engineers, digital twin specialists, and AI-augmented project managers are in acute demand as Singapore's construction sector accelerates its digital transformation.
• Public Sector, Health & Social Services — 77%: Government agencies and healthcare providers are hiring AI engineers at scale for the first time, driven by national AI strategy initiatives and the expansion of HealthTech SG programmes.
• IT & Technology — 74%: The traditional tech sector, which has absorbed AI engineers for years, is now competing with every other sector for the same talent — a structural change that means AI hiring is no longer just a tech-sector problem.

The cross-industry nature of the shortage is the most consequential finding for employers. If you are a fintech company in Singapore, you are no longer competing only with other fintech firms for AI engineers. You are competing with MAS-regulated banks, healthcare startups, government agencies, utilities, and — as we have covered extensively — Google's SG$5 billion engineering centre. The competition surface has expanded from your sector to the entire economy.

What Singapore Employers Are Actually Doing About It

ManpowerGroup asked employers how they are responding to the shortage. The most common approaches, ranked by adoption:

1. Upskilling and reskilling (26%): The most popular response. Employers are investing in internal programmes to convert existing engineers into AI-capable engineers. The SkillsFuture framework is the primary vehicle, with subsidies covering 70% of qualifying course costs and rising to 90% for SMEs.
2. Flexible work arrangements (23%): Remote and hybrid policies are being used as a competitive lever, particularly against large firms that are mandating return-to-office. For SMEs without the salary budget to match Google or Grab, flexibility is the highest-ROI retention tool available.
3. Salary increases (21%): Straightforward but unsustainable as a standalone strategy. AI/ML engineers in Singapore already command SG$72,000-200,000 with a 20-30% premium over general SWE. Continuous salary escalation without productivity improvements creates a cost spiral that most SMEs cannot sustain.
4. Broadening talent sourcing (19%): Skills-first hiring that drops degree requirements, accepts non-traditional backgrounds (bootcamp graduates, career changers, ASEAN-trained engineers), and evaluates candidates on demonstrated capability rather than CV pedigree.

The Government Response: Budget 2026 and the Policy Toolkit

The Singapore government has responded to the AI skills crisis with a suite of measures in Budget 2026 and through existing agencies. Here is what employers can access right now:

• 400% tax deduction on qualifying AI spending, capped at SG$50,000 per company per year. This covers AI tools, cloud compute, and training costs. At a 17% corporate tax rate, the effective savings are approximately SG$8,500 per year — not transformative, but it offsets the cost of one engineer's SkillsFuture AI programme.
• SkillsFuture Enterprise Credit of up to SG$10,000 per employer for workforce transformation, including AI training. This is stackable with the training subsidies below.
• Training subsidies covering 70% of AI course costs, rising to 90% for SMEs. A typical 4-month AI/ML reskilling programme costs SG$12,000-18,000. After subsidies, an SME pays SG$1,200-1,800 per engineer — an extraordinary value proposition that is still underutilised.
• AI Singapore Apprenticeship Programme (AIAP): A 9-month programme that trains and deploys AI engineers to participating companies. Targets 1,000 placements per year. IMDA and EDB co-fund eligible hires through the TechSkills Accelerator (TeSA) programme.

The policy toolkit is comprehensive. The challenge is execution speed. Most employers we speak to are aware of SkillsFuture in principle but have not mapped the specific subsidies to their specific hiring gaps. If you have three unfilled Python developer roles and two of those roles could be filled by reskilling existing engineers with a 4-month AI programme subsidised at 90%, the net cost of closing those two positions is under SG$4,000 total. The third position — the one that needs a senior AI engineer with production deployment experience — is where staff augmentation or international sourcing becomes the faster path.

Skills-First Hiring: The Structural Solution

The ManpowerGroup data points to a structural solution that only 19% of employers are currently adopting: skills-first hiring. Drop the degree requirement. Evaluate candidates on demonstrated capability. Accept engineers from bootcamps, ASEAN universities, career-change programmes, and non-traditional backgrounds who can demonstrate the specific skills your team needs.

The resistance to this approach is cultural, not rational. Singapore's hiring culture has historically placed enormous weight on university pedigree — NUS, NTU, SMU, SUTD. This is a reliable heuristic when the talent pool is large and the skills required are well-defined. It is a catastrophically poor heuristic when 95% of employers cannot fill their roles and the skills required (AI model development, LLM fine-tuning, RAG pipeline engineering) were not part of any university curriculum three years ago.

Here is what skills-first hiring looks like in practice for AI/ML roles:

• Replace the degree filter with a portfolio review. Ask candidates for three things: a GitHub repository with AI/ML code they have written, a deployed model they can demonstrate, and a write-up of a technical decision they made and why. This takes 20 minutes to evaluate and tells you more about their capability than a degree from any university.
• Compensate the technical assessment. Pay SG$150-200 for a completed take-home exercise. This filters for serious candidates and signals respect for their time — a differentiation lever that large firms almost never use because their HR departments consider it "unprecedented."
• Evaluate for the human skills that ManpowerGroup identified as most sought-after. Add a 20-minute cross-functional scenario exercise: give the candidate an AI integration problem and a non-technical stakeholder (played by a PM or designer) who they need to align with. Adaptability, communication, and collaboration are visible in this exercise within the first 5 minutes.

What to Do This Week: A Concrete Action Plan

The ManpowerGroup data is a diagnosis. Here is the treatment plan — five actions you can take this week to start closing the gap:

1. Audit your open AI/ML roles for degree requirements. If any of them specify "Bachelor's degree in Computer Science or equivalent," replace it with "Demonstrated experience building, training, or deploying machine learning models. Portfolio required." This single change will increase your applicant pool by 30-40% based on the data from our network.
2. Apply for SkillsFuture subsidies for your existing engineers. Identify 2-3 engineers on your team with strong Python foundations who could complete a 4-month AI reskilling programme. The application process takes 2-3 hours. The subsidy covers 70-90% of costs. The ROI is an AI-capable engineer who already knows your systems.
3. Compress your AI hiring pipeline to 7 days. Map every step in your current process. Identify where candidates are waiting. Eliminate any step that takes more than 48 hours. Pay for the technical assessment. Make the offer within 7 days of first contact.
4. Contact AI Singapore about AIAP placements. The programme deploys trained AI engineers to participating companies. If you are eligible, you get a 9-month-trained AI engineer with IMDA/EDB co-funding. The application window is continuous.
5. Bridge the gap with staff augmentation. While your upskilling and AIAP placements mature (3-9 months), use a platform like HireDeveloper.sg to bring in pre-vetted remote AI engineers who can start contributing within 14 days. This is not a replacement for local hiring — it is a bridge that keeps your delivery pipeline running while your structural solutions take effect.