You’ve been in London for two years now. Do you feel part of Europe’s tech scene?

I’m an outsider, I still don’t really understand it. The UK is really strong on fintech but on the whole it’s been underwhelming given how strong the educational system is, how talented young people are, how many people would like to move here and invest here. 

There needs to be a pretty staggering, unified investment from governments to inspire investment.

You charted how the ‘four horsemen’ built their monopolies, and the same dynamic pervades AI infrastructure. The scale and speed of adoption has forced total reliance on those who were there first and had that runway.

Can governments protect a competitive AI market, or have the big companies already created structures to evade accountability?

It’s mostly the latter. I think things are getting better, the DOJ and the FTC have finally woken up. But what we call OpenAI should be called Microsoft AI. Microsoft controls OpenAI and Perplexity and Anthropic have big investments from those companies. It’s frightening how concentrated this industry is and how the majority of the games are accreting to the same players. We’re seeing an even greater concentration in power, which is terrible for the economy and society writ large.

We need much more competition. These companies need to be broken up.

What would the journey look like to get to that point?

We’re seeing the beginnings of antitrust activity: the DOJ and the FTC suits against Google to break them up, the EU suit against Apple to regulate the app store. The issue is it’s very hard to manage or mandate a breakup outside of the country where it’s headquartered.

People think of regulation as stifling; actually it oxygenates the marketplace and forces standards that make the rules easier for new entrants to understand.

The EU has led innovation on regulation. GDPR wasn’t elegant but it was a start, it was a place to begin and start improving from. If you look at the Digital Services Act, the EU has actually been pretty innovative around regulation. At least they’ve had kind of the backbone to start and recognise that this is the only industry in history of this size or importance that has almost zero regulation.

I guess, but Microsoft was being sued by the DOJ in 2001 — the year I was born — and it hasn’t worked.

What’s interesting is: that case to break up Microsoft, yes it was overturned but if it hadn’t happened, we’d all be saying “Bing it”. The reason that Microsoft stopped bundling and demanding that everyone on Microsoft Office used their search was because of DOJ scrutiny and that’s how Google emerged. It’s not actual breakups that create a lot of the value, it’s scrutiny. 

Look at AT&T, it was broken into seven companies and all seven ended up being worth more than the original AT&T within about 11 years. I’m an Amazon shareholder and I hope that Amazon gets broken up because I think I’ll make more money. A lot of these companies would be worth more if they traded on their own than stuck into a conglomerate structure.

Back in 2017, you were warning that business models powered by ad revenue were on borrowed time. Where will market trends take the information economy next? Have you seen any interesting new models for media?

One thing I like about AI is it’s more of a subscription model. The original sin of the internet was that it was ad supported, so it became all about attention which then leads to misinformation. Novelty’s more interesting than truth, it’s interesting to report that mRNA vaccines alter your DNA even though it’s not true.

From 1945 to 1995, it used to be that sex sells, and then the algorithms figured out there’s something that sells better than sex and that’s rage. So they tried to incite violence and a coarsing of our discourse. The ad model has been a disaster for our society.

If GenAI models run on a string of algorithms, won’t they have the same kind of driver?

To a certain extent, if it’s going to crawl online data it’ll pick up on that coarseness. But the subscription model is more about how do I add value as opposed to how do I keep your attention. I think it’s healthier to say, all right, pay $10 a month and then I’ll focus on getting you the right answer as opposed to doing anything just to keep you on the screen.

Your tech and business predictions are renowned. How do you make them?

I put a lot of data behind it, I try to be fearless and to turn it into an interesting story. The key is to inspire a conversation that might result in better solutions. But I have to say that the secret sauce is to have a great team that pulls together great data and puts it out in a compelling way.

The Algebra of Wealth by Scott Galloway is published by Torva.