Singaporean enterprises are moving into a more demanding phase of AI adoption: proving that deployments can deliver measurable returns while meeting security, compliance and infrastructure requirements. A new Dell Technologies study suggests the biggest obstacles are no longer enthusiasm for artificial intelligence, but data access, legacy systems, energy constraints and the lack of coordinated AI, data and security strategies.
Singapore’s enterprise AI market is entering a more pragmatic phase, with companies increasingly focused on whether artificial intelligence can move from pilot projects into secure, scalable and measurable production deployments.
That is the central finding from Dell Technologies’ Modern Enterprise Readiness Study 2026, which surveyed 2,950 business and IT decision-makers across 35 countries, including 100 respondents in Singapore. The study points to a widening gap between organizations’ ambitions for AI and their ability to build the infrastructure, governance and operational processes needed to support it.
Nearly six in 10 Singapore respondents, or 58%, say their organizations cannot keep pace with the speed of technological change. At the same time, 60% lack an actionable roadmap covering AI, data and security.
The economic environment is making that gap harder to ignore. Some 93% of respondents say their organizations are becoming more selective and focused on outcomes when investing in technology. Another 79% say they will not proceed with new AI initiatives without demonstrated business outcomes and a defined risk-management plan.
Those findings illustrate a broader shift in enterprise AI strategy. Companies are moving away from simply experimenting with generative AI toward building repeatable systems that can be governed, secured and operated at scale.
Data Is Becoming the AI Gatekeeper
Data governance is emerging as one of the most important constraints on enterprise AI adoption.
According to the Dell survey, 80% of Singapore respondents believe their data and intellectual property are too valuable to place in third-party generative AI tools. The same proportion identifies regulatory compliance as a major concern.
Security and compliance issues have already affected deployment decisions. Dell reports that 79% of Singapore organizations have slowed or paused AI adoption because of those concerns.
This creates a difficult balance for IT teams. Generative AI applications depend on access to enterprise information to become useful, but organizations must control where that information is stored, how models access it and what happens to the resulting data.
The result is growing interest in private AI environments, hybrid cloud architectures, retrieval-augmented generation, local inference and stronger data-governance frameworks. These approaches allow enterprises to bring AI closer to proprietary data while maintaining greater control over security and compliance.
Storage, Not Compute, Emerges as a Bottleneck
The infrastructure challenge may also be different from what organizations initially expected.
In Singapore, 82% of respondents identify storage performance and data access as their primary technical barriers to AI deployment, rather than raw compute capacity. Meanwhile, 76% say their data centers are not yet prepared for demanding AI workloads.
AI infrastructure requires more than accelerators. Training and inference workloads can generate substantial data movement, while enterprise applications need reliable access to increasingly large datasets. Storage architecture, networking, cooling, backup and cyber recovery therefore become part of the AI stack.
Dell says 88% of Singapore respondents plan to consolidate onto fewer, more capable technology platforms. A further 94% require integrated cyber-recovery capabilities as a baseline.
That emphasis reflects a broader evolution in machine learning infrastructure. As enterprises move AI workloads into production, resilience and recoverability become just as important as model performance.
Energy is another constraint. Some 89% of respondents identify energy availability as an important purchasing consideration, while 86% say they will scale AI only when they can control its energy footprint.
For AI cloud platforms and data-center operators, that could increase demand for more efficient accelerators, optimized model architectures and infrastructure capable of delivering greater inference performance per watt.
Shadow AI Complicates Governance
Enterprise AI strategies are also running into an old IT problem with a new name: shadow IT.
Dell reports that 92% of Singapore respondents have employees using personal devices to access AI tools for work. Seventy-three percent have discovered unauthorized AI applications on company devices, while 77% say employees are expensing AI tools through departmental credit cards without IT oversight.
The numbers suggest centralized AI policies are struggling to keep pace with employee demand.
This is significant for enterprise AI applications because uncontrolled tools can introduce risks involving confidential information, intellectual property, data retention and regulatory compliance. At the same time, blocking AI outright may simply encourage employees to find less visible alternatives.
Endpoint infrastructure is part of the equation. Eighty-seven percent of respondents say legacy PCs are limiting their ability to adopt AI-enabled applications, while 93% believe AI PCs will become central to future workplace strategies.
AI PCs could increasingly provide enterprises with local inference capabilities for selected workloads, reducing dependence on cloud services while offering tighter control over sensitive information.
From Technology Vendors to Strategic Partners
The research also points to changing expectations around technology providers.
Dell says 83% of organizations across the APJC region—Asia Pacific and Japan—are replacing transactional vendors with strategic partners that can help create multi-year technology roadmaps and share responsibility for outcomes. Ninety percent believe existing partners should do more, while 95% expect security to be embedded into technology proposals.
That shift mirrors the maturation of enterprise AI.
Organizations no longer need another isolated AI tool. They need integrated combinations of AI platforms, cloud infrastructure, data management, cybersecurity, AI development frameworks and endpoint hardware.
For technology companies such as Microsoft, Google, Amazon and NVIDIA, that creates an opportunity but also raises the bar. Enterprises increasingly expect AI infrastructure and software to fit into existing governance and operational environments rather than becoming another disconnected technology layer.
Singapore’s AI market therefore appears to be entering its infrastructure phase. The next competitive advantage may not come from deploying the most impressive model, but from building an environment where AI can operate reliably, securely and economically.
The Dell research is based on a relatively small Singapore sample of 100 decision-makers, and its findings represent respondent sentiment rather than an independent measurement of AI deployment. Still, the results highlight a challenge facing enterprises well beyond Singapore: AI readiness is increasingly an infrastructure and governance problem, not simply a model-selection problem.
Market Landscape
Enterprise AI adoption is moving toward production-scale deployment, increasing pressure on data infrastructure, cybersecurity, energy efficiency and endpoint modernization.
The Singapore findings fit a wider global investment cycle. Gartner forecasts worldwide AI spending of approximately $2.6 trillion in 2026, with AI infrastructure accounting for the largest share of spending. The shift suggests enterprises are investing not only in AI applications and models, but also in the infrastructure required to operate them reliably.
For Singaporean organizations, the key competitive issues are likely to include AI governance, private and hybrid AI, data sovereignty, AI-ready data centers, AI PCs, cyber recovery, inference efficiency and energy management.
Top Insights
- Singapore enterprises are moving from AI experimentation toward deployments that require measurable ROI, defined risk controls and production-ready infrastructure.
- Data access and storage performance are emerging as major AI bottlenecks, challenging the assumption that compute capacity is always the primary constraint.
- Shadow AI is widespread, creating new governance risks as employees adopt consumer AI tools faster than organizations can establish controls.
- Energy availability is becoming an AI infrastructure consideration, potentially accelerating investment in efficient models, accelerators and data-center architectures.
- Technology vendors are increasingly expected to act as strategic AI partners rather than transactional suppliers of standalone hardware or software.
Power Tomorrow’s Intelligence — Build It with TechEdgeAI
