Warnings about ‘irrational exuberance’ are everywhere all at once. In recent weeks, several chief executives of major US banks have raised red flags, echoed swiftly by central banks from London to Washington. The International Monetary Fund has added its own cautionary note, and the financial media have chimed in almost unanimously. The Economist sounded the alarm over the weekend, the World Economic Forum’s chair weighed in, and found bubble talk irresistible. So, no one can claim that investors have not been warned.
Alan Greenspan’s famous phrase ‘irrational exuberance’ dates back to 1996 – a full four years before the dot-com bubble actually burst. True, Nasdaq valuations today look demanding, but comparisons to the late 1990s go only so far. Back then, policy errors surrounding the ‘Y2K bug’ and cheap liquidity helped inflate a speculative frenzy. Today’s dynamic is different. The surge in technology spending is driven not by hype but by geopolitics – a modern replay of the 1960s, after the Sputnik launch sparked a global race for technological supremacy. The current wave of investment in AI and digital infrastructure may prove equally transformative.
Cloud Computing & AI: Hyperscalers fuel the AI engine
Beyond the US Fed’s interest rate decision, last week’s focus was on hyperscalers’ earnings, where capex and guidance remain the core drivers of AI momentum. Hyperscalers serve as anchor clients for high-performance computing chipmakers, with their capex directly feeding upstream revenues across the semiconductor and memory supply chain. Recent earnings reaffirm an accelerating AI race. Each major hyperscaler lifted its capex guidance for 2026, with market expectations now exceeding USD500bn, underscoring a structural commitment to a continued expansion of the infrastructure. The hyperscaler capex cycle is the primary engine of AI development.
Investor sentiment remains cautious in parts of the market, especially in relation to firms that are only loosely tied to the AI infrastructure narrative. Three dynamics stand out:
- AI dominates US equity indices, with nearly half of market capitalisation concentrated in a few large-cap technology firms; hyperscalers are expanding capital commitments, reinforcing momentum; and the upstream market remains structurally concentrated – memory, semiconductor equipment, and chip manufacturing are dominated by only a few players.
- Monetisation pressure is more acute in frontier labs and AI software, where valuation multiples are elevated and revenues remain elusive. This divergence between infrastructure-led growth and software monetisation risk is a key tension.
- While concerns grow around monetisation timelines and rising debt-financed capex, demand for compute continues to surge. It is essential to note that the bulk of capex has been financed through cash flows, not debt. However, off-balance-sheet financing for some mega-clusters adds nuance to this cash flow prevalence.
Encouragingly, AI is translating into tangible revenue growth in adjacent sectors. Advertising platforms are performing strongly despite billion-dollar base effects, and hyperscalers are seeing robust growth in data-centre-related business, such as cloud computing offerings. The capex cycle remains the clearest signal of AI infrastructure momentum and is unlikely to slow significantly in the foreseeable future.
What does this mean for investors?
Of course, exuberance rarely ends quietly. But dismissing the AI boom outright could mean missing the formative stages of a structural revolution. As one seasoned investor once put it: pumping up a bubble is often the most rewarding phase. Whether this truly is a bubble will only become evident in hindsight – its bursting will be the proof of the pudding. Historically, realising that the party was over cost investors roughly 20% of their prior gains initially. In the meantime, it seems wiser to focus on fundamentals and ignore the noise. Calling the top may sound clever, but it is often premature – and this party may have far longer to run.
While monetisation challenges persist downstream, the strategic imperative to invest remains intact. Upstream firms will ultimately define the scale of the AI boom, in the meantime they have not stopped dealmaking activities financing the very ecosystem that supports them. The stakes are high, but so is the potential – we remain constructive on the Cloud Computing & AI thematic.