Nvidia CEO Jen-Hsun Huang views AI as a commodity that can be produced in 'AI factories.'

Even though I've studied philosophy and have a lot of experience with it, I was still surprised to hear Jen-Hsun Huang, Nvidia's CEO, give her latest take on AI. This is because AI is produced just like electricity. It's therefore a commodity or resource.

In a recent earnings report (via Motley Fool), the Huangster said: "Just as we generate electricity, now we'll be generating AI." These generators will run 24/7 if there are a lot of customers, just like the electricity consumers.

Huang is adamant about the idea, even if it sounds like marketing gimmick. He continued, "Today, AI services run 24/7, just as an AI factory." We're going see this new system online. I call it an AI Factory because that's as close to what it really is.

It's okay, but maybeit is just hyperbole.

"When we talk about generative AI, what we mean is that these data centres are actually AI factories. They're generating some thing."

Okay, it seems like a serious claim. Let's give him his due.

The knee-jerk response is to say that AI is not comparable to electricity as a commodity. This reaction is best encapsulated in the absurd image of topping up your AI card in the store, then popping it into the AI meter at your home to bring your AI online.

But that's really not fair, isn't it? It's a bit caricatured, but it does make some sence. "Resource" is a term that can refer to money, materials, capital and even human (you know, "human resources"). "Commodity", on the other hand, can be anything that is traded, bought or sold.

In this sense, almost anything can be considered a commodity or resource. The scandalous huckster inside me wants to agree. "There's always a price", I say as I smile to reveal the glimmer of my golden tooth.

The proverbial Huangian may claim that resources are no less ephemeral than money. They can be debated. In a hypothetical "Twigland", where people are insatiably hungry for twigs and trees, especially those with twigglesome branches, would be the number one resource. It all comes down what we value.

Huang may have a good idea of what we value in AI. Huang says we're witnessing the beginning of "widespread shift" that will see us "moving away from coding which runs on CPUs and creating neural networks that run on GPUs". He also states that "there is no company that won't do machine learning".

In this case, AI computing could become soubiquitous it would be considered twigs by Twiglanders. This conclusion would be logical if CPU processing is really second to machine learning in neural network processing.

Would it still be a like electricity?

Where do we draw the lines if we say yes? There seems to be no meaningful difference between a "resource", like electricity, and anything else that we value. Resources must have some weight behind them, i.e., they have to be far enough down the chain to be used for multiple things, such as electricity powering my computer and traffic lights.

Huang might say that this is exactly the point. AI will be used for many vitally important things, including medical treatments, home computing and research.

Maybe the only way to win this argument is to cut off power to the AI data centers. Mr. Huang, we'll see if AI can compete with electricity.