It's hard to argue with the headline point. AI is one EXPENSIVE business to build, and it's a huge on-going expense to RUN.
The cost of most brands and models of computer, gaming console, phone, and SSD has already GONE UP, and projections are for prices to stay high and perhaps go higher over the next two or three years. All because data centers are consuming so much memory and other electronics that there just isn't enough LEFT on the market for Apple, Dell, and everyone else.
And building a chip fab is a multi-year and multi-billion-dollar project, so the chip famine won't be going away soon. And if/when the AI industry gets sated and stops buying at such scale, those new fabs will be short on customers. Damn. Capitalism is scary; no wonder all the cool kids want to socialize their losses while retaining ownership of their profits.
Electricity prices are rising also, with AI sucking so many electrons out of the grid that utilities are having trouble keeping up.
How large are these monsters? Brave AI has this to say:
Current AI data centers have transitioned into the "Gigawatt Era," with the largest facilities consuming as much power as major metropolitan cities. The largest AI data center by IT power is Colossus 2 (owned by xAI in Memphis, TN), which draws approximately 946 MW of power, followed closely by Anthropic-Amazon New Carlisle at 910 MW and Microsoft Fairwater Atlanta at 636 MW.
In terms of physical scale, these campuses are massive, with some spanning thousands of acres—OpenAI’s Stargate project in Abilene, Texas, covers 875 acres (3.5 square kilometers).
Does any of that sound like "free" or "cheap?"
https://www.oftwominds.com/blogjuly26/AI-subsidies7-26.html
Let's pull all this into an undeniable conclusion: AI is based on massively subsidizing users' costs.
What's already abundantly clear but verboten to say as it would pop the bubble of AI valuations and triumphalism is that AI is unaffordable once the direct and indirect subsidies are withdrawn. Nothing that consumes this much electricity and requires such an immense scale of costly processing and memory capacity can be low-cost, never mind free.
The major AI platforms and vendors are subsidizing corporate and individual users in the hopes that they can achieve AI sector dominance --and the pricing power that comes with it--via the network effect, the dominance generated by having the majority of users bound by habit or dependence to your platform or tools. [Editorial comment: I very much doubt that's the only reason, but . . . that's another story, and scarier than this one . . .]
This battle for network effect dominance is playing out in full view:
AI Giants Are Handing Out Tons of Free Computing Power to Grab Startup Share: (wsj.com)
Pitched battle for business users comes as AI companies seek lasting streams of revenue.
Hans Ibarra, a founder building an AI-voice startup, has found himself on the receiving end of a big opportunity: Top artificial-intelligence companies such as OpenAI, Anthropic and others desperate to win his business are ramping up discounts.
Across Silicon Valley, startup founders like Ibarra are enjoying a wave of computing credits and fielding competing offers from AI-model makers racing to land new enterprise customers. Cursor, the AI-coding company bought by Elon Musk's SpaceX, offered a 75% discount through July 5.
"If I'm choosing between a really cheap Chinese model that I actually have to pay for, and a very expensive Anthropic model that I don't have to pay for, I'm going to pick the Anthropic model," Acker said. "I'm always going to pick the one for which I have free credits."
Meanwhile, back in the real world of costs, AI Costs More Than The People It Replaced (forbes.com)(via Tom D.)
It turns out that experienced human workers doing the work right in the first place is cheaper than having AI run a probability distribution process that needs vetting and corrections. And remember, AI isn't actually "intelligent," it's just a probability distribution using natural language.
. . . There are many other hidden subsidies within the AI machinery. There are corporate tax write-off subsidies, energy subsidies, tax credit subsidies for building data centers, and so on. If these were stripped out, what would the real unsubsidized costs of AI be? No one knows, but they would be higher than what's presented as the cost now.
(more)
AI is like most other emerging technologies, its gonna go through a period of deep subsidization while its in the growth stage since no country can really afford to fall TOO far behind and let a rival nation get too far ahead of them.
So that's not really news. Even traditional "Heavy industry" runs off of subsidies for NEW construction. A new oil refinery or steel mill goes up? You better believe there's subsidies involved SOMEWHERE. It may not be the feds subsidizing it (though they often do), but the state and local governments will bend over backwards to get those plants up and running. Multi decade tax agreenments that either freeze or cancel things like property and excise taxes, state taxes, etc. etc. That's just the nature of large investments. Its why companies doing stuff like this will often shop around and see what state, city, county, etc. will give them the best deal before moving forward.
As for AI usage, it was REALLY overhyped as this weird thing that would replace EVERYTHING. Its obviously NOT, but it DOES have it's legitimate uses that are never gonna go away now that the genie's out of the bottle. Anthropic for example, as far as I'm aware is the most profitable of the "pure play" AI companies (spaceX doesn't count since they also do rockets and starlink, as pure play). Likewise its been PROVEN at this point that Ai is valuable, if nothing else, as a data and information tool. Google, very briefly LOST market share for the search industry and fell below 90% for the first time since they basically became a thing. That's half the reason gemini (the google AI) exists. Data and information are INFINITELY more valuable than ANY AI company's direct profits, so even IF these AI companies fail completely, the tech giants aren't gonna let them go away. They'll snap them up in a fire sale and keep them propped up JUST for the data they bring in on the people using them.
Don't get me wrong I don't agree with the whole "bulldoze every neighborhood and turn it into a data center" schtick, the complete opposite really. But you might as well accept, AI is here to stay, the exact same way people argued against telephones, the internet, and cars because the infrastructure had to be subsidized for all three before they became "normalized". Ai isn't going anywhere at this point.
I agree, AI isn't going anywhere. It's very useful and new uses will continue to crop up. Some will be awesome. (Not always in a good way, either).
A lot of investor's and taxpayer's money IS going somewhere, however -- down the drain. As with many other new technologies, some companies will do well and a larger number will do poorly or go completely bust. AI seems more over-sold at this point, especially given the costs involved, than most other new tech areas have been.
And I think THAT is partly, maybe heavily, because the Surveillance State wants to know everything about everyone all the time, and that requires massive data storage and huge compute resources.