This is the first of two posts answering questions in the truncated 2025 “Ask Arthur” series. This year, I forgot to promote the series, and so, there will be only two posts. But… at least there are two posts? Yeah, that’s it.
Most years' first question is usually from my pal Roger Green, and that’s true this year, too. In fact, he asked the final question for the final post, too. Roger’s question is in two parts. First, he asked:
Are we going to have an AI bubble? You may have seen this graphic where Nvidia is in this loop with the companies they are selling chips to; it looks unsustainable to me because things don't always get bigger and bigger.
Short answer: Yes, it’s a bubble, and it will eventually burst. Okay, that’s that, then! Just kidding: It’s obviously more complicated than that.
Nearly everything in the tech world is built with hype as much as actual products or services, and it’s been that way for decades. Also, there’s no such thing as a market that can’t crash, and so, it’s entirely possible it will happen to AI (speaking in general terms).
There are many tech and investment commentators who have their own reasons for warning of a bubble for the AI industry, and they’re quite persuasive, however, I’ve never found the “tech bros” particularly—what’s the right word? Let’s say persuasive, served with a side dish of scepticism in incredulousness sauce. Most of the tech bros are showmen, and it’s in their insterests to, shall we say, oversell what they’re offerening. That’s true in a lot of industries, of course.
While I haven’t seen the specific graphic Roger was referring to, I think his scepticism is exactly right, and for me, it comes down to one thing: Where’s the income? We keep hearing how the market capitalisation of Nvidia is reaching ever higher unprecedented levels, and that it’s because of all the cards they’re selling to AI companies. AI companies may offer paid accounts to users, and other companies are trying to force their customers/users into using AI, companies like Microsoft, Apple, Google (Chrome in particular, but also for search), among others. So: Where’s all the profits? Not one financial commentator has reported how the tech companies are making massive profits from all their expenditures on AI. At the same time, there’s noticeable reistance from consumers who resent having AI forced on them, particularly when there are still no regulations or saftey measures.
At the same time, there are places where AI could change everything, like medical research, for example, because of its ability to evaluate, say, medicines much faster than humans ever could. However, without safeguards or regulations, we’ll see more “AI slop” foisted on us, and deep fakes used to destroy people or promoote propaganda. Ordinary people are increasingly worried about all that—and they couldn’t possibly care less about corporate profits or the salesmanship of the tech bros.
Add it all up, and there are plenty of reasons to think the AI bubble could burst, and also reasons to think is definitely will.
In the second part of Roger’s question, he asks:
AI is an incredible energy suck; is it worth it, or will this end up hurting consumers?
It’s already hurting consumers. Part of the reason that electricity proces are so high in the USA is the skyrockeing demand for electricity from AI data centres—though the situation would be dramatically better if the current US regime hadn’t made the idiotic move to ban new renewable energy projects, particularly because it’s far cheaper and faster to build clean, renewable generation than to build fossil fuel power plants. This represents an existential threat to the AI industry, particularly as consumer anger grows at having to pay high power bills because of AI data centres.
Also, the data centres often have high demands for water at a time in which climate change is making droughts more frequent. If all that wasn’t enough, the centres usually offer little employment in areas where they’re located—not counting the short-term construction jobs, of course.
Maybe the tech bros should use AI to solve some of the problems they’re creating. If they don’t pay attention to the needs of ordinary people and ignore ordinary people’s suspicion of, and resistance to, AI generally, they may end up popping their own bubble.
Thanks to Roger for today’s question! The next installment will be on Wednesday, December 31.
All posts in this series are tagged “AAA-25”. All previous posts from every “Ask Arthur” series are tagged, appropriately enough, ”Ask Arthur”.
Previously in the 2025 series:
”Doing the annual inquisition AAAgain?” – The first post in this year’s series.

No comments:
Post a Comment