I've spent my whole career in software engineering. But lately I've been thinking that software is cheap and I should focus more on the business side of things.

Why is software cheap? Well, apparently there's this thing called supply and demand. Because there's a huge supply of software engineers -- somewhere between 20 and 30 million -- companies don't have to pay very much for an engineer's time.

Clearly if I'm going to sell something, I should focus on an area where supply is low and demand is high. How can I identify such things? Well, they would have a high price. Ok, so what things have a high price? What's the most expensive thing I could buy?

• I could buy a houseplant for $20.
• I could buy a gallon of Ethylene for $48.
• I could buy a Macbook Pro for $2800.

Hmm...

• I could buy a Tesla for $80,000.
• I could buy a Ferrari for $360,000.

(well, I mean I couldn't, but a rich person could...)

• I could buy a house in San Francisco for $1,000,000.

• I could buy Boeing 737 for $30,000,000

• I could buy a rocketship for $60,000,000

Now we're talking...

• The new World Trade Center cost $3.8 billion to build. I wonder how much it would cost to buy?

• The International Space Station cost $150 billion to build. I assume one could offer each of the respective countries $100 billion for their "share" of the station and end up owning the whole thing?

What costs more than $150 billion?

Companies.

• A company called Vodafone bought another company called Mannesmann for $300 billion.

I think large corporations are the most expensive thing one could buy.

So the most expensive thing a person could buy as of November 2019 would be Microsoft, Apple, or Amazon, and they cost about $1 trillion each.

Ok, so giant companies that generate huge profits are in very short supply and there is very high demand for them. Hmm, maybe those YCombinator folks are on to something...

But wait, price isn't the only factor. You also need to know the quantity demanded. Apple sells 200 million iPhones a year for like $500 each. That comes out to $100 billion in annual revenue.

Hmm... what about actual apples? $4 billion worth of apples sold per year...

Actually, if we consider companies as products, they win here too: there were $3.9 trillion worth of companies purchased in 2018. So I guess that's a second vote for building a company that builds companies being a very good idea.

Is there any market bigger than $3.9 trillion?

Can I somehow just get a list of every possible market, and then sort it by size in descending order? Well it turns out you can -- it's called GDP per Industry and it's reported regularly by various government organizations like the US Bureau of Economic Analysis. They provide both current and historical data; that means you can consider not only the dollar size of each sector, but also the growth trends.