Bryne Hobart asks "How many trillion-dollar companies should there be?" Market cap is functionally humanity's best guess at expected total dividends returned to shareholders. A company's profits come from its revenue, and its revenue comes from how much entities (probably mostly consumers) pay for its services. A trillion dollar company is one that pays out at least a trillion dollars to its shareholders, so its revenues must be over a trillion dollars, so it must be providing more than a trillion dollars of value to consumers.
There are 6 trillion dollar companies today: Apple, Microsoft, Saudi Aramco, Alphabet, Amazon, and Nvidia. Microsoft currently pays out $0.75 per share per quarter. It has 7.5 billion outstanding shares, so that's about $20B in dividends per year. It's worth $3T so humanity expects 150 years of current earnings to persist.
Value provided to consumers isn't all that matters though! The light bulb and electricity provided much more value than Microsoft did, but its inventors didn't become quadrillionaires. Capturing some portion of that value is critical!
Bryne says you need some sort of runaway network effect. He summarizes it as: when the thing you figured out first eventually becomes obvious to everyone else, have you set up a system that ensures that you're still accumulating new and valuable secrets faster than anyone else?
- Facebook and Apple leverage the in-network effect (if your friends are on it, you won't try anything else).
- Tiktok also does this but it's less social. They use user data to make the app much better for the user and marginally better for everyone else. Making it better for everyone else means you don't try other apps
- Amazon does this with improving its deliverability speed to crazy levels with insights on user location
- I don't think AWS and Azure have unstoppable network effects? They just have economies of scale.
- Microsoft Office (Microsoft's biggest revenue stream) just uses proprietary software network effects. I don't think this has unstoppable network effects?
- Saudi Arabia doesn't have network effects. They just lucked into one of the largest crude oil reserves happening to exist under their ancestral lands, so they win.
This is a form of economic rent. These firms have found a "natural resource" (some tool that's amenable to unstoppable network effects) and charge as much as they can for it. Their profits can't be competed away, so they get to capture lots of the value they create. It's a rent in the Georgian sense: owners extracting more value out of a scarce resource than they "deserve" based on the resources expended in creating/improving it. (Thanks to Lars Doucet and Scott Alexander for making the connection to Georgism.)
So I wonder: is this worth thinking about before eg starting a company? Is being satisfied with producing lots of value for society foolish because you're missing opportunities to charge rent?
One approach is to embrace the inevitability of your profits being competed away. Suppose you think ultrasound neuromodulation devices should be much better than they currently are. You can go build the first really good product, charge a lot for it until competitors crop up to eat away at your margins, and then leave satisfied. You wanted, after all, better devices and now there's a healthy competition pushing the state-of-the-art!
Well, but you are still leaving money on the table by not finding a way to monopolize your advantage. Suppose there's an innovation in neuromodulation that costs $100B to realize. If your profits are eaten up before you can invest $100B, this innovation will never be realized! The system is broken and doesn't regulate rents, so it feels irrational to just ignore that inefficiency...
Vamshi pointed out that the greatest inventions don't seem to come about from immense investment by a few firms; they're usually the result of some genius working for very little because he can't help himself. Matt Clancy points out that rediscovery is likely if the idea seems important looking forward; it's cause areas and ideas that are only important in retrospect that aren't likely to be independenly rediscovered. That's good news since even the most ambitious companies probably can't justify entirely pie-in-the-sky research.
Also, is the economy that competitive? We only have ten billion people who can compete, and probably less than half of them have direct access to the global economy. There are all sorts of inefficiencies that wouldn't exist in a world with 100 trillion people.
Thanks Anthony Russo and Vamshi Balanaga for conversations on this.