My “mishmash” Amazon isn’t exactly cheap right now, at least in those pump-investor posts where it’s compared to hype stocks.
5 Likes
Measured by the EV/EBITDA multiple, they look quite reasonably priced. The difference compared to P/E ratios stems from the large cash reserves of most.
10 Likes
Cash reserves are one explanation. 
Many, but especially Amazon, have quite high depreciation and amortization, which is why EV/EBIT might be better from a shareholder’s perspective. However, these businesses cannot stop their investments abruptly; instead, due to the AI race, they are forced to shovel an ever-growing portion of their cash flow into increasing computing power.
For example, Amazon’s EBITDA for the last 12 months is 140 billion dollars, but EBIT is “only” 80 billion. Depreciation and amortization are as much as 60 billion dollars! Depreciation is a real cost for shareholders, reflecting something like the collapse in a car owner’s asset value the moment the car rolls out of the dealership.
Many have also started taking on debt for investments, which weakens their net cash position. For example, Amazon’s cash reserves have risen from 84 → 94 billion dollars over five years, but at the same time, debt (excluding lease liabilities) has risen from 35 → 58 billion dollars.
Meta is dramatic in this regard with Zuck going “all in” on the AI boom. As recently as 2020, Meta had 60 billion dollars in cash and no debt at all. Currently, the company has 44 billion in cash, but 30 billion in debt! 
All in all, this is an interesting change. Just a few years ago, the mega-techs (except for Amazon) were known as capital-light, pseudo-monopolistic platform businesses that churned out plenty of cash flow. Today, they are starting to pile debt onto their balance sheets, they are becoming increasingly capital-intensive, and cash reserves are not growing or are even dwindling.
27 Likes
Amazon’s depreciation and amortization are indeed very high, and that is the main explanation for the low EV/EBITDA ratio. Below are the AI’s explanations for the depreciation.
8 Likes
I’m not that familiar with Meta, but does it offer Meta AI to other companies or consumers for a fee? Alphabet’s, Amazon’s, and Microsoft’s AI investments, however, bring them money as other companies use Azure, GCP, or AWS, or even their ready-made chatbots (Gemini, Copilot).
In that sense, I wouldn’t be too worried about their over-investments, because even if it’s a bubble (which I think it is), I believe they can still run their old hardware to the end of its life and just reduce replacement investments to match the new demand situation. Weak demand can, of course, also be compensated for through price cuts.
Assuming, of course, they correctly identify a drop in demand. If erroneous over-investments continue despite a permanent drop in demand, then the damage to shareholders is naturally significant.
My reflection focuses on Meta. Below is a table produced by Gemini showing Meta’s results and profit distribution.
| Year | Net Income (Profit) | Buybacks | Dividends | Total Returned |
|---|---|---|---|---|
| 2020 | 29.1 | 6.3 | 0 | 6.3 |
| 2021 | 39.4 | 44.5 | 0 | 44.5 |
| 2022 | 23.2 | 28.0 | 0 | 28.0 |
| 2023 | 39.1 | 19.8 | 0 | 19.8 |
| 2024 | 62.4 | 30.1 | 5.1 | 35.2 |
| 2025* | 56.0 | 31.0 | 5.3 | 36.3 |
| TOTAL | 249.2 | 159.7 | 10.4 | 170.1 |
Meta is an absolutely massive cash machine, and even though it has returned 170 BUSD to its owners, roughly speaking, investments have taken 80+46=126 BUSD, and in my opinion, the profitability of these investments is considerably more questionable than with the aforementioned trio.
Of course, that includes Zuckerberg’s Metaverse nonsense, for which the market rightly punished Meta.
If we look at Apple, it is said to have been left out in the cold in the AI race without its own LLM, but I don’t think it’s necessarily such a bad and risky move. Apple sells an enormous number of devices, so it certainly has pricing power against Alphabet, OpenAI, Anthropic, you name it, if it decides to implement “Siri 2.0” as a white-label solution.
To speculate a bit more, if I remember correctly, authorities have restricted Alphabet from buying its way into being the default search engine for browsers (which, by the way, has a huge impact on Mozilla, among others). Apple has also received massive compensation from Alphabet for years, but could it be that Alphabet compensates with, for example, a customized white-label version of Gemini if Apple continues to use Google as the default search engine? It would be quite a cost-effective way to quickly get one of the best models on the market into the iPhone, iPad, and Mac.
8 Likes
I wrote a status update on the Mag7 companies for Piksu.
13 Likes
In the US, tech CAPEX is running hot, and the big culprit behind this is, of course, AI stuff. Historically, the situation is quite significant, as you can read from the tweet. I thought this would fit well in this thread. 
https://x.com/KobeissiLetter/status/2009429665421885480
9 Likes
The tweet below complements other posts in this thread regarding the valuations of the companies, etc.
So here is another tweet regarding Mag7 valuations 
2 Likes
And sure enough, it turns out that Siri will be built on Gemini. There’s no mention of the deal’s value, but for Alphabet, this will likely drop straight to the bottom line, as Apple will be responsible for operating the models.
In case anyone is worried, it seems Google won’t be able to snoop on the data:
As previously reported, a customized version of the Gemini AI model developed by Google will run on Apple’s Private Cloud Compute system servers, enabling privacy-respecting operations.
News:
20 Likes
Chinese tech stocks are attracting investors with their efficiency and rapid AI innovation. Furthermore, earnings growth for the industry’s largest companies is projected to accelerate to the point that it may surpass US tech—the Magnificent Seven—this year for the first time in years.
The Chinese markets are being bolstered by domestic IPOs and the country’s pursuit of technological self-sufficiency. China’s strong and strengthening position in hardware and software development makes it a key player in the AI transition, which will likely further fuel investor expectations.
https://x.com/MacroCharts/status/2012435741083623457
This Bloomberg article has more on the subject, but it is apparently behind a paywall—at least if you have used up your “free articles.”
“AI is a multi-year global growth driver, and North Asia’s technology ecosystem spanning hardware, software, and infrastructure positions the region at the forefront of this trend,” said Gary Tan, a portfolio manager at Allspring Global Investments in Singapore.
1 Like
It would be strange if this deal gets cleared by the EU authorities.