Photo by Marius Wennersten © DNB Asset Management

By Marius Wennersten, Portfolio Manager at DNB Asset Management

After several years of heavy investment in AI, the technology sector is entering a new phase. The cycle is broadening, and the risk-return profile is shifting as value moves along the value chain as attention moves from capacity expansion to actual usage and monetisation, higher up the technology platform stack.

The emergence of agentic AI is a central driver of this transformation. We are now seeing clear signs of acceleration in adoption, not just in investment. Recurring revenues at some of the leading players, such as OpenAI and Anthropic, have risen sharply over a short period, reflecting a pronounced increase in activity as AI agents are deployed more widely to carry out concrete operational tasks.

The long-term opportunity remains substantial. Global spending on labour far exceeds that allocated to data centres and software, underscoring the scale of the efficiency gains AI could unlock over time. That said, the monetisation path is likely to be non-linear, and value creation will be unevenly distributed.

Agentic AI is shifting value creation up the chain

At this stage, returns have been concentrated mainly in infrastructure. Semiconductor and hardware suppliers have been the chief beneficiaries of the investment cycle, with AI infrastructure stocks significantly outperforming their peers. As a result, we believe valuations in this segment are a significantly higher premium of “AI optimism” than those of the platform and application layers.

We expect agentic AI to become a key mechanism for transferring value further up the hierarchy. AI agents need to be coordinated, deployed and embedded into business processes, and many companies are choosing to rely on their existing cloud and application partners to do so, where data, compute power, security and identity management are already integrated. If this trend continues, the value created by the shift should accrue not only to infrastructure, but also to the platform and application layers.

We believe that the three leading cloud platforms — Amazon Web Services, Microsoft Azure and Google Cloud — are well positioned. Cloud growth has accelerated again over the past few quarters as prior investments are now being converted into revenue. These platforms benefit from close relationships with enterprise customers and strong operating leverage, while retaining the flexibility to reallocate capacity if AI-related spending slows. We continue to see the risk-return profile of this segment as attractive.

Software: under pressure, but not at the end of the cycle

The software segment remains the subject of intense debate. SaaS has materially underperformed global indices over the past year, weighed down by a range of structural concerns and by the lack of any tangible improvement in fundamentals.

Some of these concerns are valid. Disintermediation, bundling, consolidation among players and lower switching costs can all put pressure on incumbents, particularly as new architectures reshape how problems are solved. Customer service is one example: the way this function will be delivered in 2030 is likely to look very different from 2020, and AI-native players could capture part of that shift. Even so, established companies retain competitive advantages linked to embedded data, existing integrations and the depth of their processes.

We also believe some of the arguments have been overstated. The idea that AI-assisted programming undermines the value of enterprise software overlooks where that value truly comes from. In most cases, it does not lie in code generation, but in domain expertise, workflow integration and the structuring of complex processes. Companies such as SAP derive their economic value from the depth of their processes and their integration into customers operations, not from their ability to write code.

Against this backdrop, the market correction is creating pockets where valuations appear excessively low relative to fundamentals. In some cases, high-quality software companies are trading at single-digit revenue multiples despite high gross margins, sustained growth and significant long-term potential for operating margin expansion.

From build-out cycle to adoption cycle

The AI cycle is evolving from a build-out phase into an adoption phase. Attention is shifting from infrastructure expansion to monetisation across the entire value chain, and we identify the most attractive opportunities where this transition remains underappreciated.

We retain a constructive view on the technology sector, supported by its long-standing ability to deliver above-average earnings growth and to drive productivity gains across the broader economy. We see agentic AI as the next stage in this productivity cycle. At the same time, valuation dispersion across the sector remains wide, which calls for an approach that is both constructive and disciplined. We favour companies whose long-term earnings and cash flow generation potential appears undervalued, and we avoid segments where expectations and valuations look excessive.

Find all our Economy articles