Using artificial intelligence requires a lot of power. Like, a *lot* of power. Like, “the reason your energy bill is now 200 percent what it was is because A.I. companies are using the power grid to generate videos of Garfield making out with Odie”-levels of power.

However, these A.I. companies don’t just need electricity to be effective. They need data centers, which pretty much all of them have said they want to spread around the United States. These data centers require time, resources and, most importantly, money.

So, if OpenAI were to fulfill all of its current promises, how much would it cost? Tech researcher Ed Zitron has crunched the numbers and, according to him, it would cost… a trillion dollars. Over the next four years.

For context, that’s the GDP of Saudi Arabia.

I know what you’re thinking: That sounds like a lot of money! Zitron notes that it is, and that the company is pretty unlikely to get it — especially because OpenAI isn’t turning a profit despite bringing in $30 billion in revenue last year.

Zitron goes on to reiterate that, for all of this money, energy usage and wasted water and space, we’re really not getting a great product out of it. They can generate text and images, sure, but they still can’t do basic things like “tell you the number of R’s in the word ‘strawberry.’”

So, what do you think? Is it worth it to invest a trillion dollars into a software that can’t summarize a paper without making things up? Or should that money go to literally anything else? As someone who just paid my health insurance bill, I would tend toward the latter argument.