On reflection, some positives come out of last night's earnings call.
- Satya's pitch to investors is that a) MSFT will adapt its cost structure to the times, b) the company has ambition to own the start of the AI product cycle, c) and mainstays like Dynamics are still gaining share & M365 is still growing
- Amy's (CFO) guidance for Azure growth down 4-500 bps q-q from the 'exit' run-rate of Azure (35-36% y/y cc) probably has some sand to it. Likewise, guidance for l-s-d % growth for F4Q (June-Q) OPEX also seems to have some added sand
Positives maybe stop there.
A negative from the call is that adapting its cost structure to the times is a flexibility that investors like, and employees hate, as it implies another large cut may be needed by September if the outlook for growth into year end and CY24 isn't a happy one. This leaves MSFT internally in a high anxiety position, and it is very hard to accomplish great feats when people think they might be 'next'.
Another is that while the CFO's guidance likely has some sand to it, even adjusting for excess conservatism on some level, doesn't really get a winning outlook for the model. For example, we stick with our process of rolling Billings into revenue, so we are modeling Azure GAAP revenue growth rate decline of about 350bps Q-Q for now, much better than the implied 600-700bps guidance in the Q&A on 'exit run rate'. To think we might be wrong, and it would actually be worse isn't a song that leads to happiness when thinking about valuation. Further, we literally don't comprehend how OPEX can be flat to up in F4Q and we are modeling -7% y/y. But if guidance is conservative and the number is more like -1% y/y, it is still a FCF disappointment.
So we now see MSFT at 24x NTM FCF on our model. The thought that it could be worse on Azure and worse on OPEX doesn't leave us feeling like we could flip and endorse a long position.
We will stick with our original downside risk assessment on the stock of ~$155 to ~$195, or ~20-35% downside from yesterday's close, based on 15-19x forward FCF. And note: maybe our multiple is too low, maybe our FCF is too high. Wash.
Game on.
