The rise of AI has provoked much excitement – along with some fear. Will it really be a gamechanger for our world? Are companies investing too much in the new technology? Could this be another dotcom bubble? These are all valid concerns, but we believe the overall outlook remains positive due to many use cases, strong demand from businesses and consumers and solid corporate fundamentals.
Having invested in digital technologies for many years, we see AI as just the latest in a long line of technological advances and thus investment cycles. It follows on from mainframe computers, PCs, the Internet, mobile phones and cloud storage. The latest wave started in November 2022 with the release of ChatGPT putting the spotlight on AI in general and on generative AI in particular.
We expect this to continue to build for many years to come, in both consumer and enterprise markets.
There will be bumps on the road. We are investing in a complex, interdependent world that is attempting an enormous AI data centre buildout. Individual data centre projects can cost billions of dollars and rely on many supply chains coming together to execute them successfully. The last six months have been filled with announcements of ever-greater spending plans by major tech companies. Investors are understandably questioning the magnitude of these investments and their potential returns. They are concerned by the use of cyclical investment structures, vendor financing, debt and off-balance sheet financing.
However, taking a global thematic view of the AI investment universe, we retain an overall positive outlook. While individual companies may rise and fall, we think that the end demand signals are the most important factors in assessing where we are in the cycle and they remain strong.
Hyperscale cloud providers, such as Google, Amazon and Microsoft, are comfortable investing in excess of USD100bn each in 2026 because they have good visibility into end demand. They are also making relatively low risk investments, building "fungible fleets" of infrastructure that they can match to different workloads. And they are able to balance out demand between internal use cases (such as AI overviews and AI search for Google) and external customer workloads (either for training or inference). If they see demand slowing, they can quickly reduce their rate of investment and wait for the demand to return.
So far, demand signals and use cases appear to justify these investments. In enterprise markets we are already seeing a number of scaled AI services. Code generation is the most well-known one: thanks to AI code assistants, corporate software developers are now on average 20-40% more productive. Customer support and service is another area seeing significant hard and soft savings. As are a variety of other areas such as analytics, fraud detection, predictive maintenance and asset management optimisation.
R&D spending for selected tech companies in 2024, USD mn
Source: Pictet Asset Management, FactSet. Data covering period 01.01.2024-31.12.2024.
AI demand from consumers and businesses
More broadly, businesses are preparing themselves for an AI future. This requires an increased pace of cloud migration and digital transformation. Companies need to migrate towards more modern databases that are AI-ready (and realising significant savings as they do so). Data must have high integrity and to be streamed real time in order to deliver excellent services to consumers. Strong governance and security are required to take advantage of the increasingly "agentic" world which we are moving into, one where autonomous agents can complete tasks independently. We therefore see a broad trend of increasing corporate investment in technology.
Similarly, the market for consumer-facing AI products is rapidly growing. Following the initial explosive expansion of OpenAI’s ChatGPT we see increased adoption of, for example, Google’s AI overviews and AI search. Spotify’s AI DJ improves listener engagement and user retention, with minimal incremental cost.
E-commerce is ripe for AI innovation. Merchant platform Shopify is leading the charge – from AI chatbot assistants and product recommendations to facilitating a full agentic commerce experience, where autonomous agent can research and complete purchase transactions on behalf of users.
All of this needs major investments from hardware players. While some venture-backed companies are taking substantial operating and financial risk to get ahead of the demand curve, semiconductor and semiconductor equipment makers remain highly rational decision makers, with a close eye on return on investment (ROI). The high-performance chip designers, such as Nvidia and AMD, depend on their own supply chain to be able to meet demand. Major companies, such as the Taiwanese foundry TSMC, are working closely with their customers and also their customers’ customers to understand what true end-market demand looks like. Building semiconductor fabrication plants takes a number of years, so supply simply cannot catch up to demand overnight.
Bearish investors are making comparisons between today’s investment environment and the dotcom bubble. However, we believe they are missing some important differences. Companies such as Microsoft, Alphabet and Amazon are highly profitable global businesses. They remain supply-constrained and are investing into the demand that they see from their customers. In addition, today’s global AI supply chain is tightly concentrated in a small number of critical suppliers like TSMC and memory maker SK Hynix.
We believe that demand will remain strong in 2026 and we remain optimistic about both consumer and enterprise adoption of AI in the years to come.
This article appeared in the Neue Zürcher Zeitung’s supplement on FinTech and GenAI in November 2025.
