In today’s digital era, organisations are using AI more than ever, yet progress remains uneven. While the tools are widely adopted, few have been embedded deeply enough into core workflows to unlock meaningful enterprise value. In fact, according to Deloitte’s “State of AI in the Enterprise” 2026 report, in just one year, access to AI across the workforce has grown by 50%, expanding from under 40% to nearly 60% of employees.
However, the study also finds that while AI is delivering measurable operational gains, especially in efficiency, productivity, and cost reduction, the leap to clear top-line business outcomes is still rare. Revenue growth from AI “largely remains an aspiration”: only 20% say they are achieving revenue increases today, even though 74% hope to do so.
This challenge goes beyond adoption metrics and points to a deeper structural shift. Today, many organisations are still using AI mainly to optimise existing processes at a surface level rather than redesigning core processes or reinventing business models, so the gains tend to show up as productivity and efficiency, not revenue growth.
To fully realise its potential, AI must operate closer to where data is generated, adapt to specific environments, and function autonomously within well-defined boundaries. This evolution is often referred to by technologists as “agentic AI.” Rather than relying on a single, centralised model in the cloud to process prompts, the future points to a distributed ecosystem of specialised agents that work together, learn from each other, and take real-time action when needed, whether on a factory floor, inside a vehicle, or within secure national infrastructure. In essence, AI is shifting toward a hybrid, multi-layered architecture where centralised training and localised decision-making continuously enhance one another.
Digital sovereignty and competitive advantage: Finding the right balance
This coincides with the need to balance continuous innovation with the growing requirements of digital sovereignty. Digital sovereignty refers to an organisation’s ability to retain sufficient control and decision-making autonomy over its critical data, digital infrastructure, and AI capabilities. Yet today, global AI investment and computing capacity remain heavily concentrated in a small number of large cloud platforms and hyperscale data centres.
These challenges are increasingly evident in Singapore. The public cloud landscape remains dominated by global hyperscalers, even as local and regional players continue to build a presence through partnerships and infrastructure investments. At the same time, Singapore has attracted significant commitments from global technology companies to expand AI and data centre capacity, reinforcing its position as a regional digital hub. While this growth underscores the country’s strategic importance, it also brings into focus a critical tension. As computing power, platforms, and AI capabilities become more concentrated among a handful of global providers, how can enterprises balance innovation with the need for control, governance, and long-term digital autonomy?
Against this backdrop, organisations are rethinking how AI should be built and operated. Distributed AI architectures are emerging as a practical response to real operational needs. Three characteristics, in particular, illustrate why distributed AI is reshaping the relationship between competitiveness and sovereignty:
1. Competitive advantage in AI is built on context
The rapid evolution of large language models has lowered barriers to entry, making raw model capability increasingly commoditised. What differentiates organisations is their ability to apply AI within a rich business context, drawing on proprietary data, institutional knowledge, and domain expertise developed over time. This has shifted AI investment toward environments that align closely with operational workflows, enabling organisations to integrate intelligence directly into business processes while maintaining governance over critical data assets.
2. Latency as a defining factor
Autonomous robots, smart grids, and algorithmic trading systems require near-instant responses that distant data centres cannot reliably deliver. Bringing processing closer to the edge improves performance and resilience, while fuelling demand for regional infrastructure and specialised connectivity.
3. Energy and sustainability challenge the “bigger is better” mindset
High-density AI clusters demand vast amounts of power and generate significant waste heat. Deploying smaller facilities in locations with access to low-carbon energy, and where heat can be reused, can help reduce both emissions and operating costs.
Toward a distributed AI ecosystem strategy
Taken together, these forces are giving rise to a layered, highly interconnected AI ecosystem. Hyperscale, or gigafactory, sites centralise the training of large AI models. Regional or industry-specific centres focus on fine-tuning models using proprietary data, while real-time inference and autonomous decision-making are executed by edge AI systems located closest to operational environments. High-performance, secure networks connect these layers, ensuring that insights flow continuously across the system to drive ongoing optimisation.
Singapore is strongly positioned within this emerging landscape. With a highly skilled workforce, advanced digital infrastructure, and a policy-driven approach to innovation, it offers a solid foundation for the next phase of growth. As AI moves from experimentation to the operational core, distributed AI is emerging as a way for enterprises to balance performance, competitiveness, and control.
The future of AI will not be defined by a single platform or architecture, but by an organisation’s ability to balance centralisation and distribution in line with its strategic priorities. This is not merely a technological upgrade; it is a strategic choice that will shape competitive positioning and long-term resilience in the AI-driven economy.
