This is a follow-up to my AI Eats the World post from November. Roughly half a year later, AI is still the dominant topic of the day, and the new models from Anthropic and OpenAI are another major jump. This is part 1, covering the newest slide deck from Ben Evans and some more interesting observations. Part 2 will move beyond the data, and I will discuss my personal observations on the current AI landscape.
You will find my newest article here:
Some companies are being perceived as AI losers/victims, and their stock prices have overreacted. This is a great time to look closer. Learn about why I believe Adobe, Booking, and Salesforce are particularly interesting here:
AI is here to stay for sure. The question is whether the current hype is going too far and what kind of impact AI will have in the long run. Learn more about AI and its impact in my previous articles that you will find here:
Ben Evans updated his slide deck, and the following slides are taken from his excellent website that you will find here: https://www.ben-evans.com/presentations
The CapEx explosion keeps getting larger. Gone are the old days of the asset-light software companies like former Google and Meta.
Meta increased its capital expenditure by more than 10x since 2017. I guess nobody could have imagined the former Facebook to invest $75 billion in CapEx in a single year.
Some companies are having the time of their life, and I am happy that I have invested in these and also presented them to you. I hope you are part of the journey. Find my deep dives on AMD and Arista Networks here:
If you want to go beyond reading and actually invest alongside me, you can invest in my Wikifolio. Every position reflects the same research-driven, long-term investing approach behind these deep dives — including investments such as Arista Networks, Alphabet, ASML, AMD, and now Booking. You will find it here:
Semiconductor demand is off the charts. I am sure that there will still be ups and downs in the demand, but it will be from a much larger base in the future. The bottlenecks of this segment, TSMC and ASML, are living in golden times.
Data center construction spending overtook office construction. BEFORE including compute. That somewhat fits nicely into the narrative that AI will replace white color workers. It is already replacing the location of their work.
The bottleneck is not just GPUs anymore. Memory like DRAM is in high demand, as evidenced by the large stock price increases for Micron and SK Hynix.
Energy to power these chips is in high demand and will be a major challenge going forward. Political and community backlash (more on that later) is an underrated part of the equation in my opinion.
I was about to include a picture of a pig in python but decided against it. Look it up for yourself if you are interested. This is also a nice test case if search, as in Google Search, is dead: Would you Google that or ask ChatGPT and co? Note that I used the word Google as a synonym for internet search.
Nevertheless, revenue growth of the two leading standalone AI companies, OpenAI (ChatGPT) and Anthropic (Claude), is insane. This is the hockey stick from the business books in real life. A gigantic hockey stick.
But if the models become a commodity eventually, then you are not getting rich by investing in any of these companies. Remember the chart from Meta’s CapEx earlier on? This is the CapEx from Apple. Apple sits back, relaxes, and enjoys the AI race, and will then choose the AI of their liking for their services.
Global telecommunication stocks went almost nowhere for 15 years, even though global mobile data traffic exploded. The drop in pricing was just too strong.
I remember the old days when mobile internet just came along. As a teenager, you were afraid to go bankrupt if you accidentally turned on mobile data on your Sony Ericsson. Sony Ericsson is also a thing of the past, with the last phone launched in 2011.
That also reminds me of this Warren Buffett take in his 2007 annual letter. I am not saying that AI spend will be the same, but keep that in the back of your mind.
The worst sort of business is one that grows rapidly, requires significant capital to engender the growth, and then earns little or no money. Think airlines. Here a durable competitive advantage has proven elusive ever since the days of the Wright Brothers. Indeed, if a farsighted capitalist had been present at Kitty Hawk, he would have done his successors a huge favor by shooting Orville down. The airline industry’s demand for capital ever since that first flight has been insatiable. Investors have poured money into a bottomless pit, attracted by growth when they should have been repelled by it.
This is an interesting take and fits with what some companies have openly said: AI will eventually be a commodity - all the models are trained on the same data after all- and companies can just use the AI that fits best. Therefore rendering any competitive advantage irrelevant and causing pricing pressure.
Always keep in mind that we are still early. Even though it sometimes feels that AI has entered everyone’s life, half of the population between 14-29 does not use chatbots every week. If you are thinking about the implications of AI, you are far ahead.
So where is it actually being used in a business setting? The answer is coding. The adage “study IT or programming and you will always have a safe job” is not true anymore. In a somewhat funny plot twist, the first jobs to be replaced by AI were those that built it. Legal is an interesting one, since the legal workforce has been a relatively slow one to digitize so far.
A lot of companies are still struggling with the question: Where and how do I deploy AI? Here come the consultants who discovered this as an amazing new vertical. At the same time, they suffer from lost contracts because routine tasks can be automated by AI in the company itself.
Another important aspect: The tools are working; getting them to find the right problem is harder. That is why AI companies need help with deploying their solutions and tackling the right issues.
I love how Ben Evans phrased this paradigm change. There are new options to build solutions that were impossible before.
AI influences how we will shop significantly. AI and social media will understand us and our decisions a lot better going forward. It is questionable if that is always a good thing for the average private person.
This is what Meta CEO Mark Zuckerberg said recently
We're also working on using Spark and our upcoming models to improve our recommendation systems and core business in Facebook, Instagram, and ads. Right now, our apps primarily help people accomplish three important goals: connecting with people, learning about the world, and entertainment. But we've always wanted our apps to understand more of people's goals so we can help improve their lives in all the ways that they want. These new AI models will let us understand this in more detail. Instead of just looking at statistical patterns of what types of people engage with what content, for the first time in Meta's history we're going to be able to develop a first-principles understanding of what you care about and what each piece of content in our system is about so we can show you more useful things for what you're trying to accomplish. We'll also be able to create personalized content specifically for people to help achieve your goals as well.
This is another great chart. The music industry was supposed to be dying in the early 2010s until streaming came along. Streaming did not return the music industry to its old glory of the 2000s, but it helped to recreate a large amount of lost revenues. For the user, it changed the way we listen and discover songs completely. Two more interesting observations: Downloading music was an interesting new invention that is fading now, while physical sales are surprisingly stable. When was the last time that you bought a physical copy?
Apply this framework also for yourself. What can you change or do differently than in the past? Are there topics that you can fully automate or improve by using AI?
As always, buy great companies and enjoy the ride.
41
Invest at your own risk; this is not financial advice! This is not a recommendation to buy or sell any securities discussed in the article.
Stocks in this article:
AMD -2.49%↓ ANET -1.53%↓ MSFT -1.60%↓ NOW 1.38%↑ AMZN 1.87%↑ SNOW -4.67%↓ CRM 1.38%↑ NVDA -1.28%↓ META -0.84%↓ CSCO -0.43%↓