*Corporate leaders are starting to question whether soaring AI spending is delivering meaningful returns.*
***Why it matters:*** *Companies that rushed to embrace AI are now confronting ballooning IT costs, uncertain productivity gains and growing employee skepticism.*
***Driving the news:*** *Microsoft canceled most of its Claude Code licenses, in part over costs, according to The Verge, and Uber’s COO said AI costs are getting “harder to justify.”*
*- An AI consultant tells Axios one of their clients recently spent half a billion dollars in a single month after failing to put usage limits on Claude licenses for employees.*
*- Companies are citing AI’s ability to automate jobs as a cause for layoffs, though Anuj Kapur, CEO of CloudBees, told Axios that workforce cuts may simply be “the only lever they can pull” to offset their AI bills.*
*- Consumer sentiment around AI is also nosediving, and employees are rebelling against the use of the technology at work.*
kodos_der_henker on
Everything was sold at a loss to get people involved and dependent on a product and once it is either established so that no one really removes it, or operating at a loss not longer possible, the prices increase to finally make money
Add on that with AI (that isn’t Artificial Intelligence but a tool for intelligent, so marketed the wrong way to increase the hype) the gains are not there yet and the promised savings resulted in much higher costs.
We don’t have real numbers for the actual operation costs of the AI companies but the estimates are that 1 Million tokens cost around 100$. If true the sticker shock isn’t even there yet as you can expect ~200$ or more for 1M as a reasonable price for OpenAI and others to make the money back they already spend.
Inevitable-Dig8702 on
In other words, most CEOs are insecure lemmings.
Chill_Panda on
See this shit is why I keep ringing alarm bells at just how much energy and memory AI actually takes.
Once the data centres are up, the only running costs for building and hosting your own AI is energy and memory.
However, AI makers are still not turning a profit.
But companies are paying boat loads to these AI companies, one company sent half a billion to Claude alone, and yet Claude is not profitable.
When your only running cost is the electric bill, and companies are paying you obscene amounts of money, and you don’t turn a profit, something’s going wrong.
but-I-play-one-on-TV on
Oh, well this is fine. It’s not like the national economy is dependent on the success of like five companies that sell AI products back and forth to one another.
duckduck-a-go-go on
They’ll cut even more workers to make up for the increased cost of cutting workers.
amonra2009 on
lol, wait until they found out if they want their AI worker to be more productive they need to pay him more.
Altruistic_Hat_9990 on
turns out “we’ll just add AI to everything” is a lot more expensive than the pitch deck made it look
SmallRocks on
The *Roaring 2020’s* is going to be an interesting history read a couple decades from now.
Moontoya on
Being in IT is a Cassandrean existence
We fucking told you so and you refused to listen
jcmacon on
My favorite part of this post is the advertisement for an AI product before the comments section.
Unusual_Flounder2073 on
My coworker said out Claude costs last month were $16,000. And we have maybe 20 engineers at our smallish company ($100m revenue). And I think that is not including the work to integrate AI into our product. And it’s like 3 people driving the majority of cost.
MagicCuboid on
I feel like AI made CEO jobs so easy that they assumed it would be the case for everyone else who does actual work, so they bought in hard.
coppercactus4 on
Lol, I work for a AAA game studio ans AI has been pushed hard. With Copilot they give you the selection of which model to use. Of course I am going to select the one that is 14x the cost.
They disabled that model very recently because it costs too much.
Juicymoosie99 on
Back to using offshore labor in India and pretending we can’t afford to hire more Americans!
FanDry5374 on
So…the little people will subsidize the AI industry through intellectual “contributions”, higher environmental costs, electric costs, water usage costs and then Corporations will buy in, fire employees (more peons paying out) and end up with higher costs and worse productivity outcomes but the tech overlords will be making bank. Sounds like business as normal.
TheDoneald on
Just wait until they run out of money to burn and the real need for profit pricing comes into play. We haven’t seen anything yet as far as costs go. If and when local governments finally become hip to the lack of benefit to having a data center in your back yard and tax benefits and subsidized utilities are dropped, perhaps we can move forward as a civilization.
BurdTurglary on
They’re all gonna get bailed out, aren’t they
DonorBody on
So they’ll be paying for useless tech that’s been integrated into everything they do and they cant get rid of it, and they’re also paying for the real estate they forced everyone to come in to because they think WFH is people milking the clock, but those buildings are no longer at capacity because they laid a bunch of people off to fund the AI, and they can’t get rid of the real estate because they’ll never get back what they paid for it. Neat.
Jorycle on
My employer went from “we’re not going to jump on buzz words” and carefully evaluating LLMs to “everyone should be using LLMs for all work they do” almost overnight. There was a big meeting where a manager said something along the lines of “LLMs should be writing most of your code now, you’re just the AI babysitter.”
I think our spending went from an organization-wide $50,000 monthly cap to spending that much in just a day after the new policy.
I don’t think it’s true that it’s difficult to see the gains, but I have a feeling we’ll be scaling back to more responsible usage. At the end of the day, even if I produce 5 times more code, it can’t be shipped faster than the humans reviewing it and no one is dumb enough to merge anything that hasn’t been reviewed.
h0tel-rome0 on
CEOs next fiscal year “we’re cutting costs by dumping AI and hiring humans” and they’ll come up with some silly new business speak for it
According_Jeweler404 on
The best part is that some companies were able to lay off huge numbers of technical workers, blame it on AI, let skeleton crews absorb that displaced cognitive load, AND ensure existing workers think twice before asking for a raise or doing anything to remotely hurt their reviews.
In short, it’s been a dream really (for top-level leadership).
PartyHippopotamus on
There will be no long term savings for these companies. The AI company’s revenue capture target is the salaries that their tools replaced. Consumer costs are low now to allow for aggressive expansion and adoption. Once the services have been fully integrated, the prices will start rising. Other companies will start advertising “real people” as a competitive edge, and we’ll have come full circle.
Dash_Rendar425 on
Part of the problem is that it sucks.
Companies are just pushing, pushing, pushing this shit, and it fucking blows.
It’s also making stupid people more stupid and more lazy because they’re just putting garbage into AI and it’s producing somewhat readable emails. Which suck and half the time don’t make any sense.
pm_me_ur_fit on
I’m can’t wait to deny using AI at work for religious reasons. Thanks cool pope
25 Comments
Excerpts from [article](https://www.axios.com/2026/05/28/ai-spending-roi-enterprise-costs) by Madison Mills:
*Corporate leaders are starting to question whether soaring AI spending is delivering meaningful returns.*
***Why it matters:*** *Companies that rushed to embrace AI are now confronting ballooning IT costs, uncertain productivity gains and growing employee skepticism.*
***Driving the news:*** *Microsoft canceled most of its Claude Code licenses, in part over costs, according to The Verge, and Uber’s COO said AI costs are getting “harder to justify.”*
*- An AI consultant tells Axios one of their clients recently spent half a billion dollars in a single month after failing to put usage limits on Claude licenses for employees.*
*- Companies are citing AI’s ability to automate jobs as a cause for layoffs, though Anuj Kapur, CEO of CloudBees, told Axios that workforce cuts may simply be “the only lever they can pull” to offset their AI bills.*
*- Consumer sentiment around AI is also nosediving, and employees are rebelling against the use of the technology at work.*
Everything was sold at a loss to get people involved and dependent on a product and once it is either established so that no one really removes it, or operating at a loss not longer possible, the prices increase to finally make money
Add on that with AI (that isn’t Artificial Intelligence but a tool for intelligent, so marketed the wrong way to increase the hype) the gains are not there yet and the promised savings resulted in much higher costs.
We don’t have real numbers for the actual operation costs of the AI companies but the estimates are that 1 Million tokens cost around 100$. If true the sticker shock isn’t even there yet as you can expect ~200$ or more for 1M as a reasonable price for OpenAI and others to make the money back they already spend.
In other words, most CEOs are insecure lemmings.
See this shit is why I keep ringing alarm bells at just how much energy and memory AI actually takes.
Once the data centres are up, the only running costs for building and hosting your own AI is energy and memory.
However, AI makers are still not turning a profit.
But companies are paying boat loads to these AI companies, one company sent half a billion to Claude alone, and yet Claude is not profitable.
When your only running cost is the electric bill, and companies are paying you obscene amounts of money, and you don’t turn a profit, something’s going wrong.
Oh, well this is fine. It’s not like the national economy is dependent on the success of like five companies that sell AI products back and forth to one another.
They’ll cut even more workers to make up for the increased cost of cutting workers.
lol, wait until they found out if they want their AI worker to be more productive they need to pay him more.
turns out “we’ll just add AI to everything” is a lot more expensive than the pitch deck made it look
The *Roaring 2020’s* is going to be an interesting history read a couple decades from now.
Being in IT is a Cassandrean existence
We fucking told you so and you refused to listen
My favorite part of this post is the advertisement for an AI product before the comments section.
My coworker said out Claude costs last month were $16,000. And we have maybe 20 engineers at our smallish company ($100m revenue). And I think that is not including the work to integrate AI into our product. And it’s like 3 people driving the majority of cost.
I feel like AI made CEO jobs so easy that they assumed it would be the case for everyone else who does actual work, so they bought in hard.
Lol, I work for a AAA game studio ans AI has been pushed hard. With Copilot they give you the selection of which model to use. Of course I am going to select the one that is 14x the cost.
They disabled that model very recently because it costs too much.
Back to using offshore labor in India and pretending we can’t afford to hire more Americans!
So…the little people will subsidize the AI industry through intellectual “contributions”, higher environmental costs, electric costs, water usage costs and then Corporations will buy in, fire employees (more peons paying out) and end up with higher costs and worse productivity outcomes but the tech overlords will be making bank. Sounds like business as normal.
Just wait until they run out of money to burn and the real need for profit pricing comes into play. We haven’t seen anything yet as far as costs go. If and when local governments finally become hip to the lack of benefit to having a data center in your back yard and tax benefits and subsidized utilities are dropped, perhaps we can move forward as a civilization.
They’re all gonna get bailed out, aren’t they
So they’ll be paying for useless tech that’s been integrated into everything they do and they cant get rid of it, and they’re also paying for the real estate they forced everyone to come in to because they think WFH is people milking the clock, but those buildings are no longer at capacity because they laid a bunch of people off to fund the AI, and they can’t get rid of the real estate because they’ll never get back what they paid for it. Neat.
My employer went from “we’re not going to jump on buzz words” and carefully evaluating LLMs to “everyone should be using LLMs for all work they do” almost overnight. There was a big meeting where a manager said something along the lines of “LLMs should be writing most of your code now, you’re just the AI babysitter.”
I think our spending went from an organization-wide $50,000 monthly cap to spending that much in just a day after the new policy.
I don’t think it’s true that it’s difficult to see the gains, but I have a feeling we’ll be scaling back to more responsible usage. At the end of the day, even if I produce 5 times more code, it can’t be shipped faster than the humans reviewing it and no one is dumb enough to merge anything that hasn’t been reviewed.
CEOs next fiscal year “we’re cutting costs by dumping AI and hiring humans” and they’ll come up with some silly new business speak for it
The best part is that some companies were able to lay off huge numbers of technical workers, blame it on AI, let skeleton crews absorb that displaced cognitive load, AND ensure existing workers think twice before asking for a raise or doing anything to remotely hurt their reviews.
In short, it’s been a dream really (for top-level leadership).
There will be no long term savings for these companies. The AI company’s revenue capture target is the salaries that their tools replaced. Consumer costs are low now to allow for aggressive expansion and adoption. Once the services have been fully integrated, the prices will start rising. Other companies will start advertising “real people” as a competitive edge, and we’ll have come full circle.
Part of the problem is that it sucks.
Companies are just pushing, pushing, pushing this shit, and it fucking blows.
It’s also making stupid people more stupid and more lazy because they’re just putting garbage into AI and it’s producing somewhat readable emails. Which suck and half the time don’t make any sense.
I’m can’t wait to deny using AI at work for religious reasons. Thanks cool pope