The Next Phase of AI Growth: From Infrastructure to Enterprise Application Providers

The AI investment landscape in 2025 is undergoing a seismic shift. For years, venture capital and private equity firms prioritized funding for AI infrastructure and foundational model development, betting on the long-term potential of companies like OpenAI and Anthropic. However, as the year progressed, a clear trend emerged: capital is now flowing toward enterprise application providers, particularly those leveraging agentic AI to deliver immediate productivity gains. This shift is not merely a short-term fluctuation but a structural realignment driven by market demands, valuation dynamics, and the urgent need for tangible returns.

The Infrastructure Era: A Legacy of Long-Term Bets

In the first half of 2025, AI infrastructure remained the dominant focus for investors. Hyperscalers such as AWS, Google Cloud, and Microsoft Azure continued to pour capital into foundational models, with global AI infrastructure spending projected to reach $527 billion in 2026. These investments were justified by the belief that infrastructure would underpin the next wave of innovation. However, the returns on these bets remain distant. As of Q4 2025, public hyperscalers like Oracle faced scrutiny over their AI-related capital expenditures, with investors demanding clearer pathways to profitability. For instance, Oracle's $15 billion in new AI infrastructure spending highlighted a critical challenge: while capex is surging, the payoff in terms of revenue and profitability has lagged.

Meanwhile, the AI sector as a whole is entering a phase of reassessment. Major AI stocks faced a 12% drop in Q4 2025 as investors grew wary of speculative growth narratives. This market correction underscores a growing skepticism toward infrastructure-centric models, which often require years of R&D before generating revenue.

The Rise of Enterprise Applications: Immediate Value, Measurable ROI

In contrast, enterprise application providers are capturing investor attention with their ability to deliver immediate value. By year-end 2025, spending on AI applications had surged to $37 billion, with startups now dominating 63% of the revenue share in this space. These companies are leveraging agentic AI-autonomous systems capable of automating complex workflows-to address specific operational and customer-facing needs. For example, agentic AI is projected to grow at a 150% CAGR, reaching $51.5 billion in enterprise spend by 2028. This growth is driven by its ability to streamline decision-making and automate tasks that previously required human intervention.

The appeal of enterprise applications lies in their product-led growth (PLG) strategies, which now account for 27% of AI application spending. Unlike infrastructure, which requires significant upfront investment and long deployment cycles, PLG-driven applications allow enterprises to adopt AI solutions incrementally, scaling only as value is demonstrated. This model aligns with the current risk-averse environment, where investors prioritize companies with clear revenue traction and customer retention metrics.

Valuation Dynamics: Agentic AI Startups Outpace Hyperscalers

The valuation gap between agentic AI startups and public hyperscalers is widening. In 2025, agentic AI startups raised record sums, with neuromorphic computing firm Unconventional AI securing $475 million at a $4.5 billion valuation. Similarly, synthetic research startup Aaru achieved a $1 billion valuation in a multi-tier Series A round. These figures reflect investor confidence in the sector's potential, with organizations projecting an average ROI of 171% from agentic AI deployments.

By comparison, public hyperscalers face a valuation crunch. Despite their dominance in infrastructure, companies like Meta and Oracle are grappling with questions about the efficiency of their AI investments. Meta's partnership with Blue Owl to raise $30 billion in private credit for AI infrastructure highlights the sector's reliance on alternative financing methods. This shift signals a loss of confidence in traditional capital markets to fund infrastructure at scale, further tilting the playing field in favor of startups with clearer monetization paths.

Governance Risks and Strategic Integration

Critics argue that agentic AI's rapid growth comes with governance challenges. IDC warns that 20% of G1000 organizations could face legal or reputational risks by 2030 due to poor AI agent governance. However, these risks are not insurmountable. Enterprises adopting agentic AI are increasingly prioritizing traceability, accountability, and ethical frameworks, ensuring that innovation does not come at the cost of compliance. For venture capital and private equity firms, this represents an opportunity to invest in startups that integrate governance from the ground up, rather than retrofitting it later.

Conclusion: A Strategic Reorientation for Investors

The data is unequivocal: venture capital and private equity must now prioritize agentic AI-driven software companies over public hyperscalers. The shift from infrastructure to enterprise applications is not just a response to market conditions but a reflection of where value is being created. Startups in this space offer higher ROI, faster deployment cycles, and a clearer alignment with enterprise priorities. While hyperscalers will remain critical to the AI ecosystem, their long-term returns are increasingly uncertain in a landscape that demands immediate results.

For investors, the next phase of AI growth lies in backing companies that can bridge the gap between innovation and execution. Agentic AI startups, with their focus on productivity gains and scalable applications, are uniquely positioned to lead this transition.