AI in a new phase: From capacity to concrete use – what does this mean for you as an investor?

The discussion, moderated by Jorgen Mork, explores what this transition means for value creation and investment opportunities going forward.

The Technology Sector in Transition: From Investment to Value Creation

After several years of heavy investment in artificial intelligence, the technology sector is entering a new phase. In this webinar, Marius Wennersten, portfolio manager of DNB Technology, shares his assessment of how the AI cycle is shifting from capacity build‑out to actual usage and monetization—and what this means for value creation and investment opportunities going forward. The discussion is moderated by Marius Brun‑Haugen.

From AI Investment to Real-World Deployment

The central theme is a technology market in transition. While recent years have been dominated by large‑scale investments in data centers, semiconductors, and related infrastructure, there are now clear signs that AI adoption is accelerating. The emergence of so‑called agentic AI—systems that go beyond generating responses to actively performing tasks on behalf of users—is highlighted as a key driver of this shift.

Wennersten points to the rapid increase in revenue run‑rates at leading frontier model providers such as OpenAI and Anthropic as evidence of a meaningful step‑up in activity. This suggests that AI is increasingly being embedded in operational workflows rather than remaining confined to experimental or pilot projects. At the same time, he emphasizes that the path to broad and sustainable monetization is unlikely to be linear.

A Large and Underappreciated Productivity Opportunity

An important perspective discussed in the webinar is the scale of the long‑term opportunity. Global spending on labor is vastly larger than current investment in data centers and software, underscoring the potential productivity gains AI could unlock over time. If AI can meaningfully augment or automate parts of knowledge work, the economic impact could be substantial.

However, Wennersten cautions against assuming that value creation will materialize quickly or be evenly distributed. Historically, major technological shifts have generated significant productivity gains, but also considerable dispersion in who ultimately captures the economic value.

Value Migration Up the Stack

To date, returns in the AI cycle have been heavily concentrated in infrastructure. Semiconductor and hardware suppliers have been the primary beneficiaries of the investment boom, and AI‑related infrastructure equities have significantly outperformed their customers. As a result, much of the AI optimism is already reflected in valuations at this level of the value chain.

According to Wennersten, agentic AI could act as a catalyst for value migration toward the platform and application layers. To operate at scale, AI agents must be orchestrated, integrated into existing systems, and embedded within enterprise data and security architectures. Many companies are choosing to do this through established cloud and software partners.

Cloud Platforms Well Positioned; Software Under Pressure

The three major cloud platforms—Amazon Web Services, Microsoft Azure, and Google Cloud—are highlighted as particularly well positioned for this next phase. After a period of more moderate growth, recent quarters have shown signs of reacceleration as prior investments increasingly translate into revenue. Strong customer relationships and operational flexibility are seen as key competitive advantages.

Software, and SaaS in particular, has been among the weakest parts of the technology sector over the past year. Structural concerns related to new architectures, bundling, and potential disintermediation have weighed heavily on sentiment. Wennersten nevertheless argues that parts of this pessimism have become excessive. In many established software companies, value does not primarily reside in code generation, but in domain expertise, deeply integrated workflows, and embedded business processes.

The sell‑off has also created pockets where valuations appear disproportionately low relative to long‑term fundamentals. In several cases, high‑quality software businesses are trading at low single‑digit revenue multiples despite strong margins, solid growth profiles, and meaningful potential for operating leverage over time.

Constructive, but Disciplined

The overarching conclusion is that the AI cycle is changing character. The focus is shifting from investment toward usage and monetization across the value chain, and the most attractive opportunities may lie where this transition is not yet fully reflected in market pricing. At the same time, valuation dispersion within the technology sector remains wide, calling for a selective and disciplined approach.

Wennersten remains constructive on the sector overall, supported by technology’s historical ability to deliver superior earnings growth and drive productivity gains across the broader economy. Agentic AI is viewed as the next leg in this productivity arc. However, emphasis is placed on identifying companies where long‑term earnings power and cash generation are underappreciated, while avoiding areas where expectations and valuations have become overly demanding.

For those seeking more context and deeper insights, we refer to the webinar.