|
Aardvark DailyThe world's longest-running online daily news and commentary publication, now in its 30th year. The opinion pieces presented here are not purported to be fact but reasonable effort is made to ensure accuracy.Content copyright © 1995 - 2025 to Bruce Simpson (aka Aardvark), the logo was kindly created for Aardvark Daily by the folks at aardvark.co.uk |
Please visit the sponsor! |
The AI industry can't keep pouring billions of dollars into setup costs, as they have been for years now.
Eventually, investors want to see not just a return on that investment but significant profits yet those profits seem very unlikely.
Like a house of cards, the entire AI industry seems to be reliant on a raft of tenuous interconnections and the failure of any single one of those dependencies could bring the whole thing tumbling down.
Have we just seen the first sign that this bubble of hope is bursting?
I refer to a company called Character.ai that has just gone into bankruptcy.
Character.ai is a typical AI startup. It has attracted huge sums of money and burnt it all, without turning a single cent of profit.
According to the company's bankruptcy filings they have admitted that their business was "wildly unsustainable".
That phrase is one we might hear used a lot more in the months and years to come as investors decide to pull the plug on an industry which has become a giant money-pit of loss.
It's not that some AI companies aren't generating significant revenues, the problem is that those revenues fall far short of the costs of operation. This means that they are far from profitable and there's no real indication that they will ever be able to deliver their services at a price which would create those profits.
Huge numbers of people pay $20 a month to use ChatGPT for example but it would appear that all of those accounts run at a loss.
It's the cost of providing AI services that is the real hurdle. When someone spends $5 to create a 20 second AI video it apparently costs much more than that just in the amount of electricity used to do so. Then there are the costs of amortizing the hardware, infrastructure and realestate involved in the datacentre where that AI runs as well as the myriad of other overheads.
You can see just how expensive AI is to deliver when you consider the recently "solved" the Navier–Stokes existence and smoothness problem. How fantastic that AI managed to do this but guess how much it cost to do so?
Twenty million US dollars, that's how much.
Such a win is great PR for the industry but how many research institutions have that sort of money to throw at solving mathematics problems on a regular basis? I would wager that the answer is "almost none".
Character.ai's bankruptcy is the proof of something we've suspected all along. AI, whilst a valuable tool, is not something that can be delivered at scale at an affordable price and therefore it's not a sustainable business.
Once the investor community wakes up to this fact, as they may be about to, the flow of essential capital to the AI industry will all-but dry up. In fact the CEO of Broadcom recently cited a lack of capital as the biggest hurdle to AI growth. If that capital supply dries up even further then we can expect to see a cascading domino effect cause a wave of bankruptcies to wash across the entire industry.
That's how the AI bubble bursts.
Making life even worse for these nascent AI companies is the growing uncertainty in regards to intellectual property rights. Even while they're far from profitable, AI companies are now facing huge uncertainty in respect to their legal liabilities after training on copyrighted material scraped from the internet.
Groundbreaking AI music company Suno has lost boatloads of customers after it struck a deal with several music labels inluding WMG and BMG. As a result of this deal, designed to protect it against massive litigation, Suno had to massively reduce the rights of its customers to use the music generated by the service so many customers just bailed. Another AI business model blown to smithereens.
Finally, as I have mentioned in a previous column, the rise in capabilities of local AI is further invalidating the cloud-based services as a sensible option for many businesses. There's no need to pay a huge sum every month to a cloud-based AI provider when for just a few month's worth of such charges it is possible to build an in-house AI server that will perform more than "good enough" to do the same job.
Hold on to your hats folks... it's about to get pretty rough out there. If you thought the dot-com bust was bad you ain't seen nothing yet. Find a comfortable char and grab some popcorn!
Carpe Diem folks!
Please visit the sponsor! |
Here is a PERMANENT link to this column
Beware The Alternative Energy Scammers
The Great "Run Your Car On Water" Scam