In August 2025, I published my first Weekly fully dedicated to this topic. It became my second most read Weekly so far. I gave it the name “Artificial Intelligence meets natural stupidity – and a potential winner no one is counting on”. I introduced my readers to the growing Capex mania of the Big Tech companies, focussing on Meta, and concluding that Apple might turn out to be the winner thanks to avoiding doing stupid things. While the mania continues at even more extreme levels, the risks for a big burst have only increased.
Summary and key takeaways from today’s Weekly
– I expand and update on the AI Capex spending mania – my even spicier part two.
– Meta stock fell sharply post earnings, despite ongoing high double-digit growth.
– It fell rightly so in my view. And chances are rising, this AI Capex nonsense cracks more with each day.
The odds of being too early with such a call are very high.
The risk to look stupid and be laughed off the stage, in addition to having missed huge (paper) gains seemingly everyone else is making almost every day, especially if you are not a professional, is close to a no brainer.
The longer it takes, the more pronounced and painful it becomes.
It’s practically guaranteed these excesses take and last longer than rationally thinkable. But it is not about rationality. It is about expectations. As long as expectations stay elevated or even keep on rising, the party can go on. This is in brief what is happening around the AI Capex cycle of the Big Tech companies, which as a reminder once were celebrated for their strong, asset-light, and high margin business models.
I’m throwing another look at Meta Platforms (ISIN: US30303M1027, Ticker: META) in this Weekly – the main target in my first Weekly about this topic (see here).
I will discuss their latest Q3 2025 earnings results which were not taken positively by the market. Plus, I have a surprising insight that strengthens my arguments and entirely cracks this artificially inflated stupidity.
If you’re looking for attractive, overlooked stock ideas the majority has a blind eye on, look no further. I am offering compelling non-mainstream cases in concise 12-page member-exclusive reports.
Want to outsmart the market? Get active now! Become a supporting paid-member and receive my best stock ideas (plus updates).
The average total return of my best stock ideas is ahead of the S&P500 and the Dow Jones. With my risks-first approach (paired with high upside), I am able to find stocks with great returns.
Join me and my members on our journey to beat the markets!
both as per 05 November 2025 market close – since August 2022
Earnings keep on rising – fine! Or not?
In my previous Weekly about this topic, I highlighted that a sole focus just on the popular headline figures sales, sales growth, and earnings (growth) is not wise, even misleading.
To be more precise, in the case of earnings, they’re close to useless.
This is due to some accounting gymnastics that allow to postpone critical cost components (depreciation of key hardware the money is spent on) well into the future, making the bottom line and growth dynamic look better than it should be under more realistic and especially conservative assumptions.
Unfortunately, only the minority is questioning the sustainability of these practices – despite sustainability being so much en vogue.
Before we come to today’s core, here is a chart showing updated Capex spending by the Big Tech companies, including the recent Q3 2025. This is on a quarterly basis.
The individual levels at this stage are not important, just the parabolic increase.
My readers know that I do not like vertical moves to the upside, as recently discussed regarding gold (see here) – the correction should not be much surprising.
When will the correction – or rather a full-blown cleanup – reach out to AI?
No one exactly knows. I do not claim to know either.
My approach does not change to keep on using common sense and to remain skeptical, respectively conservative. Risk management is key. Of course it is still possible that the bars above keep on rising. That’s possible.
But this does not mean one should take all headlines at face value.
Turning to Meta, I wanted to give a brief update about the company and the stock, as the Q3 2025 results, presented last week, have sent the stock down in the double digits under heavy volume. Not the reaction one would expect from a strongly intact growth story. Note that the stock hasn’t recovered after the big drop.
The front page story is quickly told.
Sales for the quarter rose 26%, operating income by 18%. Costs quicker than sales, impacting operating income. Net earnings were hit by a non-cash tax charge, but are not further relevant for what follows. Only glancing over these figures, the impression might be the amazing AI growth story is in full swing. But the things I have criticized in my previous Weekly, continue to cast dark clouds over the case.
First of all, little surprisingly, Meta continues to invest heavily into its AI infrastructure. Over the last twelve months, more than 60 bn. USD have been spent.
For 2025, the target was upped to 70–75 bn. USD from previously 66–72 bn. USD. We didn’t get a precise guidance for next year, but reading “significant capital expenditures growth in 2026” is concerning. Why?
Meta does not have the money.
While operating and free cash flow are still strongly positive, it is evident Meta is reaching beyond its capabilities. Free cash flow has been falling for the last three quarters – despite the huge growth story with strongly double-digit top-line growth.
Memories of the 2022 disaster with the Metaverse?
Likely too far into the past.
Currently, they are meta-worsening their financials.
Liquidity on the balance sheet is evaporating and becoming virtual.
That’s why Meta has tapped the bond market with one of the biggest corporate debt offerings ever. Some say, they must be confident. I say, this is a highly questionable move which only increases the risk profile for the stock.
30 bn. USD against a market cap of over 1.5 tr. USD is not much. But it is much compared to the company’s finances.
The free cash flow statement below reveals Meta is not only buying back stock at what I’d consider to be high multiples (while FCF is nosediving).
More than half of the buyback effectively is needed only to level out the aggressive stock based compensation program that is diluting shareholders (if shares aren’t bought back on the market to satisfy these non-cash expenses).
Then, subtracting PPE investments (Capex) from operating cash flow, we have 31 bn. USD in FCF after the first nine months. However, buybacks, SBC taxes, and the dividend together caused a cash outflow of 45 bn. USD.
They are overspending by 14 bn. USD.
As the pace is only increasing, it is clear Meta needs external financing.
With plans to jack up Capex and assuming buybacks and the dividend will be kept alive, it becomes clear the 44 bn. USD in liquidity on the balance sheet won’t last for long. By the way, these 44 bn. USD are one of the lowest liquidity levels over the last five years – while sales have more than doubled.
I made the reference to the failed Metaverse for a reason.
Meta is generating the bulk of its sales through ads. This is a highly cyclical business that during economic uncertainty will quickly run dry and experience rising consumer cautiousness. Exactly this happened during 2022 when inflation caught many people by surprise. Any form of an economic, respectively ad spend slowdown, will quickly eat into Meta’s finances.
This adds another layer of risk to the story.
Not to mention that Meta excels primarily in burning cash. The Metaverse is buried, the VR adventure has already produced more than 70 bn. USD in cumulative losses.
I have heard and read one or the other voice claiming they can just stop spending if needed. Just dial down Capex, keep the powder dry, and good is.
No, it’s not.
I’d like to point you towards the following.
As shown above – this is from the most recent quarterly report – Meta has not-to-be-underestimated binding, non-cancelable commitments in the billions.
Multi-billions I should say.
So, this has become an all-in game. No stepping back possible. And, the only guarantee we have is that Meta’s cost base will remain high for the foreseeable future. Forget the asset-light business. Get accustomed to less and less free cash flow for shareholders.
If this tastes sour already, then wait for the final serving.
Remember my discussion about estimated useful lives of assets and depreciation time?
In my previous AI Weekly, I showed that all Big Tech companies have prolonged their estimated useful lives for these key assets – mainly data center and AI infrastructure.
If you don’t have that in mind, please read that, it is important to understand.
Meta was among the most aggressive, extending their depreciation periods, boosting earnings. Concretely, I wrote:
Looking back to the annual report 2022, we can see that they back then hiked their assumption even twice in 2022, first from four to 4.5 years, and then at the end of the year to five years. […] Now, Meta in a matter of two years increases their estimated useful life for their equipment by almost 50%, now 5.5 years. The next annual report would be a great opportunity to go up to six years. No seriously, this is dangerous.
This is in a nutshell the reason why earnings growth is looking so strong while at the same time cash and cash flow are going in the other direction.
I also wrote and questioned (above the previous passage):
Some assets like buildings obviously have a significantly higher estimated lifetime. However, most of the investments are unlikely to go into buildings, but rather in quickly becoming-obsolete hardware. So is it fair to assume that they use Nvidia’s GPUs for almost six years, if we assume the average? Even if the result is four years – will they remain competitive?
Now, take a breath and see the following.
Non other than Amazon (ISIN: US0231351067, Ticker: AMZN), unlike Meta at least having significant substance with AWS, now REDUCED their estimated useful lives of their servers and equipment from six to five years.
The reason is “the increased pace of technology development, particularly in the area of artificial intelligence and machine learning”.
Who’d have thought?
In their annual report, Amazon writes further – and this is truly shocking:
I haven’t seen that during the research for my first article, as I likely focused too much on Meta.
But this is bombshell.
On one side it’s not, as it is common sense for those who don’t take every non-sense headline at face value. It is highly questionable the latest hardware from today will be doing the heavy lifting in 2031 / 2032. But this is exactly what these Big Tech companies want to make you to believe by presenting “strong earnings growth”.
On the other side, this could be the final push to finally and once and for all put an end to this stupid focus on AI-related earnings. The true earnings are very likely lower.
If not now, they will soon be.
Conclusion
I expand and update on the AI Capex spending mania – my even spicier part two.
Meta stock fell sharply post earnings, despite ongoing high double-digit growth.
It fell rightly so in my view. And chances are rising, this AI Capex nonsense cracks more with each day.
By becoming a Premium or Premium PLUS Member, you get instant access to all my already published research reports as well as several updates.
Likewise, you qualify for eight, respectively four more exclusive reports with my best investment ideas plus updates on the featured businesses over the next twelve months.