From Experiment to Infrastructure: What’s Next for AI?

Why It Matters

The surge in AI‑agent investments and the move toward disciplined, infrastructure‑backed adoption will reshape enterprise productivity and create new market leaders in the coming year.

Key Takeaways

• 433 AI‑agent deals in 2025, topping all sectors
• Companies prioritize AI for high‑value, data‑ready use cases
• AI adoption now seen as mainstream, like early Google
• UK earmarks £100 bn for AI infrastructure growth zones
• Enterprise “agent rails” still missing, hindering scale‑up

Pulse Analysis

The AI‑agent market exploded in 2025, with 433 deals recorded, eclipsing even the fast‑growing med‑tech sector's 324 transactions. This volume reflects both venture capital confidence and a clear demand for autonomous software that can act across cybersecurity, supply‑chain, sales and customer service. Investors are now looking beyond single‑purpose applications toward platforms that can be customized for enterprise workflows, driving higher valuations and faster exit cycles. The momentum signals a transition from niche experimentation to a core component of the digital economy, reshaping competitive dynamics across multiple industries.

Enterprises are moving from ad‑hoc tinkering to purposeful AI roadmaps. Leaders at Box and Omnea stress aligning agents with measurable business outcomes, cleaning data, and automating high‑value tasks. Cultural adoption mirrors the early diffusion of Google, with firms segmenting users into early adopters, mainstream, and skeptics, and building internal channels to boost confidence. Senior executives, once slower to change, are now being engaged through clear value propositions and governance frameworks. This disciplined approach reduces “dumb work,” embeds AI into daily processes, and creates a scalable foundation for sustained productivity gains.

Infrastructure remains the bottleneck to truly ubiquitous AI. Cherry Ventures highlights a shortage of hardware, high‑speed networking, and robust ML frameworks needed for enterprise‑scale agents. The UK government’s £100 bn AI growth zones aim to accelerate planning permissions, grid access and data‑centre power, addressing Europe’s current ingredient gap. Yet investors note that “agentic AI rails” – the orchestration layers that let autonomous bots perform complex, human‑like tasks – are still immature. Funding pipelines for 2026 are expected to target these foundational layers, promising a more reliable backbone for the next wave of AI‑driven business transformation.