Infrastructure Gaps Hold Back Singapore Firms From Scaling AI, Study Finds

Singapore organisations are pushing hard on artificial intelligence deployment, but a significant gap between ambition and infrastructure readiness is holding back real-world impact. A new study from data streaming company Confluent, surveying over 4,600 IT leaders across 14 countries, reveals the specific bottlenecks slowing Singapore's AI scale-up — and they are not about strategy or budget.

The 2026 Data Streaming Report found that while 75% of Singapore organisations are already deploying or piloting agentic AI solutions, foundational deficits in data infrastructure, governance, and specialised talent are creating a persistent drag. Seventy-eight percent of local IT leaders cited insufficient real-time data infrastructure as a key barrier, 73% flagged fragmented data ownership, and another 73% pointed to a lack of AI management expertise. These aren't niche problems — they affect the majority of enterprises attempting to scale.

The gap is most visible in agentic AI, where 95% of Singapore leaders report struggling with data infrastructure and quality, 95% face legacy system integration challenges, and 93% cite large language model reliability as a roadblock. The result is stark: over 73% said agentic AI projects had stalled, and half of those were abandoned entirely. Separate research from Okta and Bitdefender adds a governance dimension — only about half of organisations have clear AI usage policies, and employees routinely share sensitive company data, including work emails (51%) and confidential documents (29%), with AI tools without clear guardrails.

Singapore's government has been proactive in building AI infrastructure — from the National AI Strategy 2.0 to IMDA's AI governance frameworks and the S$500 million AI investment commitment in Budget 2026. Yet the Confluent data suggests that enterprise-level readiness is not keeping pace with national policy. The bottleneck is shifting from "should we adopt AI" to "can our data systems support it" — a transition that demands real-time data pipelines, interoperable legacy systems, and workforce upskilling simultaneously.

Why it matters for Singapore: Singapore's ambition to be a global AI hub depends as much on enterprise execution as on policy. These findings show that throwing more compute or models at the problem won't work without fixing the data plumbing first. For Singapore's SMEs — the backbone of its economy — the infrastructure gap is even more acute, making targeted support for data modernisation as critical as AI skills training. The next wave of Singapore AI leadership will belong to organisations that invest in foundational data infrastructure, not just flashy use cases.