.png@webp){kind=logo}
FC Insights
The AI landscape has been accelerated rapidly, expanding from the emergence of Large Language Models (LLMs) across the entire AI stack, from infrastructure to application layers, including both agnostic and specialized solutions.While the core AI structure remains unchanged, adoption and investment in AI have shifted decisively from experimentation to real-world deployment.
Understanding the modern AI stack and the value it creates within this ecosystem is therefore critical for investors to capture opportunities in the new direction of AI models in the region.
Understanding the modern AI stack and the value it creates within this ecosystem is therefore critical for investors to capture opportunities in the new direction of AI models in the region.
The Modern AI Stack
Multiple infrastructure layers formed the building blocks of today’s AI. Based on our mapping, the AI stack can be understood across four layers:
.png@webp)
Shifting Value Concentration
In this new age, it is less likely to find emerging companies in the foundational model as it demands massive capital, long research cycles, global distribution, and access to cutting-edge compute. Hence, though structurally the architecture remains, there has been a noticeable shift in value concentration within the Modern AI Stack.
As AI adoption scales, systems become more autonomous and complex, making failures costlier and multi-agent setups become the norm. This drives lasting demand for orchestration, evaluation, and observability, the core infrastructure that addresses the real operational pain in AI solutions.
Companies that build on top of AI or pre-trained models for specific use cases are attractive investment opportunities because there is significant room for specialization. A clear niche and strong value proposition create defensible moats. Globally, startups cluster in this layer because buyers already exist and revenue comes earlier. To better understand the shifting value and demand, we break down recent enterprise spending across each layer into more specific segments.
Companies that build on top of AI or pre-trained models for specific use cases are attractive investment opportunities because there is significant room for specialization. A clear niche and strong value proposition create defensible moats. Globally, startups cluster in this layer because buyers already exist and revenue comes earlier. To better understand the shifting value and demand, we break down recent enterprise spending across each layer into more specific segments.
Where is the Demand?
Global spending indicates that enterprise AI adoption increasingly happens through embedding models into existing workflows, integrating them with enterprise data, and operating them reliably at scale. As a result, AI spend and adoption accelerate down the stack, with application-level software capturing the majority of enterprise budgets.
*Note: Layer 2 (Data Infrastructure) is not included, as it is adjacent to the AI stack. Vertical AI in our compilation includes both ‘Vertical AI’ and ‘Departmental AI’ as defined in the original report.
Source: Menlo Ventures
The modern AI stack is stable. What’s changing is where value concentrates, not the architecture itself.
1) Demand and Adoption Concentrate Up the Stack
Companies need reliable deployment, workflow integration, data connectivity, and predictable costs. As these needs expand, funding is increasingly flowing towards layers that operationalize AI in the real world.
2) Emerging Sectors in AI
Globally, there is an up-and-coming topic of Horizontal AI and Vertical AI, where more and more companies are emerging in the space. Horizontal AI, which refers to general-purpose platforms applicable across industries and functions, continues to consolidate market share as the default application layer by absorbing vertical features. However, opportunities for Vertical AI, which include industry- or function-specific solutions, can thrive where they are deeply embedded in proprietary data and critical workflows.
As demand and technology evolve, the dynamic between horizontal and vertical AI will shape market share.