🔎Big Tech is uninvestable here.
Too expensive. Too much uncertainty. Too much in wrong incentives.
About a week ago, I posted the tweet below on the valuation of Big Tech companies. I declared them uninvestable. I believe that these companies are caught in a prisoner’s dilemma that makes it unlikely for them to earn appropriate returns on their investments.
The tweet struck a nerve. Many people seem to agree with my sentiment. It got 66,000 views, 361 likes and 38 reposts which is unusually high for me. I used to hit those numbers only during my time in the Tesla cult years ago.
However, the post also got 62 comments, most of which disagreed with me, some of them in a quite hostile tone. It’s impractical to reply to every one of them. Therefore, I want to add some more general comments in a dedicated article.
TLDR Summary
Google is trading at 60x LTM free cash flow. Microsoft is trading at 59x. Meta is trading at 71x. Amazon doesn’t even know what a free cash flow is. The combined market cap of these companies is $11tn, about 16% of the entire US stock market.
In return for paying $11tn, investors receive the cash flows of some of the world’s arguably most powerful companies, which are growing their reach (and by extension their revenues) at double digit rates. That seems like a fair deal on the surface.
But the truth is also that these companies are caught in a self-destructive race for next monopoly. It’s rational for each company individually to invest aggressively to earn the next honeypot. But when all of them are doing it, they are collectively worse off. It makes them overpay and overinvest. They are speedrunning in just a few years what typically took decades to unfold in the past. This adds considerable terminal value risk, which doesn’t seem to be priced in right now.
An important aspect of the Big Tech bull case is that they could simply dial back investments if necessary and then benefit from surging cash flows. Quite a dangerous assumption. Perhaps their competitiveness depends on continuous investment. Perhaps trimming growth investment would puncture lofty valuations.
This capex spree isn’t from a position of strength as most people seem to believe, but from a position of fear to miss out on the next big thing. Whether current and future investments will pay off or not, nobody really knows. It’s an experiment in game theory with enormous stakes for everyone involved. It makes little sense to apply astronomical multiples on this experiment. There should be an uncertainty discount instead.
How I look at Big Tech
US Big Tech companies are arguably the most powerful companies in the world, today and as a group possibly in the entire economic history. The digitalization and globalization of the economy has created unprecedented economies of scale which these companies have leveraged like nobody else.
They own most of global communication and the distribution of information. Every one of us interacts with their services dozens of times and countless hours every day. They know everything about us that is commercially and politically relevant. They own us.
They also own the political process. They have been in bed with the Democratic Party for decades and Trump’s second term has certainly established their footing in the Republican Party as well. The cultural leadership of the US then broadcasts this regulatory capture globally. Whatever Americans debate today, Europeans and Asians will debate it tomorrow.
Absolutely nobody stands in their way. The dominance and reach of Big Tech is unchecked and it keeps growing. The AI industry has launched an all-out war on intellectual property rights and nobody cares. Anthropic’s insolent and hubristic behavior is a great example for that. And it is certainly not the only example.
Anthropic is testing the limits of what is acceptable on behalf of Big Tech without compromising their brands. It’s partially owned by Amazon and Google and continues to be integrated into their product offerings. They are deep in the core of the tech establishment. Anthropic’s behavior has Bezos’ blessing.
This power of Big Tech is an asset. You can’t see that in their balance sheets. But you can see that in their earnings, their market capitalization and share price performance. Whatever happens tomorrow, investors believe that Big Tech can act upon it in a way that defends their position.
However, all of this doesn’t necessarily make them attractive investments at all times. They need to act in the interest of their shareholders. And investors need to assign a reasonable price to their equity to allow for an appropriate pace of compounding.
Acting in the interest of shareholders first and foremost requires a prudent treatment of shareholder capital, most importantly retained earnings in the case of Big Tech companies today. It means investing these earnings in a way that reliably generates attractive cash returns.
Assigning a reasonable price requires the consideration of both realistic upside in cash flows and potential downside risks.
In my opinion, both of these conditions are currently not met for Big Tech. That’s why I provocatively labeled them as uninvestable. Let’s check out the latest financial and valuation metrics for these four companies to illustrate this.
Google
Before we dive into Google’s financial performance and its valuation, let me start with a question: Do you think that the introduction of ChatGPT and its copycats has improved or worsened competitive position of Google? Do they have a stronger grip on us or a weaker one?
I believe the answer is clearly that they are worse off. Their moat has always been that nobody was able to find information online as successfully as they can. They organized the internet and made it useful for companies and consumers. This ability allowed them to charge a toll on much of the internet’s traffic in the form of advertising revenues.
They are not uniquely positioned that way anymore. Much of LLM use replaces search. It’s more challenging to monetize the use of LLMs and Google doesn’t have a monopoly in LLMs. This impaired much of the value of Google’s algorithms.
Now, have they paid the price in the form of lower revenues and earnings yet? No. In 3Q25, their revenues were 48% higher than in 3Q22, the quarter before ChatGPT launched. That’s 14% annual revenue growth, which is solid for a company of that size. (It should be noted here though that 3Q22 was the first quarter after the 2022 recession, hence a sizable portion of the revenue growth in the last three years is likely cyclical recovery and can’t be extrapolated.)
Even if consumers are not clicking SEA links to the same extent anymore, they still need to find products they want to buy. The arrival of ChatGPT probably weakened the RoI on ad spending, but that doesn’t necessarily shrink or redirect advertising budgets. Google still plays a critical role in the overall product distribution process.
But the crucial difference is this: In the past, Google’s role as the internet’s chief product distributor was uncontested. Today, they need to fight for it. Primarily in the form of insane capital spending. In 3Q22, they spent $7bn in capex. In 3Q25, they spent $24bn. Capex has more than tripled since ChatGPT launched. Have revenues tripled? No.
But the share price has tripled. And why? Seriously, why does the launch of ChatGPT raise the value of Google by $2tn? Because the innovation is so great and Google can ride that as an AI leader? What about the destruction of their moat? Google’s FCF margin is falling and we don’t know where this will end.
Markets simply assume that this capex will either pay off handsomely or, if it doesn’t, Google can dial back these investments without any negative impacts to its earnings growth and without any negative impacts to the growth expectations baked into its share price.
Adjusted for share-based compensation (yes, that is a real expense), their LTM FCF multiple is 60x. This multiple has doubled since ChatGPT has been launched.
Is there any risk priced into this stock for which investors could then earn a risk premium? Because I don’t see any.
Meta
As you may recall from past articles, Meta is actually my favorite Big Tech stock from a long-term perspective. Marc Zuckerberg has a fantastic track record with his capital allocation and the company is making courageous bets under his leadership, even and especially before these bets become fashionable. He is a much stronger leader than Pichai for example.
Meta will very likely own the next iteration of the human machine interface when we graduate from smartphones to AR headsets. It’s a huge risk for Apple’s terminal value and an equally huge opportunity for Meta.
I also believe that Meta’s AR expertise will enable them to play an important role when humanoid robots when their time has come.
However, I am not overly excited about the stock at the present moment. The company earned a pitiful $6bn in post-SBC FCF in the last quarter. Taking LTM figures, their FCF margin has tanked to 14%. It used to be north of 20%.
Markets are unfazed and expect very strong FCF growth going forward. The company’s LTM FCF multiple is now 71x, about double where it used to be.
Amazon
Amazon is deemed a laggard in the AI theme so far which might reduce the risk for a valuation excess. They don’t have the connection to OpenAI that Microsoft has which puts AWS’ cloud leadership at risk. And their perceived foundational AI expertise is considered weaker than Google’s.
Their weak cash flows are certainly playing a role, too. On an LTM basis, Amazon’s free cash flow margin has dipped back into negative territory as the company is frantically trying to catch up to Microsoft in the AI infrastructure game.
In fact, the company’s entire cumulative post-SBC FCF since 2017 is negative. Can you believe it? AWS is generating more than $100bn in annual revenues and Amazon as a whole still fails to generate a positive cash flow. And yet the company is trading at a market cap of $2.7tn.
Has lack of earnings or cash flows ever prevented Amazon stock from rising? It has sometimes played a role. Usually it doesn’t. Can a prudent investor buy it in this environment? I don’t think so. There is too much terminal value risk due to the manic AI race. There is absolutely no visibility about Amazon’s long-term potential for cash flow generation.
Microsoft
Ten years ago, Microsoft was the joke of Big Tech. They missed out on mobile computing against Apple and Google. Bing never got a footing against Google in search and ads. They had nothing in social media. They were far behind AWS in cloud. They weren’t even part of the FA(A)NG acronym.
Then Steve Ballmer (who got way too much hate for his performance in my opinion) resigned and Nadella took over. He launched the cloud computing success story which is the core of what Microsoft is today.
Their biggest coup was probably the early partnership with OpenAI which ensured that Azure became the primary engine for ChatGPT. OpenAI is by far the most aggressive AI spender and Microsoft directly benefits from that. Azure’s deep learning revenue will likely be $20bn this year, about as much as Google and AWS combined.
Microsoft’s operating profits are surging as a result. But looking at earnings only tells us how lucrative past investments have been. FCF offers insight into what current investments have to earn to be worth it. On an LTM basis, their FCF margin is currently about 20-25%, significantly lower than where it was before ChatGPT was launched.
And just like they do for Google, markets consider that a temporary dip for Microsoft as well. Their LTM FCF multiple has risen to almost 60x.
Again, is there an appropriate risk premium to be earned here? Hasn’t the risk profile of Microsoft’s cash flows increased? Aren’t they today more dependent on a single big client than they were before OpenAI entered the stage? What is the price of the Sam Altman key man risk for Microsoft?
Game theory in action
I discussed the game theoretical aspects of Big Tech decision making in more detail in the article on the economics of the AI industry linked above. Big Tech executives are strongly incentivized to spend, irrespective of where the RoI is and how much certainty they have about it.
Instead of repeating my points here, I want to make an additional one. The most fascinating aspect of the current AI theme is how it suggests that Big Tech doesn’t function like you think an oligopoly would do. These companies are actually not colluding at all.
Yes, they are entering into partnerships with each other which links their fate to some extent. But first and foremost they are fiercely competing with each other, willingly destroying a ton of shareholder capital in the process. Whoever loses this battle will pay a huge price. And the winners will likely take a hit, too.
If these companies really were powerful oligopolists, wouldn’t they be much more moderate? Take more time to make sure every investment decision is worth it? Wouldn’t they use their combined power to negotiate a better deal with NVIDIA? Instead they are a) being chased by the Chinese boogeyman and b) completely surrender to Jensen Huang. Paying him whatever price he wants per GPU.
These are supposedly the strongest companies in the world. This strength grants them their valuation premium. But one meaningful innovation and they completely lose their composure. So, how strong can they really be?
Sincerely,
Rene