European enablers will be essential to the AI supercycle in 2026

JANUARY 2026

• AI infrastructure is scaling at an extraordinary pace. We observe that the world is still underbuilt relative to where we believe AI is heading.
• Europe's advanced manufacturing firms and ecosystems make the AI revolution physically possible.
• Fiscal activation across Europe, such as Germany's planned €500bn infrastructure programme, supports multi-year visibility for companies exposed to electrification, data-centre power, automation, and critical hardware supply.

'HUMANITY'S LARGEST INFRASTRUCTURE PROJECT'

While headlines have recently been awash with talk of a 'bubble', we are clear in our view going into 2026 that the AI supercycle is still only just beginning.

The world is moving rapidly toward the next major consumer breakthrough. Whether the defining device emerges in the form of glasses, wearables, or even a conversational AI companion, artificial intelligence is being deeply embedded into the everyday experience. Several large platform companies and a wave of specialised hardware innovators are already racing toward this frontier. When that moment arrives, it will likely trigger entirely new business models and services, much as the iPhone did in 2007, but we believe on a far larger scale.

Yet even before we reach that moment, the underlying AI infrastructure is scaling at an extraordinary pace (see Figure 1).

Global GPU capacity needs to increase five-fold just to meet the trajectory we are on. Token consumption (the Demand) is doubling every two to three months, but global foundry capacity capex (the Supply) is rising only around 20% per year, creating an ongoing squeeze that is fueling multi trillion dollar AI investment. Even with data centres in the US accounting for more half of private office construction and narrowing the gap (see Figure 2), the world is still underbuilt relative to where we believe AI is heading.

At the same time, AI is moving out of the data centre. Telecom equipment companies are embedding advanced processors into next generation 5G/6G base stations, marking the rise of 'edge AI' and enabling cheaper, faster inference that is closer to the end user.

Europe plays, in our view, a critical role here: it is home to what are widely recognised as world-leading network equipment manufacturers, semiconductor equipment specialists, and industrial automation firms that we consider form the backbone of this new infrastructure.

EUROPE: THE PHYSICAL, HIDDEN, INDUSTRIAL ENGINE OF THE AI AGE

This is why, in our view, Europe is far more important to the AI transition than the market currently realises. Europe's advanced manufacturing firms and ecosystems make the AI revolution physically possible. European machinery enables leading-edge chips. European telecom infrastructure enables AI at the edge. European industrial automation tech enables AI to penetrate into factories, logistics, energy networks, healthcare equipment, and advanced mobility.

The 'manufacturing continent' combines deep industrial ecosystems with a regulatory and fiscal framework that prioritises digital sovereignty, resilient infrastructure, and energy independence. We believe Europe is fundamental to the global AI and electrification value chain, not just an often cheaper alternative to the US technology exposure.

THE MACRO BACKDROP IS TURNING STRUCTURALLY SUPPORTIVE

2026 also marks the beginning of a fiscal activation across Europe. Germany's planned €500bn infrastructure programme is the most significant since Reunification, and could add 0.4-0.5% to Eurozone growth as early as this year, with a more noticeable impact from mid-late 2026 as capital actually reaches the real economy.

We are already seeing early signals: stronger defence-related order books, accelerating grid-upgrade demand, and improving pricing dynamics in energy-efficient construction materials.

This backdrop supports multi-year visibility for companies exposed to electrification, data-centre power, automation, and critical hardware supply. European corporates have generally spent the last decade becoming leaner — now they are positioned to grow again.

WINNERS, LOSERS, AND THE RETURN OF REAL STOCK-PICKING

One key theme we are watching is the widening dispersion between AI beneficiaries and those perceived to be 'left behind'. Market narratives today tend to split the world into two groups: 1) a handful of AI-native hardware and cloud players who 'win everything,' and 2) everyone else, presumed obsolete.

We think this is too binary. A number of European Software-as-a-Service (SaaS) and industrial software companies have been heavily de-rated because markets assume AI will hollow out their value. Yet the numbers do not in our view reflect that idea – many of these companies are still growing, still monetising, and are still critical in applied enterprise AI. The burden of proof lies with them, yes, but we see stock-picking opportunities as these companies demonstrate they will be enablers rather than laggards.

This environment (pockets of exuberance but also pockets of excessive pessimism) is precisely where active management matters. Being selective, understanding technological positioning, and tracking capex cycles across entire supply chains is key to capturing hidden AI leverage.

THE NEXT PHASE WILL BE PHYSICAL, INDUSTRIAL AND 'SUPER-FAST'

While 2023–2025 was more about model breakthroughs and AI cloud, we think the next phase of the AI cycle will be about deployment – into devices, networks, factories, energy systems, defence and aerospace, healthcare equipment, and everyday life. This is where we believe Europe excels.

We are moving 'super-fast', to borrow a phrase. We believe the magnitude of the transformation in front of us is vastly larger than what markets have priced into European equities today.

Looking ahead, it is our view that the AI supercycle will not slow in 2026. It will broaden across capex, industries, and borders, and we consider that Europe sits at the centre of the enabler value chain.