Cocktail Mixers, OpenAI, and Make vs. Buy Decisions
“You can get pre-mixed Manhattans and other cocktails at Friar Tucks- the liquor store on Watson.”
WHAT?!
My initial reaction was thrilled, but then I remembered.
I kinda like the process of making my own cocktails. I passed on buying the pre-mixed version (although I hear that they’re just as good as homemade).
This article is about make vs. buy decisions, and how OpenAI has a huge make vs. buy decision on its hands.
Understanding Make vs. Buy Decisions
Here’s the common definition of a make vs. buy decision from Investopedia:
“A make-or-buy decision is an act of choosing between manufacturing a product in-house or purchasing it from an external supplier.”
Say, for example, that you operate a furniture manufacture. One of your products is a dining room table that includes small drawers, and each drawer has a metal handle that is screwed into the wood.
You purchase the drawer handles from a supplier, and prices have been increased by 15% over the last 18 months. Also, you’ve had an increasing number of handles that don’t meet your manufacturing standards, and they have to be returned and replaced. Replacing handles doesn’t cost you anything, but slows down table production.
Rather than deal with higher prices and quality issues, you might decide to make the metal handles yourself.
But is that the right decision?
What is the Right Decision?
Investopedia also points out the factors to consider before you stop buying a product and start to make it:
“There are many factors at play that may tilt a company from making an item in-house or outsourcing it, such as labor costs, lack of expertise, storage costs, supplier contracts, and lack of sufficient volume.”
OpenAI has a trillion-dollar make vs. buy decision.
The Trillion Dollar Question
OpenAI relies on computer chips- a unique type of computer chip.
As the Wall Street Journal explains:
There’s a potential “shortage of AI chips needed to produce systems such as OpenAI’s ChatGPT. The biggest bottleneck in the production of AI chips has been in packaging, a manufacturing step that comes after the circuits are imprinted on silicon.”
So who makes the chips?
“There are now only three companies in the world capable of making the most cutting-edge chips—including the processors used to power AI systems—in large volumes: Taiwan Semiconductor, Samsung and Intel.”
Only three.
Well, that sounds like a real problem. Just a few suppliers means higher prices and a bigger risk of chip shortages.
Why not start making chips?
OpenAI CEO Sam Altman “has held discussions with chip makers about joining with them and using trillions of dollars to build and operate new factories, along with investments in energy and other AI infrastructure. Many of the world’s largest chip companies, including Nvidia, design their chips but outsource their production to companies such as Taiwan Semiconductor.”
Nvidia is one of the best performing stocks in recent years- an increase of 250% over the last 12 months.
But they don’t manufacture computer chips. They only design them.
OK, so what’s the cost to make your own chips?
“Building a cutting-edge chip factory typically costs at least $10 billion. But even with that, the scale Altman is discussing is extreme: Stacy Rasgon, an analyst at Bernstein Research, estimates that a little more than $1 trillion has been spent on chip-manufacturing equipment in the entire history of the industry.”
Here are some other factors:
Revenue potential: If you’re gonna spend $10 billion on a factory, do you have the revenue to justify the cost? Most AI analysts think so.
Timeline: How long will it take to get up and running?
Hiring enough experts: “There are uncertainties about finding the engineers to operate a rash of new factories.”
Every business that is making a make vs. buy decision must deal with these same factors.
The Lesson
The most important stakeholder in your business may not be employees or customers.
It may be suppliers.
If suppliers can’t make a quality product at a reasonable price and send the product on time, you may be out of business. You can’t make furniture without handles, and you can’t build AI without the right computer chips.
Think carefully before cancelling a supplier and making a component part in-house.