Big Tech: A Single Bet Dressed Up as a Diversified Basket

Our Artificial Intelligence mini-series has examined the capital expenditure supporting the AI build-out, the vulnerability of selected semiconductor leaders, the relationship between investment and monetization, and the potential portfolio exposures sitting on the other side of the trade.

In this fifth chapter, LFG+ZEST returns to the companies at the centre of the cycle.

Meta, Alphabet, Amazon, Microsoft and Oracle are commonly treated as a diversified basket of different business models.

Meta is primarily associated with advertising. Alphabet combines search, advertising and cloud. Amazon brings together retail, logistics and AWS. Microsoft is built around enterprise software and cloud, while Oracle has become increasingly connected to cloud infrastructure.

At the surface, these companies appear to represent five different exposures.

The analysis asks whether their marginal growth, capital expenditure requirements and valuation expectations are nevertheless becoming tied to the same underlying driver: the AI and cloud investment cycle.

The answer is more nuanced than a simple correlation argument.

Raw co-movement across the group is high, but part of that relationship is explained by shared seasonality. Once this effect is removed, average correlation falls meaningfully. However, a single common factor still explains approximately 71% of the variance observed across the group.

That common factor is not directly visible as one line in an income statement. It is the economic cycle connecting hyperscaler capex, cloud growth, AI-infrastructure demand, advertising efficiency and future monetization expectations.

The analysis finds that the exposure is not evenly distributed.

Meta, Microsoft and Alphabet are most closely aligned with the common factor. Amazon retains greater diversification through its retail operations, while Oracle behaves differently during the observed period. Oracle’s countercyclical behaviour, however, appears contingent on its own corporate and cloud-transition phase and should not automatically be interpreted as a structural hedge.

The research then moves from statistical co-movement to economic sustainability.

Under the stated model assumptions, committed capital expenditure implies an annual AI-revenue requirement of approximately 270 billion dollars. Estimated current AI monetization remains materially below this level, creating a gap that must be closed through higher consumer and enterprise adoption, sustained pricing and continued improvements in operating efficiency.

The framework is also applied company by company.

Meta presents the least observable monetization path because the return on AI investment must primarily emerge through its own advertising ecosystem. Alphabet and Microsoft achieve partial coverage through cloud and disclosed AI revenues. Amazon appears best covered, supported by the scale of AWS and a lower proportion of total investment attributed directly to AI.

This is not a directional call against Big Tech.

It is a structural reading of portfolio risk.

Five securities may still represent a more concentrated economic exposure than their different business models suggest. Booked diversification can therefore overstate effective diversification, particularly when several positions depend on the same marginal growth engine.

The central question is measurable: is the monetization gap closing or widening?

The monitoring framework considers paying penetration, disclosed AI revenue, cost per inference, backlog quality, pricing, capex revisions, customer concentration and AI-service margins.

Together, these indicators provide a way to assess whether the common factor remains supported—or whether the distance between committed investment and realized monetization is becoming harder to close.

This material is provided for informational purposes only and does not constitute investment advice, an offer, or a solicitation.