Microsoft earnings preview: AI spending under the microscope

On Wednesday, 28 October, Microsoft's earnings report is due, and once more the headline figures may not dominate the conversation.

That is not to downplay the results themselves. Revenue of roughly $90.6 billion is the analyst consensus, representing a near-17% year-on-year increase, alongside EPS of $4.71. Those are robust growth rates, particularly given Microsoft's existing scale.

Yet as with most major tech earnings calls, the focus among investors will be on management's commentary regarding AI initiatives rather than the numbers themselves. This is especially true following the tech stock selloff seen in the past fortnight.

Microsoft is already spending heavily to expand its AI infrastructure. Capital expenditure, inclusive of finance leases, reached approximately $41 billion in the prior quarter, and management anticipates that figure will surpass $50 billion in the current quarter — more than $50 billion across a single three-month period.

There are solid reasons for such aggressive investment. Azure revenue rose 43% year-on-year last quarter, with management guiding for roughly 45% constant-currency growth this quarter. Furthermore, underlying demand outstrips available computing capacity. Few are questioning whether businesses want more AI compute power; the demand is evident.

The real question is how much longer Microsoft must spend to meet that demand and whether the returns will eventually justify the outlay.

Building data centres and procuring advanced chips carries significant expense. Although these investments could underpin growth for years to come, they also entail substantial costs that investors cannot disregard indefinitely.

With market attention fixed on this issue, I will be monitoring Azure's growth outlook, cloud margins and any revisions to Microsoft's capital expenditure plans this week. Given the recent sensitivity in the tech sector, the implications could extend well beyond Microsoft itself.

If Azure continues to deliver strong growth and investors gain greater confidence in future returns, that could ease some of the anxiety surrounding the broader AI trade. Semiconductor firms and data centre operators in particular would benefit from such reassurance.

Conversely, if spending continues to accelerate while profit margins lag, investors are likely to question the economics of the AI boom even more forcefully.

Few doubt Microsoft's ability to produce another strong quarter. The company has offered ample reason to expect one, and with elevated expectations across big tech, blockbuster earnings have become routine.

The issue, as I see it, is that solid earnings alone may no longer satisfy investors. The market is already aware that AI demand is surging. The focus has shifted to finding evidence that the spending to capture that demand will ultimately generate returns. Microsoft's earnings will serve as another gauge of this shift in sentiment.