MSFT FH26 Top 10 Holding

MSFT FH26

MSFT reported revenues up 17%, Op Y up 21% and NPAT 23%, there was a positive Open AI MTM (omitted). These numbers were slightly better than my expectations.

The segmental breakdown showed the software-based Productivity Processes up 16%, Intelligent Cloud, which houses Azure, up 29% and More Personal Computing, which is gaming and PC’s, up 6%. The consumer businesses continue to be a disappointment. An impairment charge was made against the gaming assets. The Activation Blizzard acquisition now looks ill-timed. As MSFT is a huge software business, there is no sign of AI disruption so far in these numbers. Apart from the consumer businesses, the results were good.

MSFT highlighted the strategic position they are in to utilise and embed AI into their customer base and the progress made in selling Copilot into the base. The company mentioned that 15m paid seats of MSFT365 were in use. Although the trend is positive, will it be enough, and will execution continue? Are the relevant questions. Certainly, MFST management is clearly aware of the potential.

As agents proliferate, every customer will need new ways to deploy, manage and protect them. We believe this creates a major new category and significant growth opportunity for us

Probably worth acknowledging that MSFT key is its Windows products embedded in most of the world's companies. That opens the door for their other products, even if they are not the best products available. MSFT didn’t create anything innovative in AI from scratch itself, as GOOG and OpenAI, Anthropic have, but they are a fast follower and adapt at getting their products into customer usage.

There are concerns, and these comprise GM deterioration, ROIC deterioration and OpenAI exposure.

The GM deterioration is a function of two things. Firstly, sales mix, as IC increases as a larger % of the whole, it has lower margins. The decline of MPP with lower margins partly offset this. PB&P has the highest margins of the segments. No big deal here. The second is the inherent deterioration in the mix of Azure's margins as hardware becomes a larger share of the mix, whether this is structural or will revert over time is unknown. At this stage, it's not a huge deal.

The next part is potentially a huge deal. The chart below shows cash less SBC and capex divided by capital employed, being SHF plus net debt. The metric has been in decline for some time and has moved from outstanding returns to very good. The issue is clearly the lag between investment and returns, especially over the last few years, as AI capex ramped up. The thesis here relies on what basically all big tech CEO’s have been stating, largely in unison. What they see are outstanding returns on the older vintages; the much larger size of immature recent vintages of capital blurs the underlying returns. However, as long as older vintages mature into great returns, they are willing to continue spending. Logically, this makes sense, and it probably means that one vintage will offer poor returns. As further evidence, all management consistently state that demand exceeds supply, and that is different from past capital binges, where it was a build it and they will come exercise. The importance of these two issues being true cannot be underestimated in the investment case. Otherwise, ROIC declines and keeps declining as long as MSFT keeps spending, and that's not in the outlook.

The third issue is the OpenAI exposure. Being private, OpenAI's financials are not available, so it is piecemeal and potentially cherry-picked. We do see one large successful product, ChatGPT, but with still low monetisation. There is a range of other products, and there will continue to be new ones. OpenAI is hugely CF negative and will need large amounts of outside capital to reach profitability, which is well into the future. Current valuations are int eh $500-750B range, a huge valuation given the current state of play. MSFT has three exposures. Firstly its equity stake at 27%. On top of this is access to OpenAI IP for several years. MSFT invested about $16B, small given the size of MSFT. Getting in at an apparently low valuation makes this a low-order issue for MSFT.

The second is that OpenAI is the largest customer of Azure. MSFT disclosed that they have a $625B order book (which is enormous, even for MSFT) and that OpenAI comprised 45% of that book. MSFT's motive in disclosing this appears to be that they wanted to highlight the other 55%, which is broad-based and growing at 28%, as a cause for strength. However, that appears to have backfired and only focused the market on the OpenAI exposed backlog. A chart of the backlog appears below. Even removing OpenAI, the outlook is reasonably, if we credit OpenAI with a smaller real contribution, the backlog is fine. The next issue is probably the most important: what happens under the scenario where MSFT builds all this capacity for OpenAI, and they don’t have the capacity to pay for it, their funding dries up. MSFT management here says that the payback for the shorter dated capex (GPU/CPU) is quite short, and they will get their capital returned. Under that scenario, MSFT would experience poor returns on their capital. Even waiting for the rest of the demand to fill it up would be a poor outcome, but not company-threatening.

Many scenarios, but MSFT would surely like to see OpenAI fully funded and so derisk its capex. That, in turn, relies on OpenAI continuing to convince investors that its products have massive use cases and new products will see the same.

SUMMARY and VALUATION

My base case valuation is that MSFT generates 16% eps growth over the next 5 years and at a 25X exit multiple, using the current SP of $434, generates 11% return, attractive for a company of MSFT's stability and track record.

The big winds that threaten the 25X exit multiple are a software existential threat from AI, returns on AI capex being disappointing, and the OpenAI relationship failing to pay off. All of these need to be monitored, especially OpenAI. Maybe 2026 is the year we get more evidence of returns on AI capex as well.