The catalysts for a crash are already laid out, and it can happen sooner than most expect. AI is here to stay. If used right, chances are it will make us all more productive. That, on the other hand, does not mean it will be a good investment.

Big tech doesn’t need to win, just outspend

Magnificent 7 companies are increasing capex to their biggest ever to differentiate their tech from each other and the big AI labs, but the key realization is that they don’t have to spend it to win. It’s a defensive move for them, if they commit $50B, OpenAI and Anthropic need to go raise $100B each to stay competitive, which makes them reliant on investors’ money. As the numbers get bigger, the amount of funds that can write checks of the size required to fill such amounts gets smaller. And many of them are now getting bombed in the Gulf.

This is the reason there’s a push for IPOs, it’s because it’s the only option left to keep the funding coming.

Taking this into account, Google is extremely well positioned to weather the storm. When they announce capex expenditure, they don’t spend it overnight. They can simply deploy month by month until their competitors struggle to raise and get forced to capitulate. At that point they can just ramp down the spending and declare victory in a cornered market. They don’t need capex, they just need to make it very clear for everyone that nobody can outspend them. It is hard to picture as numbers get so big, but Alphabet (Google’s parent) is ten times more valuable than the biggest military company ^.

This also has a great implication for the Mag 7, especially Google: their capex will be a lot smaller in practice than projected, and as investors hate to see high capex in tech, the market will probably reward that if it materializes.

Apple didn’t even have to pretend, their strategy of waiting on the sidelines, while selling Mac Minis, for someone to come up with a good-enough model and just buy that when it’s done seems to be working. They may not even do that, they are now hinting at charging models for being available on Siri. Amazon is hedged with an Anthropic investment, and Meta is spending like there’s no tomorrow.

The catalyzer

We’re hitting the worst-case scenarios for the big AI labs: energy, their biggest expense, is at multi-year highs, capital from the Gulf is not available for obvious reasons, there are serious concerns about a rate hike, and RAM prices are crashing because new models won’t need as much, but labs already bought them at sky-high prices. And that last innovation came from their biggest competitor, Google.

Anthropic is already in a push to reduce costs and increase revenue. If investor money dries up, they will be forced to cut their losses and pass the true costs to their users. The question is now if customers will be willing to pay up. Independent reports state that Claude metered models are priced 5x more expensive than their subscribers pay^{, and nobody is sure if even their metered pricing is profitable. In investing, stories are way more exciting than reality: a company losing money but growing like crazy is an easier sell than a huge company losing money or with tight margins. Raising prices will for sure decrease demand and that risks killing the growth story. And even if revenue keeps growing, it doesn’t matter if there are no margins — growing revenue without profits just means burning cash faster, especially when competing against companies that can offer the same product as a loss leader bundled into their cloud platforms.}

It’s also worth mentioning that Claude’s most expensive subscription plans (Max and Max 5x, priced at $100 and $200 respectively) do not allow for yearly payments, hinting prices will go up^.

OpenAI’s endgame

OpenAI is struggling to monetize. They turned to showing ads in ChatGPT, something Sam Altman once called a “last resort”, while Anthropic is crushing them with the more profitable corporate customers and software engineers. Their shopping feature flopped and they shut down Sora, both supposed to be revenue drivers.

I wouldn’t be surprised at all if in the next couple of quarters we see OpenAI looking for an exit. It will be interesting because the sizes are now so big that we will probably know all the details. The most likely buyer is Microsoft, they already own a lot of it, and because of that, they are the most interested in showing a win. Sam Altman managed to get Microsoft so involved in OpenAI that making sure it lands on its feet is a Microsoft problem to solve. But, would shareholders vote to spend 22% of an established company’s market cap to rescue a money-burning AI lab that has lost most of its differentiators?

And independent of whether Microsoft makes money or not in their OpenAI endeavor, it kills the story: they were betting the whole growth story on AI, and if that doesn’t work out, then what’s left to justify a high stock price? They lose a big customer for their cloud services. Even worse considering that now, using the AI they helped fund, everyone can compete with their sub-par products. GitHub is a good candidate for disruption, and that’d be just the start.

How this all affects you

You may think that you’re not affected by the big labs struggling. Hell, you may even be happy that they won’t be replacing your job after all. But that is far from reality.

Investments are now so big that writing them off would certainly hurt public companies’ balance sheets, and their growth prospects. This will drag the whole market, reducing valuations and slowing M&A, which further dries up VC money and slows down investments. Just like it happened in 2022.

And this has even more ramifications, pension funds around the world will take a hit. Datacenters that were built with the expectation of growth will now be undercapacity, because as training is the most compute-intensive part of a model, if there’s no capital to train a new one, they won’t be needed. GPUs then sit idle while their value goes down as there’s no demand. Some committed GPUs may never get delivered, or even manufactured. Investment drying up is a disaster for Nvidia, now the biggest company in the world.

It could happen that datacenters are not underused, but they get to charge their customers a way lower rate than they projected before building, so everyone benefits from AI but them.

Building a datacenter is supposed to be a “safe” investment in normal times, so banks give private credit and mortgages to finance them. A write-off of those assets means that banks start realizing losses, hurting their capacity to loan, and some may even be forced to liquidate, just like we saw in 2023. And all this assumes we don’t get disruptions in manufacturing in Taiwan or global supply chains.

Of course, the content of this article is highly speculative, it may end up being that demand for models is just so high it offsets every other problem I lay. But almost all innovations go through a boom and bust cycle and I don’t see a reason this is an exception.

Thanks to Javier Silveira and Augusto Gesualdi for reviewing drafts of this post.

Enjoyed this post?

Subscribe to get new posts by email.