Profits
To Dad… In 1980 I was a child, I talked like a child, I wondered how GM could sell lots of cars and lose lots of money. As a man I'm gob smacked how big companies lose big money. Corinthian MBAs 13:11
Despite the godawful body of the “It works” Chevy Citation, the crushed burgundy velour that a 2000-era mall-walker wouldn’t don as a track suit that overcovered the optional front bench seats of the Olds 88 like the excess skin on an aging shar-pei, and John DeLorean’s Edsel-level bad Chevy Vega, GM sold $58B worth of cars in 1980. In this final year as CEO, Thomas Aquinas Murphy yep – his real name – said, “General Motors is not in the business of making cars. It is in the business of making money.” Um. Then-newbie CEO Roger Smith audaciously blamed the$763M loss on shifting consumer demand and inflation. The word you’re looking for is chutzpah. With extra spittle. Smith was GM’s first MBA-trained CEO. Yep, the truck and bus guy.
We’ve come a long way.
Executives don’t need to blame things when they lose gobs of money.
Check it. Sam Altman runs OpenAI. One hundred million people use the company’s ChatGPT product every week. This year, OpenAI will sell $5B of their smart service. They’ll lose $4B. No remorse. No excuses. No explanation. No chutzpah. Just loses. Ironically, he is converting the money-losing not-for-profit into a for-profit that will lose money for the foreseeable future.
But, OpenAI isn’t Amazon. Famously, Amazon lost gobs of money on every sale for its first couple of decades of operation. And, pretty much, they still do. But, they got so big, they sell the services that keep them running to other companies that lose money. Like investing $4B of in-kind services to OpenAI competitor, Anthropic. File this under “S” for “Stuff you can’t make up.”
This year, OpenAI will spend $3B to train its AI, $2B on computers to run its models, $400M to host the computers, and $500M on data. Smartly, they’ll spend $1B to research how to lower those costs. Somehow that research doesn’t include $700M on employee salaries or the $600M to pay executives and their presumably lavish perqs.
Maybe Sam doesn’t know he’ll run out of money before he changes humanity. Or, maybe, he’s banking his company’s future bank on Moore’s Law that compute costs will go down. And, that he won’t need to buy content. Maybe our demand for his AI services will shift to simpler fare so ChatGPT can serve up high-margin AI leftovers to inflate their profits. FWIW, Sam doesn’t have an MBA. Or, maybe, his child-like thinking is just what big companies need.