Meta & Microsoft – Everything AI

Meta Platforms Q3 24 – Boondoggle?

• Meta reported good results but the continued heavy investments into AI and The Metaverse from which returns are hard to quantify, made the market worry whether Mr Zuckerberg is going back to his old profligate ways.
• Q3 24 revenue / EPS was $40.6bn / $6.03 which was broadly in line with revenue forecasts of $40.2bn but ahead of the EPS forecast of $5.22.
• Guidance for the coming quarter was also broadly in line with expectations with revenues expected to be $45bn – $48bn which is very slightly weaker than the consensus estimate of $46bn.
• Meta’s social media properties continue to be the powerhouse with increases in time spent of 6%-8% YoY as a result of using AI to improve the relevance of content shown to users.
• This in turn increases both the opportunity to display adds and the ability to target them more effectively, making them more valuable.
• However, this is where the profitable news ended as Meta intends to continue investing heavily in AI and The Metaverse which was evident at Reality Labs which lost a massive $4.4bn in Q3 alone which is the equivalent to an eyewatering $49m every day.
• This is how Meta is able to sell pretty good VR devices at cheap prices as well as invest billions in working out how to build augmented reality of which Orion is its latest and greatest effort.
• These investments are likely to continue in 2025.
• If Mr Zuckerberg is right, then these investments will pay off handsomely but even in this case, it is going to take a long time.
• Either way, I think that the core businesses remain fairly valued, and so I have no real inclination to take a wild bet on whether Mr Zuckerberg is the next Steve Jobs or not as this should be the free option, not the investment case.

Microsoft FQ1 25 – Capacity Constrained.

• Microsoft reported another good quarter but its inability to build data centres fast enough has constrained its guidance for the coming quarter, which strangely sent the shares down 3.7% in after-hours trading.
• FQ1 25 revenues / EPS were $65.6bn / $3.30 slightly ahead of forecasts of $64.4bn / $3.10 but there was a slight disappointment when it came to expectations for Azure.
• Once again Azure was the engine of growth with 34% growth YoY, (just behind the much smaller Google Cloud which grew by 35% YoY), which given Azure’s size, is a big achievement.
• AI is responsible for the acceleration of growth from the low twenties back to the mid-thirties and revenues from AI will hit a $10bn annual run rate during FQ2 25.
• This is not as amazing as it sounds as OpenAI is spending like crazy and I estimate that it alone is making up more than half of Microsoft’s AI revenues.
• Unfortunately, Azure’s growth will drop to 31%-32% in FQ 25 which was taken badly by the market but I see this as a very strong long-term sign.
• Microsoft explained this shortfall relative to expectations by stating that it had more demand than it could handle and that it could not build the data centres fast enough to meet current demand.
• This is a good sign for Microsoft for the rest of fiscal 2025 as it is a strong sign that the $58bn annual run rate of capex is quickly producing revenues.
• It is also a good sign for Nvidia which reports on November 20^th where the market will be looking for confidence that the AI party will continue into 2025.
• Microsoft is clearly reaping the benefits of the AI boom and remains well positioned, but at 32x FY2025 PER, I think I would rather have the purer play of Nvidia which is not trading at a meaningfully higher valuation.
• However, given the choice, I choose to stick with my AI adjacencies of inference at the edge and nuclear power.