Downside at Anthropic

Margin of Hype

Share
Downside at Anthropic

Anthropic (Ticker: ANT) is finally expected to list on November 19th, on the NASDAQ.

After Anthropic's recent troubles with slowing growth and a reducing (private market) valuation – many thought the IPO would have to be suspended. We continued to believe that no matter their issues, the IPO would eventually go through.

Since Anthropic's Investor Day and its S-1 filing on October 26th: much has been said about the business, the business model, its accounting practices, the future of AI and compute demand, Anthropic's relationships with its business partners and competition from Chinese/open-weight models etc.

Technologists Vs Generalists

The two main camps seem to be the technologists on the one, characterised by an infinite love for AI and technology – but also an infinite bullishness.

And on the other side the generalists – or the AI skeptics. Those who don't buy the hype, or the idea that any new technology or innovation will result in massive profits and ROI for years to come. Note: it usually doesn't.

The technologists made so much money the past few years that they are emboldened by their P&Ls and the pace of development in AI. They are usually young(er) and don't have as much actual investing experience.

The generalists aren't really intimidated by the other guy's P&L and can survive any environment. They have decades of investing experience and they've seen it all. They survived bull and bear markets, including dot com. Their battle scars are deep and they refuse to drink the Kool-AId being passed around.

On the market participant spectrum, those two groups lie on opposite ends.

We first touched on this bifurcation in market perspectives in Flash Philo on Memory, with the meme shown below.

Technologist poster boi Leopold blew up a month later. Since then many others did too, with more to come.

Downside by PhiloπŸ¦‰

It's been some time since we've published a Downside by Philo issue. So let's repeat what the main objective of this Philoinvestor section is.

These are not necessarily short candidates. Just because I write a critical piece about a company does not mean I go out and short it at the same time.

Rather, I write a Downside piece when I have strong critical views about a company that the market misses or completely ignores.

They could or could not play out, either imminently, or much later down the line. But it's important that you consider them when you invest in said company.

Remember: preparation is key.

Downside at Anthropic

In this Downside issue we will take a wise investor approach to investing in Anthropic shares.

What would Charlie Munger say about investing in this business?

I have a list of angles to touch on, and I expect this will be a long letter – but to the point. I also suspect some of the angles we will write about won't be covered anywhere else.

Change is a Constant

Warren Buffet advised in investing in companies that benefit from slow change, not the opposite.

The reason is: to make long-term predictions about a company you need a margin of safety, not just in price, but in everything else too.

In other words, the business needs to be somewhat predictable.

In the case of Anthropic and AI Labs – they find themselves on the far right of the business predictability spectrum. That is to say, their future is largely unpredictable.

The business predictability spectrum

As Ancient Greek philosopher Heraclitus once said, ta panta rei, which translates to everything flows. But time is moving much faster nowadays, and interconnectedness complicates that even further.

And that's why Modern Geek Philosopher Philo said:

"The world is getting more and more fat tailed."

With how dynamic user patterns are, and with virtually zero switching costs – the wrecking ball of users can swing in and out of your favour. From OpenAI to Anthropic, back to OpenAI, or even anywhere else.

This lack of moat makes it impossible to predict where we will be 6 months down the line – not to mention 3 years.

And no, just because Claude Code is cool and works wonders, and you have your agents running on it does not mean Anthropic has a solid moat that's measurable enough to consider.