Yeah, I didn’t even notice that part, and it’s definitely a clear tactical move in that regard. Of course, it simplifies the big picture if there are fewer subscriptions. It doesn’t simplify things for the user if the “feature list” for a subscription is 10 pages long. Meaning, when you bundle enough things into the same pile, most of them go unused. That’s how I feel when looking at that “pizza menu” list above. There’s music creation, video creation, etc.—90% of subscribers surely won’t even try these more than once or twice. With their own TPUs and the efficiency of Gemini models, the operation stays cheap to run, and it’s possible to drive down prices. The company is profitable, though the search engine accounts for most of it. Compared to that, these OpenAI visions of being “profitable in 2030” feel wild. And Alphabet’s price cuts certainly won’t cause cheers among competitors. If someone starts paying €5/month for Gemini now, why would they even switch to a €5/month ChatGPT?
Google is indeed difficult to price. It’s down 10% from the peaks, and back then it was very expensive 
I’ve tried more to stay on board with the winners because they are doing things right. Puuilo’s results just came out and it’s the same case; it’s not cheap, and if there’s a +5% rise today it won’t get any cheaper, but it’s worth sticking with them when the right things are being done and the business is rocking.
I also agree regarding the share buybacks, although I think there was quite a lot of maneuvering involved there to ensure dilution isn’t as high as it could be. Of course, when shares have been bought back cheaper, selling can technically make sense, but it does cause headaches for investors. I would hope for the CAPEX growth boom to calm down, but I fear a drop in investments would scare investors into thinking the whole thing is starting to fizzle out. A couple more years like this and 700 billion will have been pushed into investments, and then we’ll be sitting there wondering about the returns.
So far, the numbers have supported the investments. When I originally invested in Alphabet, my thesis was Google Cloud, its growth, and achieving profitability. This was supported by real-world observations of how little, for example, large companies operated in the cloud. The cloud might not have been used at all, other than some personal small SharePoint where the desktop was backed up, etc. Now all information management can be in the cloud, and many applications also run in the cloud. The change has been huge, and in that regard, the thesis has been successful, as Google Cloud’s numbers also show.
This AI-Gemini thing, on the other hand, is very hard to predict, and when you add in Anthropic’s IPOs and Google’s ownership stakes and the overheating of the entire AI industry, it brings beads of sweat to the forehead. Fortunately, the investments support their own ecosystem as a whole, and there are actually paid subscribers for the products, like you are now 
. The threat of regulation is a major risk because AI matters only expand this “himmeli” (complex construct).