Google & AMD Q4 24 – AI Wobble?

Google Q4 24 – Core business still sound.

• Alphabet reported disappointing sales as, like Microsoft, it could not keep up with demand for its cloud services, sending the shares down 9%, creating an opportunity for those interested in growth at a reasonable price.
• Q4 24 revenues / EPS were $94.5bn / $2.15 missing revenue forecasts but beating profit expectations of $96.7bn / $2.12.
• The miss was caused by Google’s cloud division where growth was 30% YoY compared to forecasts of 35% YoY which was not helped by a massive increase in its plans for capex in 2025.
• Here Google expects to spend $75bn on capex in 2025 way ahead of forecasts of $57.9bn which disquieted the market still further.
• Despite the concerns, generative AI continues to be unable to affect the search business as Google reported $54bn in search revenues for 2024 just ahead of estimates and up 13% YoY.
• This is further evidence that despite being around for over 2 years and having hundreds of millions of users, generative AI is not yet in a position to challenge the digital search industry.
• While Google clearly has no moat at all when it comes to Gemini which is no better than anyone else’s, it does have a moat when it comes to billions of users who use Google services on a daily basis.
• If generative AI is going to take over, Google has a better chance than anyone of being a leader as it can gradually infuse its existing services with generative AI and, assuming its services are good enough, migrate its existing user base over.
• This is exactly what it is doing with search summaries and the presence of a Gemini chat assistant on Android devices powered by Gemini Nano.
• Hence, I think Google’s position and financial position look relatively secure meaning that at 23.1x 2025 PER, the shares offer reasonable value for the growth that they are delivering.
• If the slide continues and extends, then I would consider taking a position in Google.

AMD Q4 24 – Impenetrable Nvidia

• AMD reported good results but the fact that the company missed expectations in the data centre, put a dent in the AMD AI story and sent the shares down 8% in after-hours trading.
• Q4 24 revenues / EPS were $7.7bn / $1.09 broadly in line with estimates of $7.5bn / $1.08 and guided reasonably well for Q1 25 with revenues of $6.8bn – $7.4bn ($7.1bn) compared to the consensus estimate of $7.1bn.
• However, the data centre business did not fare as well as hoped and given that a lot of expectations are pinned on this division, this was a problem.
• Here, Q4 24 revenues were $3.86bn, up 69% YoY but short of expectations of $4.09bn.
• The company also gave a tepid (relative to expectations) forecast for the division where it expects “strong double-digit growth” for 2025 which in no way can be interpreted as anything like the 46% that the market is forecasting.
• Hence, estimates are likely to come down for 2025 which is why the shares fell in after-hours trading.
• The problem here is one that RFM has outlined in its research for quite some time which is that in this generation, no one is going to put a dent in Nvidia.
• This is due to Nvidia’s CUDA platform and its ability to keep its silicon at least one generation ahead of everyone else.
• While silicon development remains the control point in the AI ecosystem, it is very unlikely that anyone else is going to take share from Nvidia
• This is not what the market was expecting and is what I suspect underpins the miss relative to expectations that AMD has guided for in 2025.
• Consequently, I continue to think that the only thing that is going to trouble Nvidia for the next year or so is end demand but with the hyperscalers raising their capex forecasts (see above) to even crazier levels, demand looks like it is going to remain strong.
• This is why I continue to think that AMD looks expensive on 2025 PER of 24.0x and I see risks to the estimate meaning that the real PER ratio could be higher.
• I would have Google or Nvidia over AMD without having to think about it too hard.